Legal Case Summary

Next case is Carnegie Mellon. Miss Sullivan, you want 25 and 5? Yes, you're right. Go ahead and start when you like. Good morning, honors and may it please the court, Kathleen Sullivan from our bell. We respectfully urge the court to reverse the judgment below because CMU's patent claims are invalid as anticipated by Wursel and because Marvel does not infringe them. Alternatively, we respectfully request that you at a minimum, remain for a new trial on damages because no reasonable jury could have found that 50 cents a chip on worldwide sale represents a reasonable royalty for any domestic infringing use of CMU's claims. I'd like to cover in the living very briefly if I may. It boils down to two very simple points, your honors. Wursel anticipates both limitations in the CMU claim. The key limitations are that there be a set of signal dependent branch metric functions and that the applied trapezoidal reality of signal pamphlets. You need no further than the plain language of Wursel to find both of those limitations disclosed. I respectfully refer you to Wursel column 10, line 52. That's where you get the set of signal dependent branch metric functions. That seems for all the world to say that there is a single and a deviation based number, it doesn't matter whether it's squared or not, but I don't care about that. But it's a single one used for every single column in the trial. Assume that that signal dependence, which I'll prepare to assume, I'm going to assume here that set actually means plurality rather than a trivial set. So how have we gotten to two? Greater than your honor. The disclosure at line 52 is that the fraction derived from sigma will be greater for the one branch of zero branches. That means there's going to be two equations. It doesn't matter if you multiply the zero branches by one. That may mean that you are selecting among functions based on a signal. That doesn't mean that what you select in zero branches is a signal dependent metric. Well, you're on and out. I think this is the idea that the testimony that you select for the zeroes a non-signal dependent function, namely equation 20, which is not single dependent. I think your own witness said about four times. And you select for the one something that is single dependent, but because it's constant, it doesn't vary from time to time to time. What column 10 line 52 just was is that there will be a difference between the function for the one branch and the zero branches and the zero branches. And if I could refer your honor in column 10, the key equation 20 is the starting point. But column 10 asks that you further modify. What worth the number does is you further modify the modified branch metric function. And by multiplying by some number dependent on the standard deviation. Yes, your honor, but the standard deviation expressing the transition noise is going to differ between the zero branches and the one branches

. And is that that produces the plurality and is the set? The fraction is, it doesn't matter if you keep the zero branches constant or multiply them by one, the comparison to the one branches is going to yield two different equations, two different functions. Because the sigma fraction is going to differ for the one branches and the zero branches. Your honor, as to the plurality of, that's our argument. We would just respectfully suggest that CMU's own expert conceded both the transition noise as a form of single dependent noise. And that the, that the column 10 disposes a set, all he'd disagree about is whether both equations in the set were single dependent. We believe that's impossible to argue when he's conceded that the, that the, that signal transition noise, which is measured on both branches is single dependent. So, your honor, that's the envelope of the argument. Just to be clear about one thing, you're not suggesting that the sigma referred to, I guess not, not by letter, but standard deviation is one that is being updated over time. It is in fact the same number that is used as a multiplier against the rest of the, against equation 20 everywhere you have a one branch. Correct, your honor. But the difference is between the one branches and the zero branches and that's the set. And your honor, as to plurality, again, it's simple from the plain language of the patent column 12 line. At the beginning at lines 48, the modified metric and this by this he's referring to the further modified metric takes into account not only a presently sampled input signal, but a previously sampled signal as well. So, your honor on invalidity, we respectfully suggest that it was not a matter of the battle of the experts, the plain language of worst all coupled with McGlauff and admissions with respect to worst all makes a case for invalidity. I'd like to just stay on worst for a second. It talks about transition noise. What do you understand transition noise domain? We understand it to mean signal-dependent noise. I know that, but let's just talk about this spec and what worst will mean when he's talking about transition noise. Is that talking about when you go from a first bit to a second bit, if you flip the plurality on the second bit compared to the first bit, then there will be some resulting noise at the transition on the flip of the plurality? Yes, your honor. And where there's no transition, the noise will be different from the noise when there is a transition. I guess I'm wondering why is that signal-dependent noise based on the claim construction provided for signal-dependent noise, which was a noise structure attributable to a specific sequence of symbols, a single written symbols. So, why is the flip of plurality between two bits the same thing as a specific sequence of symbols? Because the transition or no transition is expressed in a series of bits, ones, and zero, and we believe those are covered by the term symbols. But why is symbols the same thing as bits? Because when I look at the 839 pattern and I try to understand what the discloser says, it repeatedly makes reference to symbols and written symbols with reference to these funny symbols. So, the signal-dependent noise is a single, that they plus, plus with a circle around it, minus, minus with a circle around it. And so, what the 839 pattern is telling me is that when it's talking about a specific sequence of symbols, it's talking about those particular pluses and minuses, which represent transition between two different bits. So, now I'm beginning to wonder if the best understanding of the claim construction is, it's referring to a noise structure attributable to a specific sequence of transitions. And what Worsal is trying to remedy is noise that a single trend is. So, do you see why I'm now seeing that there is daylight between Worsal as discloser on its face, in my understanding of what is being claimed when the claim references single signal-dependent? You're on a reduced to your point, but the claim construction did not construe as narrowly as you described. And there's nothing to suggest that the transition noise equations that are disclosed in Worsal don't themselves embody a sequence of symbols, because the bits represented by one and zero on the transition branches and non-transition branches, we think that does fit within the term symbol as construed. And if Dr. McLaughlin talked about how all Worsal has is a single transition, and that's not the same thing as a specific sequence of written symbols, and I didn't see anything by Dr

. Proakis countering that by saying, well, Worsal is really more than just about a particular transition of a flip of polarity between two bits. It can actually be more expansive than that, and that's really what Worsal is contemplating, it's contemplating a sequence of transitions which would require more than two bits, which would require at least three bits. Your Honor, that is a possible reading of Worsal, but at a minimum here, our interpretation of Worsal renders the willfulness judgment here inappropriate, because we certainly have an objectively reasonable defense, because there's ambiguity in the term symbol. And because we reasonably believe that there was a set of signal dependent branch metric functions, and we reasonably believe that it was applied to a plurality of signals. And the further details of the CMU patents are not necessarily relevant to the particular claims that issue here. There may be other differences between Worsal and CMU. In fact, Worsal and Self said there might be some differences, but the claims here are really quite simple. And set of signal dependent functions and plurality of signal samples we think are covered by the claim to. What do you think is the Worsal meant when he said that this work by cab check goes above and beyond Worsals on work? We don't know your Honor, but as narrowed for trial, the covariance matrices dropped out of the case, and to the extent that covariance matrices were what CMU pointed to as part of the complexity of their analysis, that may have been what he was referring to, but we don't think. We totally reviewed the work at that point. No, your Honor, he had just reviewed the disclosures. So, Your Honor, if I could, because I really do want to focus on the damages issues in the case, the extremely brief about non-infringement, again, I want to suggest that at a minimum, we had a reasonable, an objectively reasonable, non-infringement defense. Sorry, just real quickly. One last quick question on validity, and maybe I didn't catch it in your first colloquial gesture on to. So, what is, well, I'm sorry, you go ahead. We are on a non-infringement very quickly. The MNP chips perform the analysis in a post-process or not in the trellis. You might say, why does the analysis have to be in the trellis? The answer is the court's construction to which CMU stipulated of the term branch, which modifies both branch metric function and branch metric value. Do I, do I, do I remember in writer wrong that there was a quite substantial dispute in the clean construction process about whether the post-process there was going to be excluded or not, and the court ruled against Marvell and for CMU, in what way of you making a point different from what was rejected against you? You're on simply that the, to include the post-processor in the application of the function is incorrect under this stipulated claim construction about branch, because branch is defined as a transition between two nodes or states in a trellis. And that's confirmed, of course, by Castix admission that his claims operate in the trellis and not in the post-processor. So, you're on a, without, however we draw the line around the thing that is doing the detection, we think the application staff has to be performed in the trellis by virtue of the claim construction of branch. On the NLD chips, the argument is very straightforward. CMU claims both require application to a morality of signal samples, our branch metric function, which is in the chip stipulation. So, it's part of the technical documents that are binding in the case on both parties, applies the branch metric function to a signal, signal sample that's outputted by the pre-felter. And that's represented by the parameter f sub y. So, no plurality there. And finally, on simulations, your honor, we would respectfully suggest that simulations are not covered by the claims. If you look at the, okay, can I just ask you about that? It seems to me that the patent itself makes clear contrary to a kind of ordinary English reaction to the word detector. That this is not a sensor of the magnetic properties of the hard drive. It is in fact a data detector, which is why it's separated from the sensor in the figure one, at which point if, and I guess there's evidence of this, that you get data from actual hard drives and then feed it into the processor, the, the perturb, like processor. It doesn't matter whether you call that a computer or not

. It's still a relevant detector because it's detecting that data by running a process to figure out what the true bits were. Correct, Your Honor, but in real time. And in fact, the opening of the A39 patent says that the patent is directed to high density magnetic recording sequence detectors and other aspects of the claims that we detail in the brief, referred to the notion of a real time hard drive data detector, not a simulation using a computer. And the admission again of CMU's own experts confirms that reading of the patent because McLaughlin admitted that the detector is in the trip. And we'd further suggest that, you know, if CMU now wants to say, well, this is just an algorithm, it would run into Alice problems because if it's just a patent in eligible mathematical equation that doesn't have a structure, we think it would run into problems of eligibility. But Your Honor, I would like if I could because they're complex to spend the rest of my time on the damages issues. And I'd like to begin by simply, let's break it out into the base and the rate and let me begin with the problems with the royalty base. And I'd like to begin with the instruction because that's what we think is the root of the problem here. The judge instructed the jury that it was entitled to include as a proxy for the domestic infringing use. All sales, including all foreign sales, all sales of chips manufactured sold and used abroad, all sales, including foreign sales, quote resulting from Marvel alleged infringing use during the sales cycle. Now, the problem with that, is there another place in the instruction that said you can't award damages based on the foreign sales. That's correct. In the preceding sentence, Your Honor, the court instructs, you may not award damages on the basis of the foreign sales. But it goes on to say you may consider all sales resulting from Marvel alleged infringing use during the sales cycle as they means of valuing the domestic infringing use. So she says you can't literally award damages for the foreign sales. But you can use foreign sales as part of the royalty base for creating this proxy for the valuation of the domestic infringing use in the sales cycle. Now, there's two problems with this, Your Honor. I'd like to start with the first problem by including foreign sales at all. We believe it violates the long line of precedence by which this court and other courts have said as a matter of law that you may not recover US patent damages for foreign sales. It doesn't matter if it's direct or indirect. If they're in the base, they're being used as a basis for calculation here. And obviously there are other ways. Do you generally agree that the law is different in the copyright world? I can't speak to that, Your Honor. I thought there were quite a number of cases cited, fourth circuit, second circuit, ninth circuit, others that allow the use of foreign sales to calculate copyright damages for domestic activity as long as the former followed from the law. So, Your Honor, the argument in the patent context is that the territoriality principles for patents law, which may well be different from that for copyright law, requires that a US patent rights holder does not obtain in a sense a valuable worldwide license. Now, I want to just say that there isn't a single case that CMU can cite in the patent context as opposed to the copyright context. That suggests this kind of fruit of the poisonous tree theory that lets you reach out to the foreign sales. And power integrations is just the latest statement of this. Power integrations obviously have some differences from this case. That's what I'm wondering. Likewise, true that there's no case to this kind of fact that I'm going the other way in your favor

. I feel like we have here something of an in between case. I agree, Your Honor. I think there is no clear case either way. And what we would respectfully suggest is that you should make clear in this case that the principle of power integrations, which of course expresses a long line of territoriality cases in the patent context, requires you to send the case back for a new trial with an instruction that excludes the foreign trips from the royalty base. Now, let me say why I think the closest case to our side is power integrations. And I think the reason for that is power integrations obviously are involved in apparatus patent where sales were potential acts of infringement, where a method patent, the infringement here is used. That in our view makes no difference because what this court rejected in power integrations was the idea that near proclamation or reasonable foreseeability was enough to allow the patent holder to reach out into foreign jurisdictions. And include the foreign ships and the base. We think what happened here is even worse. Power integrations that least involved approximate causation notion about reasonable foreseeability. Here the instruction is just about four causation instruction. It says you can view the jury can use foreign sales as a proxy for domestic infringing youth if they resulted from. There's no qualification if they resulted from there's no definition. And that's the second error here. You're on a clean. What's the nature of the sales cycles? I mean, if you win a sales cycle, does that mean that you and your customer are essentially attached at the hips for the next four or five years and they're going to be making billions of dollars together? I mean, how does this exactly work? What are they agreeing to when there is a sales cycle win? Your honor, the answer to that question is the record does not say the record does not say for a very basic reason. We saw at many points pre trial and during the trial to exclude evidence of foreign sales as part of the royalty base. That was denied in the court repeatedly ruled both in pre trial motions, motions and women again in response to our trial objection that all sales, including foreign sales, could come into the case. So there is never an opportunity for the jury to decide what does a sales cycle mean? Where does the sales take place? There was never any joining a visual on that because the court's instructions precluded a definition of the sales cycle or a definition of where the sales took place. And that's an additional reason for you to read. There was lots of talk about design wins. There was a win when a design win occurs. Is there a contract? The record does not say you're honored. There was no evidence in the record that established where contracts were signed. It never mind where. Is there a contract when C-gate agrees with Marvel, we're going to use your chip. Is that embodied in a contract? Yes, your honor. But the evidence in the record is that a requirement contract? I can't answer that because it's not in the record. And I think this is one of the central reasons why you need to send it back. I mean, I could try to give you answers from personal knowledge, but that isn't the point. The point here is that the record does not have any evidence of what sales cycle means, where the contracts took place

. And as this court makes clear recently in Halo, a mere offer in the United States is certainly not enough to establish that a sale took place in the United States. You might want evidence about where contracts were signed, where sales were made. So you're on two separate problems with the instruction. One is it included foreign sales. You have repeatedly said as a matter of law that foreign sales can't be the basis for determining a royalty base. And you ask what case we have the closest case for us is power integrations. The only case for them, your honor, is railroad dynamics. And we think that railroad dynamics is a case that allows the inclusion in a royalty base of some foreign sales of concepts containing infringing snubber components. And it does so in the only context in which you've ever approved the use of foreign sales as part of the royalty base. And that is where the infringing concepts were domestically manufactured as a manufacturer was infringed with the underlying concern about extraterritoriality as a common sense matter is trying to prevent duplicative regulation. How could something why doesn't the railroad case present exactly that problem? The making and the selling, whether making is here and the selling is abroad and the two countries may have quite different regimes or even the same regime and want to tax it twice with damages and yet that's okay. It does present your honor and we don't necessarily believe railroad dynamics would survive this courts approach in power integrations, which suggested that the extraterritorial principle should trump the principle of fair compensation and for the reasons that you described. So I'm not suggesting railroad dynamics is correct. I'm just saying that to the extent it's so good law, it doesn't apply to our case because in our case, you've got a long attenuated chain of causation between the domestic act of infringement, the use of demonstrations and simulations in Santa Clara to create a relationship that becomes contractual at some point not specified in the record. That's very different. There's a one-to-one correspondence between the infringing manufacturer and railroad dynamics and all the other cases that CMU has cited for this principle and they form down. Where there's a failed cycle win that ultimately there's a breakdown on the way to ultimate sale and manufacturing and cooperation of the acceptors. We do your honor. It's undisputed in this case that the chips are all manufactured in a foundry in Taiwan that was completed by CMU's own expert. We also believe that when power integrations said that the act of foreign sale, manufacture or use creates an intervening cause, the breaks, the chain of causation, there is a real problem with this kind of case. The reason why the car set might be understood as an acceptable proxy is that the car set is not a consequence of the infringement in the United States. It's literally a measure on a one-to-one basis of the infringement in the United States because the thing is completed before it leaves our borders. We think that our case is dispositively different from railroad dynamics and it would take a new car bat from the territoriality principle for this court to affirm. I see my time is nearly up. I wonder if I could switch to the royalty rate and just briefly make a few quick points. I'm sorry, on the base there may be one more thing to say if I can interrupt myself. On the base, I think we may be arguing about what's wrong with the instruction. It includes foreign sales and the resulting from the but-for-closation piece is undefined, the court compounded the error by not saying what the sales cycle means and the parties were precluded from inputting evidence. I'm aware sales took place, an additional reason for remand. But on the base, I just want to say why a reason? I'm not sorry. I just want to follow up on one thing though I'm not sure if it matters

. I was not seeing a preclusion of evidence on the question of where the sales took place. Actually, you're on a district court of summary judgment and which I think I forget the citation. I thought that this court said near the end this question is an issue for trial. The district court repeatedly said, I'm letting all sales come in regardless of place. We regard that as effectively precluding or rendering futile. I'm not sure I would go on the base. On the base, just one last quick point. Even if you think this instruction is acceptable, we don't. We think you must remain for a new trial, giving clarifying principles about foreign sales and the causation standards. No reasonable jury could have found that the foreign sales would have been included in a hypothetical negotiation in this case. So just on the record here, three quick points. The parties must be assumed to bargain in the hypothetical negotiation in the shadow of the law. CMU had no evidence of any analog foreign patents here. So CMU was not in a position in the hypothetical negotiation here to bargain for a worldwide license. Second, quick point. There was no industry practice data or any other real world data point in this case whatsoever to suggest that percentage of worldwide sales was ever the kind of license given in this industry for this kind of venture project of the university to the contrary. As I won't be labor, all of the real world evidence in this case, such as the $200,000 a year offered until... The $200,000 is a appealing economic rationale of everybody getting money if the thing turns out to be a success. And you're getting basically three quarters of it and they're getting one quarter when the alternative is maybe much, much worse for you. There's no evidence that Marvel would have signed a license that $600,000 or $6,000 times more expensive for it than it could have gotten by joining the DSSC or taking the deal offered to Intel. Or taking the deal, the $2 million annual deal that CMU said posts the date of the hypothetical negotiation. 2004, 2005, CMU says a great optimistic scenario would be $2 million from Marvel. So there's no evidence that the willing license for here would have contemplated this license. And there's no evidence that the willing license see here would have contemplated it because if Marvel was faced with paying $600,000 or $6,000 times the nearest historical license, it could have moved off short for the chip cycle. So, you know, I see I'm almost out of time for my rebuttal on the right point. We covered in that brief, but I can cover it briefly if you're on our once or I can save it for Rebuttal. What about, just if I could move on or go back in a second. So, you're basically saying the hypothetical negotiation would be restricted to actual US usages and sales. Yes, you're honest

. So, what about Ms. Lawton's reference to a calculation that resulted in something like 556 million chips making it back on the US towards? Yes, you are. So, I guess that would be something that you would agree is an infringement. Well, I don't agree with that, Your Honor, but first I asked your Honor, may I preserve my rebuttal time if I answer Judge Chen's question? Well, we'll give you two minutes. Okay, thank you. The answer to that, Judge Chen, is we don't concede that if the sales cycle begins in Santa Clara, the manufacturer and sale occurs abroad and the chips come back into the United States. We don't concede that that necessarily is a property piece of the royalty base. We don't concede it because there's been intervening causes. But, Your Honor, on this record, there was also a contributory and induced infringement verdict. And so, the notion that the chips that come back into the United States might be fairly encompassed, might be fairly encompassed by a secondary liability theory might allow you to sustain that piece of the verdict. But then, at most, the liability would be for the domestic chips that come back into the country. We don't think the evidence on that was proper. It was based on industry-wide PC data and not data specific to Morval. There's a lower number that's more data specific, 329 million chips. But, Your Honor, we don't concede. Of course, we don't concede infringement. But, we don't concede that causation that takes a round trip out of the country and back in is necessarily proper. You'd have to ask, is that the direct and immediate result? Do we want to extend causal chains that far to include it for direct infringement? So, to summarize, we don't concede on direct infringement. You might sustain those chips here if you had a contributory or induced infringement theory. But, Your Honor, last one, on rates. We've said in our brief that there were rates of use. Thank you, Your Honor. If I have things to address in the rate, well, thank you very much, Your Honor. Thank you. Good morning, Your Honor. May I please the Court, Josh Rosencrands representing Carnegie Mellon University. Your Honor, if I may, I'd like to take an unconventional approach and begin with damages first and then circle back to the merits and in particular willfulness. But, just to frame it all up in one package, this case truly is extraordinary. But, not for the reasons that Marvel has recounted, you almost never see copying as blatant or as reckless as this. You almost never see a case where the accused infringer has admitted and never argued that it has anything by way of a non-infringing alternative. You almost never see such a stark before and after picture of a company struggling in a life and death struggle about to go out of business and then rising to the top of the market where the company itself says that the reason is this particular invention

. But, on the damages point, in particular, the extraterritoriality point. That point, at least, I take it to be independent of any questioning of the tightness of the causation from all of the activity in the Santa Clara. And, Taiwan and other places that the resulting chips were made and delivered and most of them used. So, why is it that even if your basic narrative is accepted that the alternative to infringement here was Marvel going out of business and the amount of 25% of their multi-billion dollar profit resulting in the loss of the profit. It was a reasonable price to pay. Nevertheless, the extraterritoriality argument is you've reached too far. I understood your honor and actually the narrative links directly to extraterritoriality. Not just was it necessary for Marvel to infring to save its life. It was necessary for Marvel to infring in Silicon Valley. That was part of the no-non-infringing alternative. And that makes this case the diametric opposite of power integrations. In power integrations, there was absolutely no linkage between the domestic act of sales and before and sales. I thought I read power integrations differently where the opinion at least rejected the theory of awarding damages on foreign sales that were connected to or related to or caused by domestic infringement. The word was foreseeable. It was a completely crackpot theory of damages in that case. They were domestic sales. The plaintiff said because it was foreseeable that there would be foreign sales, we get to claw back the law sales that we would have made in foreign territories. This court went out of its way to say that the expert admitted that its damages estimate was not actually rooted in the defendant's activity in the United States, was not related to any parts that were manufactured, used or sold in the United States. So to segue back to Judge Toronto's question, here we have that definitive linkage. This may be the unique case, the extraordinary case where that linkage is very clear precisely because here is where the sales force was, here is where the design force was, 150 Marvel design employees out of 153. Here is where the customers and their design centers were. They had no choice but to be infringing for trillions of times per chip over the course of millions of chips hand in hand over the course of several years for each design win. In the red brief, you described the patents in suit as must have cutting edge technology. Does every other chip use it? Have you sued other chip manufacturers? The answer to the first question is yes. Every other chip in a hard drive has it according to our industry expert, Bioric. It has become industry standard. Everyone is asking for it so you sued other. The answer to the second question is no, we haven't sued. It was a big lift to sue Marvel. Now understand that the reason it took so long to sue Marvel is because you can't like pry open the chip and see that someone is infringing. You have to actually get their documents to figure out the nature of the infringement

. Now I want to turn to Judge Chan's question about what the record says about a design win because it supplies the rest of the Nexus. It is a winner take all market and on page 42, 124 you have a clear quote from our industry expert that says when you win you get all the chips on the drive in that generation. He says that he has never heard of a case where some other chip gets put into the drive once you get the design win. So yes Judge Toronto it is like a requirements contract. We don't actually have a record either containing the contracts or saying C.Gaid or Western Digital or the other five for customers or something. They commit to buying either a certain amount or everything that they are going to use when they make hard drives for a certain period. The best that we have is 42, 124 online for this is our industry expert and so to my knowledge no drive company has ever taken two reach channel suppliers to production on the same drive on the same generation at the same time that essentially says winner take all. Is there a reason why there would no. The transfer of partnership entered into record and that reflects what are the terms of the agreement between Marvell and. Well so the I would say Marvell completely punted on the question of where the sales occur. We put into the record that we were all about uses it was a winner take all market we had clear testimony about that Marvell is saying worldwide sales worldwide sales that will be court of the district court said Marvell did not produce any evidence that they were sales anywhere but in Silicon Valley. This was a design win the design win was the sale and the district court said that not only that but that there was but four calls established between the infringement and the design win. In the red brief at 61 you say lot and test five that both Marvell and CNU. So I was going to frequently entered into running Royal D. Agreement around the time of the hypothetical negotiation I couldn't find any evidence in the record to support that assertion. Well see you're out of I've got pages 43 335 to 36 where we're lot and talks about it and 34 024 were all. I looked at what you had in the in the brain but it didn't seem to me to support it. What you have in the way of other. Well you're on the the three licenses that were put in the record were all licenses on the same exact chips the systems on chips that our technology was incorporated into and all three of them were what Miss Sullivan refers to was worldwide sales it was the sale of all chips and that makes total sense as Judge Toronto was saying is an economic matter. Marvell doesn't even track where the chips end up or where it's customers customers end up using the chips what Marvell cares about and what in this hypothetical negotiation the parties would care about is that it's a winner take all design market and the infringement had to happen here. So let's share in the proceeds that come from that now I do want to talk I want to talk to you about 50 cents per chip absolutely honored. So the sales to the particular company miss I don't know what's confidential on here are not so I'm just going to I think it's all free okay so the sales to master of about 137 thousand chips which were included in Miss Laughan's analysis were associated with an operating premium of 47 cents. That and those figures were from 2003 2004 but Miss Laughan says in this industry the price is always going down so why is 50 cents a chip royalty reasonable at the time of the negotiation. Well see on our few things to say about it first the same licenses that I mentioned are per chip licenses. Marvell said we don't really care much about the form of the license what matters is is on a case by case basis how much money we're going to make the difference between 50 cents a chip or a certain percentage but the other thing that Miss Laughan testifies to is yes the chip prices are going down why because as you put the manufacturer more the cost goes down. The operating premium goes up the excuse me the operating profits go up as the cost go down because you actually make more profit per chip and so that's why it's reasonable and the only question here then becomes whether or methodology is reasonable or more specifically whether the district court aired in adopting in approving of her methodology is a matter of abuse of discretion. But before we have to deeply in the 50 cent per chip and I'm happy to go back I do want to make one other point about the question I think judge chin asked about whether there is a different standard in the copyright act rather than the patent act. I'm sorry just trying to I'm sure we're both interested in answer. So the answer is no there is one rule of extracurricular reality the copyright act has that rule in it is called the predicate act doctrine the predicate act doctrine says that when you when the predicate act here is the infringement that's what is necessary for it to occur in to it for it to occur in the United States but in keeping with rules and railroad dynamics once you have that predicate act in the United States you are allowed to measure damages based upon benefit the classic way of measuring benefit is based upon sales and it doesn't matter once you are allowed to do that. To measure based even on not infringing sales it simply doesn't matter why the sales don't in fringe whether they don't in fringe because this is a method patent and that's an apparatus or because this is a patent on domestic use and the ultimate sales were not infringing because they were

. So do you think in a case like Microsoft versus Asian teams where the Supreme Court said all of those foreign uses and sales are not infringing the Supreme Court nevertheless would have said it's okay to sweep in all of those sales into a reasonable royalty negotiation. Yes not only would the Supreme Court say it but the Supreme Court did say it in rules in rules the active infringement was domestic and what the plaintiff was allowed to get were royalties on the foreign sales as well. I want to ask it again because I want to make sure you say it again hundreds of millions of sales of windows abroad that Supreme Court ruled do not infringed those would never let be considered in a reasonable royalty negotiation between 18 and 20 Microsoft. If there is the clear nexus that we have established here between the domestic uses and the sales regardless of where they were and again I have to emphasize there's no evidence in this record that the sales were abroad that just report went out of its way to say. The Supreme Court did but you I thought quite conspicuously say practically nothing about that point in your brief all you do on page 83 is quote what J a 250 where the district court's conclusion but you don't make the argument that in fact under normal location of sales principles which I realize are not complete given the footnote and halo that nevertheless these sales actually did occur in the court. We didn't for lack of space but that is clearly true you know but but that is clearly true we are not distancing ourselves from what the district court said the location of these sales the only thing we've got in this record is that the sales that is from more about to its customers were not just negotiated not just design not just priced out but consummate it in the United States. But the reason we don't make much of that is because our theory was not about where the sales were consummated or theory was about the domestic uses and the domestic uses were essential I mean we're not we're not talking about an extravagant theory here we're talking about an extraordinary possibly unique linkage where everything that happened in the United States over the course of three to four years per cycle every active infringement had to happen in the United States. Precisely because that's where Marvell's customers are that's where the design centers are. I have to think about only the question of whether it was permissible to award this kind of damage based on the causation theory of the direct infringement related to the testing designing in the sales cycle and I don't have to ask the other question of whether or not all of these sales of these 10 billion chips actually happened in the United States. That's correct your honor that our theory was not about where the sales were or are theory under ningo and spectral it is and union carbide is once you have the active infringement that once you have an active infringement you can value the benefit to the infringer based upon the resulting sales even though in every one of those cases the sales were not infringing and when you marry those cases with gold and especially railroad dynamics railroad dynamics. When you've got the active infringement in the United States you are allowed to get the you are allowed to measure benefit based upon the non infringing sales including the non infringing sales in other countries. So you can ask you this suppose for purposes of this question that we were to think that the jury instruction was improper on extraterritoriality grounds and that a new trial was needed as a result because we don't know if the jury found on any other basis in such a new trial. So I take your position would be that at that point there could be a much more well joined evidentiary dispute about location of sales. Yes there could be but I need to emphasize one point about your hypothetical your honor. I realize you disagree. But even if the court were include that let's just be clear on what instructions were on the table there was our instruction the resulting from instruction which comes directly out of for example this court's opinion in king instruments which we didn't say but it says resulting from or Marvel's instruction which is sole cause only cause which is even more than just the same. So it's extreme than an EMVR instruction so what's what's the law on the question if their instruction was incorrect but so was yours and yours was adopted do they get to get a reversal they don't your honor what they get is the benefit of the jury improperly you're saying the jury was improperly instructed but forget about it. What I'm saying is if hypothetically the court concludes that the jury had the wrong instruction and the reason it was wrong is because it should have been a but for a for you know but but for instruction and Marvel never asked for you but for instruction absolutely Marvel waived that argument if you know the supposed to be error is it needed to have some sort of territorial limit in it. Is that preserved or waived or what it depends what the territory limit was what I mean Marvel's objection was that the district court aired and saying the act of infringement has to occur in the United States that is what the district court said do not count as infringement and right but you can count yes that's preserved but you can count all fail the entire benefit now that comes directly out of gold and railroad dinners. So you can count the benefits of all non infringing sales and by the way that's exactly how parties would be negotiating which is what we've already discussed. Well except that the hypothetical negotiation is hypothetical in a number of ways it's not tied to the specific parties as we recently said and not completely tied somewhat related to the position of particular parties but not completely tied otherwise for various reasons. It's also under a set of hypothetical assumptions including everybody knows that this patent is infringed everybody knows that it's valid and the question I guess is why isn't another one of those hypothetical constraints you may negotiate only about intra territorial activity whatever that may mean. So I was with you the extra territorial is tied at the hip in a causation way to the domestic activity. I was with you up until that last ten. Sure you're negotiating about infringement in the United States that's what the predicate act doctrine is about that's the only doctrine that the Supreme Court has ever used when it talks about extra territoriality. I mean Marvel has never cited a single case that is said damages need to be limited or the benefit needs to be measured only with respect to acts that excuse me only with respect to sales that are domestic. I mean Sheldon which is versus MGM which is the learned hand opinion that talks about the price of the predicate act doctrine that was about making of a negative in the United States. But what ended up happening because the negative was made in the United States and exported abroad was all these copies were made abroad those acts were all abroad but because the benefit to the infringer the infringer in infringed here and because the benefit to the infringer is measured by sales the act of infringement is measured by all of those sales. So to reach a different conclusion here would mean that the court has to conclude that there's a special patent law rule governing extraterritoriality that doesn't apply in any other context. Now I did promise to turn it on the panel. At page 50 in the blue brief the argument has made that I think you probably agree that at least actual licenses are probative evidence of the proper form of law structure. And then they go on to say the only evidence of any contemporaneous actual licensing was flat fee. And I wrote a note to say I think that's right but I wanted to give you a chance to discuss it. So what I think they say is that the only evidence with respect to these specific states, the only evidence of any contemporaneous actual licensing showed that CMU contemplated and received flat fee sums right. That's false, Your Honor. We put into evidence. That's why I wanted to do this. Yes. We put into evidence a license that was a running royalty with a 1998 license. It wasn't on this specific patent but I need to underscore that Ms. Lawton did consider all of those other licenses and she concluded in the words of the district court that they are radically different. There is a huge difference between on the one hand a license that you that you provide to what are essentially partners in a joint venture. Let's go together and figure out how to solve these problems. It'll cost you a quarter of a million a year to be a member. But you're also giving hundreds of millions of dollars worth of research support and participating in the exercise of identifying and solving problems versus Morvail. And by the way, none of them actually know what you're going to invent. Of course, is that by the way true of that last point in your last sentence? By the way, point is that true of all of the licenses that... The more bell points for the... Yes, all of the DSF3 and then there's the... Pre-invention. Yes. Then there's the Intel offer. Now the Intel offer was a completely different scenario. Intel was not in the read channel business. So Morvail went to Intel and said, we know we have this license that is worthless to you, but it would mean a great deal to us if you would license it so we could announce to the world. That was post-invention

. And then they go on to say the only evidence of any contemporaneous actual licensing was flat fee. And I wrote a note to say I think that's right but I wanted to give you a chance to discuss it. So what I think they say is that the only evidence with respect to these specific states, the only evidence of any contemporaneous actual licensing showed that CMU contemplated and received flat fee sums right. That's false, Your Honor. We put into evidence. That's why I wanted to do this. Yes. We put into evidence a license that was a running royalty with a 1998 license. It wasn't on this specific patent but I need to underscore that Ms. Lawton did consider all of those other licenses and she concluded in the words of the district court that they are radically different. There is a huge difference between on the one hand a license that you that you provide to what are essentially partners in a joint venture. Let's go together and figure out how to solve these problems. It'll cost you a quarter of a million a year to be a member. But you're also giving hundreds of millions of dollars worth of research support and participating in the exercise of identifying and solving problems versus Morvail. And by the way, none of them actually know what you're going to invent. Of course, is that by the way true of that last point in your last sentence? By the way, point is that true of all of the licenses that... The more bell points for the... Yes, all of the DSF3 and then there's the... Pre-invention. Yes. Then there's the Intel offer. Now the Intel offer was a completely different scenario. Intel was not in the read channel business. So Morvail went to Intel and said, we know we have this license that is worthless to you, but it would mean a great deal to us if you would license it so we could announce to the world. That was post-invention. That is post-invention. Yes. Now if it's okay, I did promise to turn to the merits. And in particular, let me look at the merits through the lens of the willfulness because... I want to ask you just one question about a topic that Ms. Sullivan, for not bring up the latches, put aside for the minute our pending on-bunch proceeding. What do you do about the principal case? Sort of a revoc. Yes. Thank you for saying it. I'm sure I missed for that. Which I think says in terms and sites and quotes, if you earlier, decisions that say it is true that the defendant's bad conduct, inequitable conduct, can count against latches, but only if it is in some way responsible for the patent holder's delay. And we don't have that. Well, so first of all, I don't leave sort of a sort of excited other cases for that proposition, but it is an inventorship case and this court's law in the infringement context is completely different. The copying law. The copying by itself can be a... Yes. In the infringement context. And that's... I mean, that's unbonged, that's Alchermann, which says it explicitly, conscious copying may be a factored way against the defendant. Gaster chair, an explicit holding, bot, an explicit holding. And the point is that inventorship actions are just different from infringement actions. Right. They're different. Well, because in the inventorship context, if the inventorship context, the inventor who is the defendant will always have unclean hands under that scenario. They will always know that they're cutting someone out of the action

. That is post-invention. Yes. Now if it's okay, I did promise to turn to the merits. And in particular, let me look at the merits through the lens of the willfulness because... I want to ask you just one question about a topic that Ms. Sullivan, for not bring up the latches, put aside for the minute our pending on-bunch proceeding. What do you do about the principal case? Sort of a revoc. Yes. Thank you for saying it. I'm sure I missed for that. Which I think says in terms and sites and quotes, if you earlier, decisions that say it is true that the defendant's bad conduct, inequitable conduct, can count against latches, but only if it is in some way responsible for the patent holder's delay. And we don't have that. Well, so first of all, I don't leave sort of a sort of excited other cases for that proposition, but it is an inventorship case and this court's law in the infringement context is completely different. The copying law. The copying by itself can be a... Yes. In the infringement context. And that's... I mean, that's unbonged, that's Alchermann, which says it explicitly, conscious copying may be a factored way against the defendant. Gaster chair, an explicit holding, bot, an explicit holding. And the point is that inventorship actions are just different from infringement actions. Right. They're different. Well, because in the inventorship context, if the inventorship context, the inventor who is the defendant will always have unclean hands under that scenario. They will always know that they're cutting someone out of the action. And so, a repick says, well, that means that you would never have a... you would never be able to survive a latches defense. Here we've got the opposite. And the infringement context, you can be a blatant infringer, because you know about the patents, for example, as with the case here, or you could be someone who accidentally trips across the line. And, by the way, under Marvellous Theory, no one would ever be able to present and win a latches claim because in the patent infringement context, the infringement, the copying, however, is just as never the cause of the delay. But I keep promising to turn to the merits. And let's look at the court has questions about more questions about latches. So, the bottom line here is that these defenses were not just wrong, but as the jury and the district court both found they were unreasonably wrong. Marvellous Top Infringement Argument was all about drawing boxes on ship diagrams and saying, this is what you should focus on. Don't focus on that. Even when Marvellous Own Documents described the ship otherwise. And this, as in answer to one of the questions that was asked, was exactly the argument that the district court raised and rejected in claim construction. The argument that Marvell is making. So, the argument that it has to be done in a trailer? Yes. Yes. Exactly what the district court said. Don't give me the seven word email. But the response to that, I think I'm remembering correctly, is nevertheless the claim construction adopted requires that this calculation be done for the branch. And if it's being done for a branch, you're in the trellis. I'll metaphorically in the address. So, okay, but then that leads to the argument. Therefore, it's got to be in, therefore a post-processor cannot work. Now, now, McLaughlin was very clear. Our technical expert was very clear. It's going on in the trellis in the MMP. Marvell kept pointing to, and if you look at page. At page 15 of the opening brief, you've got this great diagram. And McLaughlin, this is of the MMP. It's what Marvell calls the MMP

. And so, a repick says, well, that means that you would never have a... you would never be able to survive a latches defense. Here we've got the opposite. And the infringement context, you can be a blatant infringer, because you know about the patents, for example, as with the case here, or you could be someone who accidentally trips across the line. And, by the way, under Marvellous Theory, no one would ever be able to present and win a latches claim because in the patent infringement context, the infringement, the copying, however, is just as never the cause of the delay. But I keep promising to turn to the merits. And let's look at the court has questions about more questions about latches. So, the bottom line here is that these defenses were not just wrong, but as the jury and the district court both found they were unreasonably wrong. Marvellous Top Infringement Argument was all about drawing boxes on ship diagrams and saying, this is what you should focus on. Don't focus on that. Even when Marvellous Own Documents described the ship otherwise. And this, as in answer to one of the questions that was asked, was exactly the argument that the district court raised and rejected in claim construction. The argument that Marvell is making. So, the argument that it has to be done in a trailer? Yes. Yes. Exactly what the district court said. Don't give me the seven word email. But the response to that, I think I'm remembering correctly, is nevertheless the claim construction adopted requires that this calculation be done for the branch. And if it's being done for a branch, you're in the trellis. I'll metaphorically in the address. So, okay, but then that leads to the argument. Therefore, it's got to be in, therefore a post-processor cannot work. Now, now, McLaughlin was very clear. Our technical expert was very clear. It's going on in the trellis in the MMP. Marvell kept pointing to, and if you look at page. At page 15 of the opening brief, you've got this great diagram. And McLaughlin, this is of the MMP. It's what Marvell calls the MMP. It's what Marvell calls the detector. Marvell kept saying, look, it says the turbid detector. It's going on in there, but it's not going on in the MMP. And McLaughlin kept saying, this is 46.588. No. I'm talking about what's happening in the MMP. There is a second turbid detector with a second trellis. It's called a pruned trellis. And it's all actually going on in the MMP. Not that it matters because this whole thing Marvell calls the detector. And then as to validity, Marvell's position was simply to, depending upon, rewriting a formula that our expert said is ridiculously misleading, rewriting it in a way that contradicted the plain language of the very column that Mr. Sullivan just to check my understanding. The ridiculous aspect, I think you would say, of the rewriting was not using a variance instead of the standard deviation, but putting the index, putting the index, the subscript indexing the subscript to suggest that it was differing from time to time. Absolutely. The subscript is really small and really hard to see, but it didn't put one over on the jury or the district court. It is incredibly important. In column 10, it begins by saying on line 26, that sigma squared, it is a constant. It's 0.055. It could be something else, but it's a constant that he describes. And he goes further down in column 10, and he refers to it as a constant. And it is not different from branch to branch. And I have to emphasize. What are you doing in the column 10? Column 10 at the bottom of the, for the bottom carrier graph, line 55 by a fraction, which depends on the transition noise standard deviation. Yes, correct. It's the, so line 26, it's talking about the standard deviation, that being a fraction, 0.055. He's talking down here about each branch of the metric, so it's the same branches. They're all being multiplied by the same number. You can choose a number, a number, that is different from 0

. It's what Marvell calls the detector. Marvell kept saying, look, it says the turbid detector. It's going on in there, but it's not going on in the MMP. And McLaughlin kept saying, this is 46.588. No. I'm talking about what's happening in the MMP. There is a second turbid detector with a second trellis. It's called a pruned trellis. And it's all actually going on in the MMP. Not that it matters because this whole thing Marvell calls the detector. And then as to validity, Marvell's position was simply to, depending upon, rewriting a formula that our expert said is ridiculously misleading, rewriting it in a way that contradicted the plain language of the very column that Mr. Sullivan just to check my understanding. The ridiculous aspect, I think you would say, of the rewriting was not using a variance instead of the standard deviation, but putting the index, putting the index, the subscript indexing the subscript to suggest that it was differing from time to time. Absolutely. The subscript is really small and really hard to see, but it didn't put one over on the jury or the district court. It is incredibly important. In column 10, it begins by saying on line 26, that sigma squared, it is a constant. It's 0.055. It could be something else, but it's a constant that he describes. And he goes further down in column 10, and he refers to it as a constant. And it is not different from branch to branch. And I have to emphasize. What are you doing in the column 10? Column 10 at the bottom of the, for the bottom carrier graph, line 55 by a fraction, which depends on the transition noise standard deviation. Yes, correct. It's the, so line 26, it's talking about the standard deviation, that being a fraction, 0.055. He's talking down here about each branch of the metric, so it's the same branches. They're all being multiplied by the same number. You can choose a number, a number, that is different from 0.055, but it's the same number for every branch. And I have to underscore that, Miss Sullivan just changed the theory of, which lines were you agreeing with? So the first line was column 10, line 26, and then when you go further... This is in the unfurther modified example, right? Yes, so that's the standard deviation. Standard deviation is a fraction, and then when you go down to the further modified, he's talking about the standard deviation. It's still a fraction, and it's further modified by multiplying by a scalar. It's a fraction. It could be sigma squared. I mean, it could be sigma, it could be sigma squared for all V-Tayer, but it's the same fraction every time it's multiplied. And I do want to underscore that the theory of infringement changed today. Marvel admitted in the district court, so there are two equations. There's equation 20, and there's the equation 20 that's multiplied by a constant. What can call them to... I mean, they are two different equations, arguably. Marvel and its expert admitted in the district court, equation 20 is not signal dependent. The whole argument was, well, if equation 20 is not signal dependent, you can't make it signal dependent by multiplying it by a fraction. And Miss Sullivan today said... And so, Marvel's entire theory was no-no-no. It is the further modified equation. That's the one that represents the set, precisely because of this B-com-NT. That's the set. Miss Sullivan today just said, no, that equation 20 is further modified is not the set. The set is equation 20. That's one element of the set. The second element of the set is the further modified. And which one

.055, but it's the same number for every branch. And I have to underscore that, Miss Sullivan just changed the theory of, which lines were you agreeing with? So the first line was column 10, line 26, and then when you go further... This is in the unfurther modified example, right? Yes, so that's the standard deviation. Standard deviation is a fraction, and then when you go down to the further modified, he's talking about the standard deviation. It's still a fraction, and it's further modified by multiplying by a scalar. It's a fraction. It could be sigma squared. I mean, it could be sigma, it could be sigma squared for all V-Tayer, but it's the same fraction every time it's multiplied. And I do want to underscore that the theory of infringement changed today. Marvel admitted in the district court, so there are two equations. There's equation 20, and there's the equation 20 that's multiplied by a constant. What can call them to... I mean, they are two different equations, arguably. Marvel and its expert admitted in the district court, equation 20 is not signal dependent. The whole argument was, well, if equation 20 is not signal dependent, you can't make it signal dependent by multiplying it by a fraction. And Miss Sullivan today said... And so, Marvel's entire theory was no-no-no. It is the further modified equation. That's the one that represents the set, precisely because of this B-com-NT. That's the set. Miss Sullivan today just said, no, that equation 20 is further modified is not the set. The set is equation 20. That's one element of the set. The second element of the set is the further modified. And which one... And which one... And choose depends on the particular value, whether it's a 1 or a 0, whether it's a transition or not. Which one you choose depends upon whether... The selection is signal dependent, though what you select in half the case. So, so, no, your honor, because a transition is not the same as a sequence of symbols. So, you can tell that there has been a transition without... But is that a different point? I thought you disagreed. Even assuming that a particular transition is a signal and a function that you apply when you have that signal is signal dependent. Even if you assume that, that when you apply exactly the same thing to all of the one, all of the ones, all of the transitions, and don't to the zeros, that the choice based on signal does not make what you choose into something signal dependent. So, at the end of the day, you still have just one signal dependent function. I would say yes to that, but I am making it quite... I think you are very clear, but I haven't over... But it's also an additional argument, which is being able to detect a transition, which is what in column 1 is described as the prior art, which is no good, that is not the same as being able to detect a sequence of signals. Because, for the conversation that Judge Chen was having with Ms. Sullivan, being able to detect whether it's a transition only tells you whether it's... It doesn't tell you whether it's north to north or south to south. Whether it's north to north or south to south depends upon the sequence of symbols and the district courts unappealed claim construction was that the signal dependent branch metric function has to be

... And which one... And choose depends on the particular value, whether it's a 1 or a 0, whether it's a transition or not. Which one you choose depends upon whether... The selection is signal dependent, though what you select in half the case. So, so, no, your honor, because a transition is not the same as a sequence of symbols. So, you can tell that there has been a transition without... But is that a different point? I thought you disagreed. Even assuming that a particular transition is a signal and a function that you apply when you have that signal is signal dependent. Even if you assume that, that when you apply exactly the same thing to all of the one, all of the ones, all of the transitions, and don't to the zeros, that the choice based on signal does not make what you choose into something signal dependent. So, at the end of the day, you still have just one signal dependent function. I would say yes to that, but I am making it quite... I think you are very clear, but I haven't over... But it's also an additional argument, which is being able to detect a transition, which is what in column 1 is described as the prior art, which is no good, that is not the same as being able to detect a sequence of signals. Because, for the conversation that Judge Chen was having with Ms. Sullivan, being able to detect whether it's a transition only tells you whether it's... It doesn't tell you whether it's north to north or south to south. Whether it's north to north or south to south depends upon the sequence of symbols and the district courts unappealed claim construction was that the signal dependent branch metric function has to be... able to determine the sequence of signals, symbols, I should say. Thank you, Your Honours. You've been very patient, and we respectfully request that this court affirm the district court and all respect. I'm going to give you six minutes and so on. Thank you, Your Honours. Thank you. Just Toronto, you asked my friend if you corrected the erroneous instruction and sent the case back for remand, could a robust evidentiary record be joined on the meaning of sales cycle and the location that sales and the answer is absolutely yes. It could be, and it has been, in post-Marvel cases, like the France Telecom case just finished in the northern district of California. There was no evidence to support my friend's assertions about sales actually taking place in the United States or a sales cycle, necessarily being located in the United States. There was no evidence to suggest that Marvel didn't have not infringed no evidence out of the country. I took it from the district court's re-orcel pages, around 248 to 250 of the joint appendix, that a no evidence was really on your side. There were plenty of testimony that said all the negotiations took place. These were designed wins. The point that Mr. Rosencrantz pointed us to a little while ago that said once you have one of these, all of the chips are bought from the same source, and what was no evidence was anything to the contrary from your side. The instruction kept the jury from deciding where the sales were located by... That's a different point. What I'm suggesting is a properly instructed jury. You can't accept the trial record or its omissions without a proper instruction that gets the jury to see the relationship between any sales. In real life, in fact, the contractions in Santa Clara are subject to a lot of other contingencies, like whether they'll be contracts signed abroad. But your own or my key point is the instruction was erroneous. I want to just go to the point about Mr. Rosencrantz that we didn't object to the but-for-language of the instruction. That's incorrect. We strenuously objected to it. We proffered an alternative instruction that said solely resulting from the domestic infringing is not what's denied. So the instruction was incorrect both as to excluding the extiterritorial chips and as to the weak standard of causation

... able to determine the sequence of signals, symbols, I should say. Thank you, Your Honours. You've been very patient, and we respectfully request that this court affirm the district court and all respect. I'm going to give you six minutes and so on. Thank you, Your Honours. Thank you. Just Toronto, you asked my friend if you corrected the erroneous instruction and sent the case back for remand, could a robust evidentiary record be joined on the meaning of sales cycle and the location that sales and the answer is absolutely yes. It could be, and it has been, in post-Marvel cases, like the France Telecom case just finished in the northern district of California. There was no evidence to support my friend's assertions about sales actually taking place in the United States or a sales cycle, necessarily being located in the United States. There was no evidence to suggest that Marvel didn't have not infringed no evidence out of the country. I took it from the district court's re-orcel pages, around 248 to 250 of the joint appendix, that a no evidence was really on your side. There were plenty of testimony that said all the negotiations took place. These were designed wins. The point that Mr. Rosencrantz pointed us to a little while ago that said once you have one of these, all of the chips are bought from the same source, and what was no evidence was anything to the contrary from your side. The instruction kept the jury from deciding where the sales were located by... That's a different point. What I'm suggesting is a properly instructed jury. You can't accept the trial record or its omissions without a proper instruction that gets the jury to see the relationship between any sales. In real life, in fact, the contractions in Santa Clara are subject to a lot of other contingencies, like whether they'll be contracts signed abroad. But your own or my key point is the instruction was erroneous. I want to just go to the point about Mr. Rosencrantz that we didn't object to the but-for-language of the instruction. That's incorrect. We strenuously objected to it. We proffered an alternative instruction that said solely resulting from the domestic infringing is not what's denied. So the instruction was incorrect both as to excluding the extiterritorial chips and as to the weak standard of causation. And if you send it back, we could join. In this case, it was Marvell's birth, of course, to prove the damages. We were not given an opportunity to debate where the sales took place because they were all folded in. So I would argue that we were effectively precluded. Your Honor, Judge Chen, you asked my friend, well, wouldn't AT&T versus Microsoft? Me and that if we've redone as a reasonable royal case, all the foreign windows sales would come in. And the answer is absolutely yes. He tried to say, oh, this is a causal nexus case. And there's no causal nexus in Microsoft. That's not true. The master document, the golden source in Microsoft was, of course, causally related to all the foreign sales. He proved our point that this would have sweeping consequences if reasonable royalties could in-run territoriality by bringing in all foreign sales. Judge Wallach, you asked my friend, where's the evidence that the licenses for running royalties were comparable? There is none. Ms. Lawton, their experts did not say that the running royalty licenses were comparable. And in fact, our damages expert, Mr. Hoffman, said all were agreed they were not. The comparable licenses were all flat feet. And we're not saying the jury was bound to find a flat feet. But certainly where they've dealt it is 600 to 6000 times. The jury verdict in this case, surely the historical licenses are probative of the unreasonableness of the jury award. And finally, sorry, two more just quickly, your honor. Oh, you gave me six minutes also now. The latches point, your honor. Judge Toronto, we do think latches under the court's current law warn vacator here. This is a rare case of which her latches ruling were vacated under current law. Give me just on that. Although not in the latches ruling, I think later in time, the district court said, the only reason I denied latches was willfulness. So in your view, latches falls with willfulness. Absolutely correct, your honor. If you look at her opinion, she says the only reason for for for for for a remand that is the district court might well if the latches doctrine survived, still be able on a remand to say put aside now willfulness. I still think that latches doesn't far the presuit damages

. And if you send it back, we could join. In this case, it was Marvell's birth, of course, to prove the damages. We were not given an opportunity to debate where the sales took place because they were all folded in. So I would argue that we were effectively precluded. Your Honor, Judge Chen, you asked my friend, well, wouldn't AT&T versus Microsoft? Me and that if we've redone as a reasonable royal case, all the foreign windows sales would come in. And the answer is absolutely yes. He tried to say, oh, this is a causal nexus case. And there's no causal nexus in Microsoft. That's not true. The master document, the golden source in Microsoft was, of course, causally related to all the foreign sales. He proved our point that this would have sweeping consequences if reasonable royalties could in-run territoriality by bringing in all foreign sales. Judge Wallach, you asked my friend, where's the evidence that the licenses for running royalties were comparable? There is none. Ms. Lawton, their experts did not say that the running royalty licenses were comparable. And in fact, our damages expert, Mr. Hoffman, said all were agreed they were not. The comparable licenses were all flat feet. And we're not saying the jury was bound to find a flat feet. But certainly where they've dealt it is 600 to 6000 times. The jury verdict in this case, surely the historical licenses are probative of the unreasonableness of the jury award. And finally, sorry, two more just quickly, your honor. Oh, you gave me six minutes also now. The latches point, your honor. Judge Toronto, we do think latches under the court's current law warn vacator here. This is a rare case of which her latches ruling were vacated under current law. Give me just on that. Although not in the latches ruling, I think later in time, the district court said, the only reason I denied latches was willfulness. So in your view, latches falls with willfulness. Absolutely correct, your honor. If you look at her opinion, she says the only reason for for for for for a remand that is the district court might well if the latches doctrine survived, still be able on a remand to say put aside now willfulness. I still think that latches doesn't far the presuit damages. Well, your honor, you heard, I don't see you getting to a judgment of no presuit damages just because the district court said my reason was dependent on willfulness. If you took that out, the district court could come to the same conclusion on the different basis possibly. That is possible, your honor, but we would suggest that a minimum you should remand because she found all elements of latches. The district court found that there was no causation under seroderovic and we didn't cause their delay. And therefore the only ground was egregious conduct which equated with willfulness. If you throw out willfulness, you should throw out latches and if anything remand to see if there's some other equitable basis that can override the elements, we don't see that there can be. Conscious, everything about latches with co-extences with willfulness, it was all that actual knowledge. So that you would stand or fall together? They do. We believe if we went on willfulness latches should go out. We believe if you could go on willfulness. Then latches should still go out because she found all the elements and her. Yes, your honor, let me explain. We believe we have an independent basis to throw out latches even if you don't roll with us on willfulness because egregious conduct must cause the delay. And she found as they find it that are supposed to not cause any delay. Both statements and offerments and gesture chair that your friend quoted that seems to suggest that there isn't necessarily a causation element between the unclaimed hands and latches. Your honor, we think if that's endowed the proper remedy is remand here unless her see if she can come up with some sort of unclear hands that isn't dependent on willfulness. Last point, your honor, sigma is not constant. Column 10 uses the word greater. Mr. Rosenbrandt is simply in the threat. Sigma is not constant. No, no, no, uses the word greater. Column 10 line 52. The fraction for the standard deviation for the transition noise is greater for the one branch that is zero branching in. But it's the same one you're using everywhere. It's not varying from time to time. Correct, right. But not that there's two functions, not one, orality is a set. With respect to your honor, you should reverse an amendment. Remand for a new trial in which you'll guide the jury not to create sleeping extitorial reach and in which the 50th set ship, which is detailed in our brief, is redone for proper apportionment under this court's presence. Thank you, Your Honor

. Thank you, smellful. Thank you