GCABBH Meeting — Closing Remarks

0:00 I was just asked to make some remarks, as this is the last meeting of the Binary Black Old Man Challenge Alliance, and so it seemed appropriate to have some remarks on this occasion. I was quite struck by Scott's description of his experience with Skibby's, that before the current version, everybody was a master, and the system got confused and died. That rather reminds me of the external world's view, or the view that some of the external world has of this collaboration. so that's a sick joke but unfortunately there's some truth to it and I do think that one of the important things to be done over the next year is to make the external world aware of the great amount of progress that actually has been achieved so that although the Grand Challenge Alliance didn't achieve its most highly publicized goal of trying to get, achieving a working binary merger code and beginning to build a catalog of merger waveforms. Still, a very enormous amount of progress has been made for that goal, and it's very plausible that that goal will be achieved in the next few years, thanks to the large amount of effort and successes that there have been within this alliance. It's interesting to look back five years and compare where this community was then to where you are now. Now, I can remind you that I'm not really a card-carrying member of the community. I don't really speak your language, though I tried to on occasion. But anyway, from my point of view, when I look back five years, there was nobody doing any 3 plus 1 numerical relativity. It was all 2 plus 1. Now there has been extensive experience in 3 plus 1 It's understood that there are a variety of different problems in 3 plus 1 than there were in 2 plus 1 Of course, it's one common problem, stability But the instabilities are of a different character than they were in 2 plus 1

2:30 Five years ago, there were no ideas for how to build viable interboundary conditions Now, there are several different very promising variants of apparent horizon boundary conditions as a result of the work by this alliance, several very promising implementations. There were no outer boundary modules. There was perturbative extraction of waveforms to some somewhat credible degree of accuracy. I've seen five years ago in 2 plus 1, no injection of the information from those waveforms back in to provide outer boundaries on the evolutionary codes. Now both the perturbative and the characteristic approaches have basically achieved what's needed to have outer boundary modules demonstrated in. 2 plus 1 codes five years ago were still plagued to a fair degree by instabilities, and now the 3 plus 1 codes are plagued to a fair degree by instabilities that we have seen here at this meeting, that there is real promise that those instabilities may be coming under control. And I feel never so much more upbeat about that than I think we ever have been before. There was no computational science infrastructure for this field that was common to different research groups and usable by new people coming into the field five years ago. Now we have DAG. Even the Caltech fluid mechanicians are using DAG, I'm told. So DAG has spread from this project to much of the rest of the world and is being used by people, and many people who never heard of this project The Cactus code infrastructure is in no small measure an outgrowth of the efforts in this project, although it was developed at Potsdam after the group moved there. There was no understanding five years ago in this field of the importance of hyperbolicity to numerical solutions of Einstein's equations, and there were no first-order symmetric hyperbolic formulations of the Einstein equations.

5:00 Now there are a number of Fosch formulations. There are Fosch evolutionary codes, and the ideas for these evolutionary codes have been used, ideas that come from those codes have been used as key ingredients in the stabilization of the A.D. Young code that we heard about from the Illinois group this morning. Five years ago, there was little communication between the various research groups working in this field. Now there's a lot of communication. Five years ago, there was little commonality of language and little agreement on what were the most serious stumbling blocks or building a 3D code for the binary black hole problem. Now I think most everybody would agree on one of the most serious stumbling blocks. You may not agree on what's the best way to approach beating them down, but there is a commonality of language, there's a communication, and there is a real sense of community. It was certainly not there five years ago. This in large measure is due to the young people, the young generation, and also to the fact that the young generation now gets shifted from one research group to another, and help provide cross-fertilization and communication. So I really do see an enormous change, and it really does appear to me that you are well on your way toward achieving the goal of having a working code that can generate the kinds of waveforms that LIGO needs. Of course, in the meantime, LIGO has been moving along. Construction of the facilities is finished in Hanford, Washington. It's nearly finished in Livingston, Louisiana. Detectors are being mounded in the facilities as we speak in Hanford. They will begin to operate at poor sensitivity within a year or so. And by three years from now, the first serious scientific approach will begin for gravitational waves. And the waves from the merger of two black holes in a binary system are quite, I personally think that's most likely to be the first source that is seen. It could be seen as early as 2002, just a bit over three years from now.

7:30 However, without guidance from numerical relativity, the best estimates, and these are due to Scott Hughes' name and Flanagan, the best estimates are that we will lose a factor of 2 in signal-to-noise ratio, which means we'll lose about a factor of 10 in event rate, binary black hole mergers, if we don't have significant guidance from numerical relativity. And that's really going to hurt. So I'd like to wind up with offering a wager, or re-offering a wager, which Richard and I had discussed some time ago, but I offered to all comers in this alliance a wager along the following lines, and you want to listen carefully to this, that LIGO will collect data that contain binary black hole merger waves at a strength that's strong enough that they could be unequivocally detected with matched builders based on numerical relativity simulations so it so the data will be collected the binary black hole waveforms will be sitting there in the noisy data and if we had the merger waveforms that this community is gearing up to give to us we could extract them and have an unequivocally case so they're sitting there in the data and that they will be there in the data this community gives us a working code that's capable of producing the corresponding waveforms to be used in the data analysis. Now, what odds are you? I'm giving even odds. I'm giving even odds. How many points are you? Pardon? How many points? First is free. These things have to be haggled out. The thing may not be resolved until either the detectors are improved further or you get us the waveforms. Because in order to find those waves, it may be necessary to have the waveforms. So if you don't have the waveforms for us, if you come along with

10:00 the waveforms and then people go back into the data from the year before you were able to produce the waveforms, and they find the signals. You lost. Of course, in some sense, you're betting against the combination of nature and the experimenters. But that does give you at least three years. And whether it gives you more than three years or not depends, I think, much more on nature than on the experimenters. So we give you waveforms. You don't find any waveforms in the data. code just a little more carefully also I'm only asking that you have a code that's capable of producing the waveforms to find the signals we met we may need until further upgrade of the detectors we may need a catalog of waveforms and it may be that the computational powers that they're yet able to give it. And so it may take even more time to decide who is what and who is not. But anyway, I think the basic idea is there. How many input parameters can we have? How many input parameters? Well, all I ask is that you have a code that is capable of doing this for the generic case. Yeah, but it's the thing that with enough parameters, I can probably generate any waveform you want. Well, you also have to, you also have to have, well, yeah. I guess we haven't really discussed initial data, but I think this should include having good enough initial data that basically you're able to generate initial data that has the spins of the two black holes and their masses. So this is the input parameters, and you've been able to, the initial data are posed early enough that all of the noise in the initial data has been able to clear a way soon enough to get out of the waveforms. So I think this should include having viable initial data. So I offer this as a challenge to those of you who want to sign up. I'm happy to have the other side have signatures of 40 people if you wish. Maybe I'll put out a pad of paper here and people can sign up if they wish. On the other side of the bed.

12:30 But then having signed up... I'm going to sign up on your side. Do you want Richard on my side of the bed? I'm quite happy to have you on my side of the bed. Okay, so it's Richard and me, and you. We're all going to be on their side. this gives you three years until ligo can begin collecting data and and then the odds are pretty good that nature will not provide signals strong enough to be seen with ligo one in which case you have something more like five years so it takes until ligo two five or six years so you I think that's a reasonable challenge. I'm not asking you to produce the catalogs, just have a couple of the codes that are capable of producing the catalogs. I think both of us would be very delighted. I think we all ought to take the bet just for the incentive to come wrong. Well, anyway, here's the sign-up. She writes her name in legible form. I know, they go up in a gallery of bets at Caltech. What are we doing? What's the worst thing about? I hadn't thought about this. Rich and I will take you all out to a celebratory dinner. You can wait to see the dinner. I would like to make a comment also on the aspect of the collaborative structure that we have had in the last five years. I mean, it's very easy to criticize how things were handled, but we have to go back five years and to see that it was not an easy task. And let me just give you some evidence for that. That, I mean, some evidence that perhaps not everybody in this room would have worked.

15:00 In 1993, when was the first Penn State meeting of the lines, that was my first PI meeting. And I went back and listened to the tapes of that meeting because of the news. Don't worry, those tapes are safe. You will take Ken Starr. But it was shocking to me to see that it was going to be a very difficult task. And I think with time things, especially in the last couple of years, has definitely gone in the opposite direction. I mean, we have had a better sense of toleration. So I hope that we can focus also on the positive part and thank one of the key players on this, which is Richard. I should say thank you on behalf of the American taxpayers I will let you sign up and I think maybe Richard and I will discuss the stakes with Richard Matzner and maybe if we can agree among the three of us you'll have to accept whatever we decide before you discuss the stakes which side is Richard I have to validate the previous I have only one more comment which is again to thank the advisory committee for their help now we only have one representative so people have to accept all of them. Thanks. Well, thank you on behalf of the advisory committee. I must say I am impressed with what was accomplished given the difficulty of the problem. You've made an enormous stride. It seems to me that the success is within sight, though perhaps if it's excessive life, it's also within sight. And I want to especially single out and thank, thanks not exactly the right word, but point out the contributions that we've made by the postdocs and students, the guys who did all the work over and over again, long hours, fighting instabilities, rewriting codes, cleaning it out,

17:30 and as we can see, as we saw over the past couple of days, that's been pretty successful, and we can at least now see that maybe there's a light at the end of the tunnel and a way to beat our way towards that light. That's essentially all that I have to say. I am going to be happy. I already did sign the con side. And I'll try to come up with something that's equivalent to the Texas Lottery, where, you know, more money to put in and more money to put in. So that does that. No, we'll try to keep that friendly. Anyway, that's all I have to say. Thank you all for participating for such a long time in what's turned out to be a really difficult but clearly very important project. and for the amazing amount of work and progress that's been done on this very hard physical subject. Thanks.