part of what we are talking about is what we call it period affect so the young people we spoke with our sort of as you correctly identified the were sort of in the period as technology is expanding but before it became something every household has to have. mount dial, broadband now. we have facebook and twitter and all that kind of stuff. i'm always amazed people say i facebook somebody. it's become a verb, we facebook each other. i don't. i stay away from facebook. but i think it's different for several reasons. i think these young people coming of age now should be more prepared. i still think there is work to be done. it's not that they completely lack the skills needed, some of the guys we spoke with did because they just were not available. so they found themselves in a

workplace that needed that stuff like i didn't do any of that. so, that is part of the answer. the other part of the answer is i think building a digital infrastructure also allows you to do two things. it makes you more attractive. it makes things better for people who live in small towns but it also makes it possible that again and you can compete for those types of industries and i will give you a couple of examples that are on the one examples. the google plant that's opening mayor of west des moines i guess or has opened probably at this stage it's called cloud computing, whenever that is apparently it's lots of computers together in a cloud. [laughter] ..

the danger is that then you focus on what they call also thanh takes so you're dependent on landing a big plans. maytag is a good example of what happens when your elephant dies. it can almost kill you. there's a danger with that too. >> just a kind of piggyback off that.

i'm wondering what impact it will have as obviously we've had a major shift in the majority of our industries in the service are now. and we're having huge increase in the number of times that can be done from home. and so a lot of people are choosing to stay at home and work from home and of course the internet provides that opportunity. but what impact that might have on the number of people that choose to stay and order return after college because those opportunities are increasing. and how we can maybe capitalize on that opportunity. to absolutely. >> we talked to a bunch of people in the state about billing those opportunities. we talked to people in the generation iowa commission. these are young people who have chosen to stay after college. you see people, right? one of their pet peeves is their first two perkin on 2007 compass is a group formed by the state legislature, basically to attract more people back after

graduation. in their first report they said the digital infrastructure in iowa is shambolic, but there are so many technology bald spots that it's not even funny. phone reception or a dial-up now, it's not broadband but dial-up is an aspiration and not a reality. and again that's something that is relatively easy to do. you know, as a technology develops, it's much easier to build up that kind of digital infrastructure. you are right, having the thing simply is not only to attract people back, that crazy thing, what kind of stayers do jobs like that allowed to telecommute? why not? you know, it's not just, i want to get people away from build the goodies just to get people

back. we should all benefit from those things. i mean, that's coming you know, right hematite for everybody is a good thing. and another thing about a place like us or shearer, this town. we're standing in a beautiful new library. you do have your wonderful aquatic center, which i swam in on occasion. and there are a tremendous deity is any town like this. it shows what is possible when you have people who are cynically minded and engaged. you know, so i believe that small town should develop their infrastructure for everybody. it's a boon when people are attracted back by it, but it should be developed by anybody is a no-brainer it should happen much easier. no expletives, no

shouting. [applause] >> thank you very much. it's a great pleasure to be here and for those who don't like the numbers approach to things,

probably if you buy a car the way i recommend an "the predictioneer's game," you'll save a thousand dollars, maybe more. i've done this eight times with great success. it works. it's just a little bit of strategic thinking changes the lay of the land. so prediction, which is a subject i want to talk about tonight and i will end by making some predictions, perhaps we'll be depressing, but that's life here it prediction is not anything new, of course. people have always been trying to predict, you know, they stare up at the stars and try to figure out what they mean. nostradamus stared at a bowl of water to figure out what was going to happen in the world and then wrote some mumbo-jumbo austrians which only after the event people could look at and say he predicted that, very smart guy. cut up sheets, very unpleasant for the sheep and probably not very informative. all of those methods have one

thing in common, they don't work. so i want to talk about is a way to predict and to engineer future events that does work. and when i say it does work, i want to put a marker down. as mentioned, the central intelligence agency has evaluated my models and a couple thousand applications. and it's not just that they've concluded that the models are more accurate than their analyst, which indeed they have concluded, which doesn't make their analyst terribly happy with me. they concluded its right about 90% of the time and it is -- it hits the bull's-eye more than twice as often as their analyst. but academics have also evaluated these models. i've published hundreds of predictions in the peer-reviewed journal articles for the prediction are about events that had not happened at the time the articles were accepted or published. so no question that they were

real predictions. academics who always have an incentive to shoot down others have also reviewed that work and come to a similar conclusion about 90% accuracy. so i want you to think that in mind that there is a track record in that track record is over a very large number of cases. unlike predict when asked what is the method? i certainly have no great wisdom or insight into the world. what i have is a very powerful tool game theory. now what is game theory? game theory is just a way of thinking about how people compete with each other. if you think about the problem a physicist has, we call that for odd reasons, the hard science, the easy sciences. what is physicist have to worry about what they have to worry about how particles will interact with each other. now if it happens, if i took this been and they threw it at the wall, no molecules in the pen are going to rearrange themselves in anticipation of hitting the wall.

but if i stepped off of this podium and grabbed one of you and try to throw you against the wall, there would be a lot of molecules rearrange him. would respond strategically. that's what game theory is about. game theory is about working about the strategic responses of people, anticipating what they will do, and choosing your best moves given what you think others will do. my game theory makes a few very simple assumptions about people. and sometimes people have problems with these assumptions, so i want to be clear about what they do and don't mean. game theory assumes people are rational. i rational, all that is meant is that people try to do what they believe is in their best interest. now, let's be careful here rationality does not mean having perfect foresight, doesn't even

necessarily mean having good foresight. rationality does not mean looking at all of the various options that a person has and choosing the best one. indeed, if you choose to look at everything you could do, you're probably behaving irrationally because the cost of continuing to surge eventually is going to exceed the expected gain of the continued search. and when the cost exceeds the game, it makes no sense to keep on looking. so rationality just says people do what they believe is in their best interest. now that means that people have value and people have beliefs. values are the things that they want and people are free of course to choose their values. they are shaped by their families, by their experience, by their cultural environment that they live in. there's nothing rational or

irrational about what people want. there is something rational about what they do to achieve what they want. so, they have values and they have beliefs. they have police, for example, about how other people will react to them. what sorts of people others are. either the kinds of people that want to be tough, bully them? are the sorts of people who want to cooperate and try to find compromises? that means also they face uncertainty. they know their beliefs are just that, beliefs. they're not certain knowledge about how others will respond. and they make choices under constraints, besidesin the uncey have some pool of resources, they are limited by what they have. if you are the king of nepal, you might aspire to conquer europe. you just don't have the wherewithal. you are constrained, you know that you are constrained, you can't do it. so they have values, beliefs,

and they faced constraints. and they have to work out what is the optimal way to behave given that they are constrained and that there are other people trying to defeat them. this is, in essence, the problem of chess. is the problem of ridge, it is the problem of poker. so if you think about a game like poker, for example, is it about how good the cards are that you hold? generally not. it's better to hold good cards, of course. but you can play a very successful game of poker, holding poor cards. poker is about signaling people through betting. and foreign affairs is about signaling people. there are two kinds of signals, cheap talk and costly signals. i'm always very proud of my undergraduate students. i teach at nyu in the introduction of international

relations. i'm always proud when they come into class after the president or some other foreign-policy leader makes an important statement and say, wow that was cheap talk. for example, they remember a few months ago the north koreans for change were poorly behaved and president obama maties each that there will be dire consequences for the north korean violation of our agreements. there will be economic heightened economic sanctions against north korea. and my students understood, this is just like saying blah blah law. we don't trade with north korea. what will we sanction them with economically when we don't do business with them? on the other hand, sometimes when a president says there'll be dire consequences, as when president obama has to attend sanctions against iran for its nuclear program, which i will come back to.

that can have real effects. that's a costly signal. not just because it costs uranian something, if we cut back bank transactions and other activities that the iranians do. because they cost some of our friends money because they do business with iran. and if they abide in the sanctions that we threaten, that's costly to them, which means they will be a little annoyed. and so the president is taking a political risk. and because it is costly, we can take the statement as credible. and of course, the iranians have to work out, is the cost of giving into the sanction -- giving in to the threat, the demand for change of behavior greater than the cost of living with the economic losses of the sanctions? and if they conclude that the threat of sanctions is worse -- but the actual sanctions would be worse for them, then changing

their policy then they would change their policy. if they conclude that the sanctions will cost them nearly as much as it will cost them politically, to give in, then they won't give in. so this is why threatening sanctions can be credible and enforcing on implementing sanctions are almost never successful he cut implementation means that as artie worked out that they can bear those costs. that's game theory reasoning. it's working backwards. thinking about what the other guy going to do and then deciding what your actions should be. okay, so if one wants to construct a model to analyze foreign-policy problems, how do we start? well, for one thing, we want to think about who are all the individuals and groups that will try to shape a decision on a given policy? not just, were the key decision-makers who will make the policy?

for example, if we're thinking about what to do about iranians nuclear program, we know that the president of the united states will make a policy choice. but let's face it, president obama probably doesn't know a lot about iran, so he is going to turn to his advisors he's got. the secretary of state, he has national security adviser, he has the secretary of defense. he has a secretary of energy and so forth. he's going to turn to these people and ask, what do you think we should do about iran? what would be the effective strategy? and let's face it, stephen hsu is a very, very, very smart man. he used to be my next-door neighbor. i babysat for his children. he's got a nobel prize in physics. he knows about particle interacting. he does not know a lot about people interacting strategically and he probably doesn't know much about iran.

hillary clinton operably doesn't know that much about iran either. so they have advisors, date turn to their advisors who do know about iran, about energy policy, about nuclear development and so forth. and those people may exert a great deal of influence in what bubbles up to be the recommendations that the president receives and that eventually translates into his policy. so if we ignore those people farther down the ladder, we are going to get the analysis wrong. because they are the people shaping the views. and of course, that's just one side. the iranians have their own set of advisors. i expect that ali khamenei doesn't know a lot about uranium to richmond. he knows more than he cares, but is probably not an expert. he turns to other people to get advice. so if we want to analyze problems, we need to pay attention to a large number of people who are shaping the

choices. people who have advisors, people who are lobbyists, people aren't interested, people who can organize street demonstrations, anybody who can shape how others will think about the choice they will make. real decision-makers who don't have the benefit of a computer model sadly tend to have to take intellectual shortcuts. so they don't pay attention to all those many, many influencers. they say well, there are 50, 60, 80, 200 people, trying to shape things. just consider the following problem. imagine you just had five decision-makers, kerry, jane, sally, george, and john. harry wants to know what jane, sally, george, and john are thinking about a given issue.

and jane wants to know what harry, sally, george, and john are thinking. and probably harry is interested to know what jane thinks. and maybe even harry wants to know what jane thinks the sally inks george is saying to john. so if we go to all others, there are five factorials possible interactions of interest. that's 120. actually, only 60 of them are really important, at least according to my model. and a smart person can keep that straight in their head. if you have ten decision-makers, still a very small problem. there are 3.6 million possible interactions, a ten factorial. no one can keep that in their head. again, we don't carry about all of those. i carry. now think about writing down a model. so a model is a symbol of

equations, its mass. but the math means that the logic is transparent. you can work through exactly how one goes from input information to results. and the computer is as wonderful instrument. unlike all of us, it requires no sleep, no vacation, no coffee breaks, doesn't belong to a union that says seven hours a day, hour for lunch, no more. it will work 24 hours a day. it can crunch through all of the possible interactions. so it is certainly not smarter than a smart decision-maker. but it is better equipped to pay attention to all of the subtle, nuanced details that real people have to skip over and unfortunately sometimes, as a result, make errors. hence, i believe, the explanation for why the model gets about 90% right and intelligence analysts get a lot

less rates. because they have to take the intellectual shortcuts. okay, so if we keep track of all that information, and what are the data that we need? we're going to work out what people are going to do. we'll be surprised how little you actually need to know. you need to know what does each player say it wants, what's his position on the issue? not what in its heart of hearts that they really want. very hard to know that. what is the position it has staked out because that is a strategically chosen position. you can work backwards from that to draw some interesting conclusions. how important is the issue to them? how willing are they to drop what they are doing and pay attention to the issue when it comes out instead of something else that they have to deal with? how high of priorities? how results are they, that is how much are they committed to

finding an agreement, even if it's not the outcome they desire, compared to being committed to sticking to their position even if it means coming to no agreement, going down in a blaze of glory? so if you think about litigation, a mediator is somebody who doesn't care what the resolution of a dispute that brought in four days. what they care about is there is a resolution because they are self-interested. and their future employment is a mediator depending on their successfully resolving disputes. doesn't matter how this dispute is resolved even if the party agrees. whereas, a defendant and a plaintiff were resolved care about how it is settled. they are less flexible. people vary on those dimensions. and finally, we need one more piece of information about everybody. how much clout could they bring to bear? how persuasive could they be if

they chose to be? so what do they say they want? how focused i than the problem? how results are they? how influential could they be? four numbers for each player. if we know those four numbers, we can plug that into a game like my computer model, which by the weight you want to try it, you can play this game and you can download it. @www.prediction year's game.com. zukin fiddle with it. you can't engineer. i have in mind fantasia and mickey mouse's sorcerer's apprentice. the version that's up at the moment is going to come down because we want a web-based version you can solve on the web. anyway, you can play the game just four numbers for each party. you notice what i have not included. nothing about emotions. game theory does not deal with emotions except if a motion are

used strategically to gain advantage. a real, raw emotion has no place. i have said nothing about history. i love history, and read a lot of history. i've written some about history. but i've not mentioned history as an input to the model. i've not mentioned culture. that is the very things that most people think are central to working out what's going to happen in the world. now i believe that history and culture shape those four numbers for each of the players. but if we have the four numbers, we don't know how to know what shape it is. it's like walking in on a chess game. you come into the middle of a chess game. if blacks move. you don't know how the board got where it is, but you can work out the next move for black to make if you're reasonable chess player. you don't have to know how the game got there.

you know that the players are self-interested. i want to win and therefore you can work out what they're going to do. okay, so why shouldn't anyone believe this we have a chakra kurd and i'm now going to dare to be embarrassed, the title of the chapter of prediction year's game, dare to be embarrassed. i believe firmly that people think they have a method for engineering the future they should come out and state what they think is going to happen and if there's time, why rather than after the fact they i knew that was going to happen. i realized that. or after the fact provide an explanation. it's incredibly easy to fit facts to known outcomes. the hard challenge is to dare to be embarrassed and make the predictions before the event happened. so let me talk a little bit about the upcoming copenhagen meeting designed to try to

control greenhouse gas emissions and global warming. my analyses indicate that we make it an agreement out of copenhagen. it will be meaningless. why will it be meaningless? it is the nature of universal treaties to be meaningless. with few exceptions. so we should step back, i will expand when a moment, we should step back and ask why is that that virtually every world leader that went to the united nations and made speeches called for the essential universal agreement coming out of copenhagen? without its being universal, we couldn't do anything. i don't quite know why we can't do get a lateral action, i'll come back to that. why is copenhagen going to fail? what is the nature of the universal agreement? so it has to be an agreement that just about everybody will find, otherwise it's not universal. why will they find?

universal agreement asked people to do one of two things. either they asked them not to change their behavior, in which case they're perfectly happy to sign a comply because it has nothing to do anything. or ask them to fundamentally change their behavior and then lacks teeth. it either has no monitoring mechanism to detect cheating or if it has a monitoring mechanism has no way to enforce punishment. so as we think ahead to copenhagen, let me reflect back on the kyoto. 175 countries signed kyoto. the united states did not come about 175 dead. now, people were ecstatic about such a broad-based agreement here it really for the first time we were going to take charge of global warming. unfortunately they didn't read the fine print. so the 175 signatories to kyoto,

137 have to do nothing except report on what they have done. we are proud to report this year we have done nothing. we are at full compliance with the kyoto protocol. as we promised, we have reported. 137 out of 175 had to do nothing. the host country came forward and we deeply regret we are unable to meet the greenhouse gas emission standards that we signed. and the british did the same animosity others. and what was the consequence? none. there's no punishment in the protocol. there is no consequence for signing and then violating the agreement. and that is the nature of universal agreement. either asked people to do nothing so of course they comply, or they often to do something serious and then they shut their eyes is that they can't see whether or not they're

actually doing it. or they are a violation. now we have to ask them why the people promote universal treaties instead of taking unilateral action. and the answer is quite straightforward. if you can say, well, we can't cut our greenhouse gas emissions because after all india and china and brazil and some other rapidly developing countries are not cutting bears. it would be unfair to put our citizens, face a burden, while these other people continue to pollute, forgetting that, you know, western europe and the united states has been polluting for 200 years and china and india are just getting started. instead of stepping back and saying i have a different plan. we will unilaterally impose very high taxes on fossil fuel use.

two main fossil fuel use is that our heavy polluters. gasoline, you know that, fertilizer is a bigger greenhouse gas emitter than gasoline. so, if we put a heavy tax on fertilizer on gasoline it will be more expensive to drive big cars and will be more expensive to drive produced at the supermarket. now we can figure out why this doesn't. we can use those tax dollars to help compensate the indians, chinese, and so forth for switching to cleaner energy sources. and we could use some of the leftover money to sponsor research among cleaner energy users. it's true that you all know that foreign aid is very unpopular because what happens is the voter looks at this and says, i'm not going to vote for

somebody who is taxing me to give money to people in a faraway country that i've never been to, never going to visit. and so they say no and therefore the politicians don't make these proposals. instead they propose they view universal treaty and that we all feel good when it gets signed and feel disappointed when it doesn't get enforced and don't work out. what was the strategic reasoning behind that? it was the politicians wanted to get reelected and you are not going to reelect them if they did the right ring. and if you doubt that, i will very quickly tell you, i ask my students, how many of you would give up your cell phone use to help poor people in poor countries around the world. and the answer is none of them. although, they're very eager to help poor people in poor countries, but their eagerness is for somebody else to pay for it. and we are the same. let me talk a little bit about

iran. so back in february, i gave a talk that you can listen to online at a ted conference and i made three predictions about iran. i predicted that they will not build a nuclear bomb, but they will build -- make enough weapons grade fuel so they can build the bomb, but won't. i predicted that the theocratic regime but it died only cloning in the next couple of years be replaced by a more or less petty dictatorship, run by jeff ari, the head of the revolutionary guard with the support of the group called the finance who control much of the money. and iran and with the coombe clerics rising in influence. so basically producing a more dramatic regime. and i predicted that the

students dissidents and the coombe clerics would be rising in power as homage in the chad recently went flat or declined. that was back in february based on that analysis and quite a bit earlier. with my few remaining moments let me review what he know so far. we all waiting to see what iran is going to say this week about the international agencies proposal of transferring a large amount of the vast majority of enriched uranium to russia. here's what we know. on september 9th, front page of "the new york times" is a report that said that the president had just been informed by american intelligence that the iranians had made a big spread forward towards building a bomb and then to liberally, that the exact word used, deliberately stopped short of

making a bomb. and on october 14th of this year, "the new york times" reported that the international atomic energy agency had reported that there was no evidence that iran was attempting to build a bomb, although they clearly had enriched a lot of uranium. why wouldn't they want to build a bomb? they get all the upside in terms of satisfying nationalist urges, in terms of reasserting their role in the middle east, in terms of reasserting their prestige as a potential exporter of fundamental islam. remember they're the ones who export shia islam. they gain all of that by showing they know how, but don't actually do it. and they avoid the wrath of the rest of the international community. what else have the newspapers

reported? well, we know in june, which is by the way you can look at this online, was when the high maney power started to turn down. we know that there was a huge set of demonstrations against the regime after the elections. they are continuing the american press does not report very much on these demonstrations. i urge you to read the bbc which does. there are demonstrations almost daily and lots of cities in iran by students against the regime. hi mimi sufficiently weakened having never faced an sufficiently weakened that when the trials of the students and dissidents from june were about to start having been couched by ahmadinejad and others like people who were the dupes of a foreign agitators. they were working on behalf of a foreign agitators. no less a person of the supreme leader ali khan maney of

intervention behind the student demonstrations, undermining trying to shore up his political position. and now they're reports by ron experts that say it looks that jeff ari is becoming the dictator of iran. let me wrap up. patrickgame theory is a transpat form of logic, that with data can help to predict the future, not a pie in the sky claim. it's been done, i've been doing it, others have been doing it for 30 years. it works. not everything is predict bull, but an awful lot is predict the ball. i at least can predict markets because my models are concerned with problems that involve the opportunity for negotiation and the threat of coercion. and if you can predict, then

there is a very good prospect that you can find better solutions to problems that you can engineer, you can find better approaches that will change or whether people perceive the situation they're in, change their beliefs and produce better outcomes here at thank you. [applause] >> i'm nancy jarvis cure the world affairs council of northern california. speaking with bruce bueno de mesquita about his new book, "the predictioneer's game." it's now time for us to take questions from our audience. this is a great first question. madeleine albright describes global affairs as more like a billiard game and not like poker or chess, where the arrangement of the game is constantly changing and dynamic. how do you account for dynamic shifts were changes in one variable might have wildly different effects on players? >> wonderful question.

so my models are themselves dynamic. the data coming in, as i mentioned, the position of the players and so forth, all of those variables within the logic of the model are changing in response to the strategic laying out of the game within the model's logic. people power go up and down, people's positions moves this way or that way. people focus on the issue ships and so forth. the world is a complicated place and it's also filled with uncertainty. and the models are designed to try to capture much of that uncertainty. so there are multiple dimensions of uncertainty modeled by the program i've written. and in addition, the software though not the online version because it's amplified, but the software allows the analyst to reduce random shocks to the data. to ask the question, how robust is a given result against unforeseen earthquakes?

how big a change doesn't take in order for things to come out differently? and that is essentially the problem i think that madeline albright is thinking about, although i have to point out that those random shocks don't happen on a billiard ball and a billiards game. it's more like tilting an amusement park, tilting the machine, where things can slip around. >> you're talking about a black swan event. >> i'd like to observe that rare event in fact have happened, rarely. and we should not lose the 90% to get the rare event. but we can at least a test for where events. >> and you give us a look at the dark side, leon saber uses of your model? >> well, i'm not sure what will constitute unsavory. i guess that is a matter of taste. i have had some unsavory quest which i talked about in my book.

and in the past year i've had some more, both unsavory and bizarre. many years ago, i was approached by somebody who alleged to represent wal-mart qadhafi a very modest man by the way still a colonel after all these years, still not promoted. suggesting that they wanted me to do analysis on how to remove him in egypt and of course i declined and informed authorities in the united states. this past year i was approached by somebody who is interested in my analyzing how the warlords could take control of afghanistan. and i said, well if you can produce a letter from me from the secretary of state or the president or the head of the foreign affairs committee and the house or the senate that says this is in the interest of the united states, a letter like that. and if you can't, i won't. they couldn't. i was also asked busier whether

i would analyze what would be the right venue and timing for announcing that the earth was regularly being visited by aliens. this game from a person who is a very distinguished scientist running a distinguished research institute in the hard sciences. and so i wrote back and basically said, you know, if you could provide a letter from the president etc. and i'll analyze it, otherwise it'll think so. those would be the unsavory or bizarre. i don't do those. >> if your equation use inputs from analysts, how can they ultimately predict better than analyst in general? how about garbage in garbage out? >> wonderful question. and several answers. first, i teach an undergraduate course called solving foreign crises. my students generally don't have

access to analysts or experts, so they told the web, which their mother back than i am. they generate the information of websites. you don't have to have the analyst are the expert. why would the model be better than the analyst? well it would be for exactly what i talk about the d'amato keeps track of more nuanced information. what about garbage in garbage out? there's a wonderful check on whether the data coming in are garbage. so remember the model is dynamic, is predicting change through time, but the first-place is just a reflection of what the data for the current state of the world is. if the data are garbage, they're not going to reflect a reasonable understanding of the current state of the world. and so then we know it's garbage. if the data are pretty good, that they do reflect the way we understand the current state of the world on a particular problem, then from that point forward it's the model's logic that is taking over and is

generating change in the variables and changes in the predictions. and so they think to argue with is the logic in that the data. >> there's someone in the audience who's willing to pick on you just a little bit. and you said before we started that was all right with you. so, you say that history doesn't affect your variables, but yet when you discuss the copenhagen summit, the first thing you referred to was the historic that historically international treaties have no teeth. >> i did not say that history does not affect my variables. i said history is not built into the model. i said quite a silly that and culture probably shaped the values of the players hopped on the universe. and therefore history is a way to understand how the variables got where they are but however they got there is less important than moving forward. the kyoto example helps us to see the logic of universal

treaties and to understand why they fail. it's not that because that's kyoto copenhagen will fail. kyoto failed for the same reasons that copenhagen failed. kyoto is not shaping the failure of copenhagen. >> you will like this question because this is part of what you do. if an outcome can be predicted, cannot knowledge or awareness be used or result in change that brings about a different outcome? how can game theory be used for positive change, or can it? >> wonderful question. of course again or it wouldn't be doing this. so if you think about a policy problem, i will illustrate this with iran. if you think about a policy problem, they're basically a few things that one can do. one can take a more moderate position than people anticipate or a harder line position than people anticipate. one can take a united position or divided edition, good cop,

bad cop. one can pay less attention to the issue than expected for the level that was expected. one can show more flexibility or more resolve. you can manipulate the variables. what we can do is simulate what happens if we are more moderate or more extreme or what have you and see how the problem unfolds to see if we get a better result. and that's exactly what i do in my consulting life when i tried to engineer outcomes. so let me talk about iran. i did a study, the details of which i won't go into, on iran for the american government. and when i presented the results of that study in august of 2007, it was on nuclear policy. it was presented to a pretty

senior audience, including the person who had chief intelligence responsibility for nuclear proliferation issues. and he was very upset with my results. and argued vigorously with them. and i argued vigorously back and he pointed out the data where they are, so the issue was the logic and here was what i saw as the problem in his argument. could you tommy where the feelings were in the logic of my model. in the meeting broke up rather acrimoniously. i got an e-mail two days later saying i go to a lot of briefings, ignore them. but you get to have for the last two nights because they could not dismiss your claim that i had come to my conclusions on weak logic. correlation is most assuredly not causation.

i'm not trying to suggest any credit here. i'm just stating a sequence of events. this is the person who had chief responsibility later in 2007 before writing the new national intelligence estimate on iran that concluded there wasn't much evidence that they were trying to build a bomb since 2005. back to me is how a model should be used. he confronted the logic. he confronted the other that it led to. he argued with it, he thought about it, he came to the conclusion that maybe there was some real merit in that decision and it was different from the position he had staked out and he adjusted his views. i'm sure yet about the other inputs. i'm taking no credit here. i'm just stating that's the way to use analytics to do good in the world. to force people to think through, how does you are right that decision because i can show you exactly how this model arrived at its implications.

and sometimes he gets people to change their view. >> as a follower to your comments about working with the government, the question is, can you tell us how game theory can be useful to an intelligence agencies like are the efficient for personnel recruitment? >> so this is a really interesting question. i'm at the moment on a committee of the national academy of sciences that was requested by the intelligence community to try to think about how to bring more social science methods like game theory, network analysis and so forth into the intelligence assessment process. there are members on the committee who specialize in organization theory and they have some very interesting ideas about recruiting people to be more effective intelligence

analysts. i don't have much insight into that. on the other hand, my models have been used to, for example, identify people in senior positions in foreign governments who might be more sympathetic to u.s. concerns than we had thought and to have leverage that to help them provide us with assistance. that's i think a very important thing. >> this is a good question. where have you and your model been wrong? please give us specific examples. [laughter] >> absolutely. 10% of the time it's wrong. actually, in "the predictioneer's game" i start up with a chapter called the worst ever prediction for my model. 1993 i was engaged to work out what the health care reforms were going to look out when the clintons first came into office. when bill clinton and hillary

clinton set up task force. i forget the exact details. i predicted something like 27 specific details of what would come out. i got 100% wrong, every single issue wrong. and i learned a lot from that. that's wafers are to introduce random shocks because at least in my model went wrong was that the model indicated that the then chairman of the house ways committee with the key person, the tipping point person who could put together the right compromise to get through. and he was indicted on a think 17 felony counts and his focus on health care which is zero. of coarse humor to present. more recently, i did a book with some colleagues in 1996 called red flag over hong kong which was about the transfer of hong kong to china. and i had a lot of predictions, most of which are right, but one of the predictions, which was wrong, if you're running a hedge fund would've cost you a lot of money.

it was that china by, i think it was 2000, would decouple the hong kong dollar, would be paid a hong kong dollar from the u.s. dollar. and they came very close by the way, but they didn't do it. so if you have been a hedge fund investor, spent a lot of money on that deep again, you would've lost your shirt. i was wrong. i have not had an egregious aerie or of late. actually, in the chapter dare to be embarrassed, there's a section on pakistan in the big picture prediction so far have been right, but there are some important details that have not been right. the model predicted that the pakistani government would try to cut a deal with the taliban, who were greatly raising empower according to the analysis unless the united states increased foreign aid to $1.5 billion,

that was the exact number in the book. and to the extent that we did not tie strings to that $1.5 billion, then they would go after the taliban. but the book also predicted that there was a consequential probability that the military would potentially, not a sure thing, but would launch a coup in july. they didn't. on the other hand we just passed a $1.5 billion aid bill. they protested the strings were attached. senator kerry just announced, no, no, no we don't intend to interfere with the sovereignty of pakistan, just tell us how you spent the money so they can use it for their own purposes. and as a result, they are now going to war restfully after the taliban. >> i think the 1.5 billion approved in september proves that congress has been reading the prediction year's game. maybe you can get a translator. you touched on this in your talk

earlier, but perhaps you could explain it a little bit more. if people had used her motto, would have predicted the recent financial crisis? >> so i'm going to give two answers to that. no, yes. no to the extent that the financial crisis was the consequence of individual greed and risk-taking by bankers, mortgage brokers, and so on and so forth. yes to the extent that congressional regulatory policy, that for example during the clinton years and then the ante was upped during the bush years, clinton promoted the idea that 67% of americans should be able to own homes. bush upped the ante to 70%,

which pretty much assure that people would be buying homes who couldn't afford them. so to the extent that the regulatory environment contributed, that port would've been predicted because that was political decisions, the individual part was basically just bad market behavior. >> i object to the statement that game theory supports the idea that people compete rather than cooperate to advance their best interests. the study on the prisoner's dilemma indicate that that may not be the case. how do you react to that? >> i never said that, so i agree with the comment. game theory, in fact my game is about the conditions under which people cooperate, negotiate in the shadow of the threat of coercion. game theory is comfortable with

of cooperative behavior and non-cooperative, competitive behavior. self interest does not preclude cooperating and there's nothing in my model or in my book or in my talk, i hope, that indicated that cooperation is not part of the self interest. operation is not always in one's self-interest. so if we take the example of the prisoner's dilemma, the one shop prisoners dilemma has no comment. it has a dominant strategy that those parties defect. if we look at a finitely repeated for an a known number of times, there's also no cooperative outcome. if we look at the infinitely repeated prisoner's dilemma, if the reward payoff is worth more than twice the temptation or take advantage of the other guy or be taken advantage of, then

there is a cooperative outcome. there is by the way an infinite number of equilibria in which cooperation is one. and if that is the case that the payoff for cooperating is not double the taking advantage than equilibrium is to alternate. i punch you in the nose and i get a big payoff and then you punch me in the know than i did a big payoff. even that is a more complicated program and turn problem. if you know the stories of the prisoner's dilemma, the cooperative outcome is to terminals precluding against the interests of the people. >> regarding the israeli-palestinian divide, you say that all land for peace arrangements fail due to the time inconsistency problem. can you explain? >> yes, excellent question.

somebody has read the book. so let's suppose you and the israelis and you are the palestinians. and you say to me well, if you give us land i'll give you peace. i give you land and then you say, is that all the land? i thought you were going to give me more. it's not worth my while to give you peace for what you've given me. you had no credible commitments. there was no cost to you for taking the land and then reneging. this is pretty much the story of gaza. flip it around. how do palestinians enjoy the israelis. you say to me, disarm, give up violence, and then we will make easter to you. and so i lay down my arms and i say, okay, now let's resolve the territorial issues and you say to me, you are no threat to me. you have no arms. why would i give you any of my

land? this is again, lack of credible commitments. this is the time inconsistency problem. what was the one thing that the ira was most lax to comply with? once they're gone, you have no credible threat with which to enforce the deal. so land for peace and peace for land are both foreign formula that are doomed for failure. i talk about a mechanism that could incentivize both sides to issue violence without any need for they are trusting each other or getting along with each other. you have to redo it to find out about that. ..