Political Science 30: Politics and Strategy, Lec 2, UCLA



okay I have um yeah three announcements to make first of all sections are going to meet this week except for June haunts sections on June is out of the country until next week I believe if you're in his section you've already heard from him his sections will not meet those sections our section 1 F 1 J and 1l it's the Friday 9:00 a.m. one one of the 11 a.m. ones and the 1 p.m. one on Tuesday the tas names were not yet posted by the sections on the course web page now they are ok so if you're not sure which TA is yours you just need to look to see what section number you're in go to the course webpage and you'll see all the TA name ok so June students wait till next week everybody else students of Adam Florie and Emily go to your sections this afternoon or tomorrow ok second announcement is some updates about enrollment so actually two pieces of information there one expected one a little bit new as expected I was able to wait let in the waitlist so if you were on the waitlist as of like 9 o'clock this morning or so you're in the class you're fine what I didn't expect is that in letting in the waitlist I completely closed the class ok and that that is how it is it's actually fine letting in the waitlist has this way over the limit for both the room and the number of TAS we have it means if you are not on the waiting list your chance of getting in the class would from very low to zero ok so if you're in that situation now there's really no point in continuing to stay go look for another class and I hope to see you in some future Poli Sci 30 last announcement concerns those of you who are enrolled in the AAP program the academic advancement program I think is what aap stands for anyway it's a tutoring program I think if you're eligible for it you know the signups will have started today they'll continue through Friday of third week let me put up the contact information for the AAP tutor his name is Chris Chris jau chris has a class himself that at this time so that's why he's not coming to make the announcement but he's got a sign-up sheet in Campbell Hall I'll put his email up here and his cell phone are EGU s and you Jeannie calm six to six five three three two oh five eight okay so as I said on Tuesday I'm just going to jump right in and set up a game and solve it for you and in doing so I'm going to expect I'm not going to expect you to necessarily follow every little step' or see the logic why I'm doing everything but if I think if I go through a complete example and then go back and say okay why did we do it this way not that way that it'll just be a lot clearer rather than talking an abstract we'll have a concrete example to work from okay so that's that is what I'm going to do and in doing that over here I'm going to put the Holy Trinity of game theory okay every time you set up a game every time you take a situation in the world that you're wondering about and try to put it into the game theory form to analyze it what you are doing is you're looking for three things okay I don't need to write it again you're looking for these three ingredients what are the preferences okay preferences have of really two components when we say we're identifying preferences we're really identifying who are the players okay who are the different people with different goals who are both able to affect the outcome okay and what do they want the second thing we want to identify specifically are the actions or the strategies those words are not the same thing and in the course of today I hope you'll get to see the differences between them in the relationship that they have but right now what we're wondering about is what can each player do what choices do they have notice there's an order here first we decide who the players are then for each player we think about what do they want for each player what can they do what choices can they make the third ingredient which is often the most work in setting up a game is to figure out what the outcomes are okay what can happen in this situation not just what we think is going to happen but what are the other possible things that don't happen something that you're going to I hope see very clearly this week and next is how important things that don't occur are for our understanding of what really does occur if we want to understand why sometimes there's a war we have to think about what else could happen okay to the other side if we want to understand why there's peas we have to understand why sometimes there war okay so whatever we're seeing in the world if we want to understand why it happens we have to think about what else could happen okay so the key with outcome is all the possible things that can happen not just what really does happen the other possibilities the roads not taken the dogs that don't bark those are going to large in our understanding of strategy and what happens is tragic situations okay all right still pretty abstract here let's switch gears and start talking about a specific case with specific players preferences actions and outcomes okay so what we're going to do now is did I leave point B off my outline yeah I did well I'll add it right now I think there's space I'll add it in a different color and it's fine point B is this nice example all right point B will do an asterisk here I'll send that up skyward for a while point B is our example okay this example is about incumbents incumbent politicians challengers the people who are running against the politicians already in office and I will say here is a campaign strategy campaign strategy is pretty broad really what I have in mind here is fundraising strategy okay how many people here have um been involved in any kind of campaign a good number okay um those of you who are involved was there more or less fundraising than you expected was that a bigger or a smaller deal whore way bigger right and that's that's you normally the experience people have when they first get involved in politics is they didn't believe how much time they were going to spend asking other people for money and it's not something that most people like to do very much including politicians okay so even if you talk to very seasoned very professional politicians people who've been in the US Congress for years been through lots and lots of campaigns talk to them about what they like and don't like about their job you know they're politicians they're pretty upbeat people they'll have something positive to say about just about everything but they are hard pressed to say that they like fundraising okay it's a lot of work it's tedious it's thankless in many respects but it's a huge huge part of politics okay well there's this several puzzles about this and one particular puzzle moons especially large in the context of congressional elections so in this example I hope you see that this example is becoming more and more specific as I go that's off going to be a feature of this class in this example the incumbents we're thinking of our members of Congress it applies pretty well to members of state legislatures as well in particular state legislatures like California where on the state legislature has a lot of power and is it's a pretty good job to be a member of it okay so these incumbent legislators and the challengers are the challengers who run against a sitting incumbent now I'm really emphasizing that the story I'm telling here is one that pertains to elections in which there is an incumbent okay not to open seats okay strategies campaign strategy and fundraising strategy also in open seats is different okay but what we're going to look at here is just fundraising strategy for incumbents okay and what I'm going to do is I'm going to talk to you a little bit about puzzle and the puzzle is the fact that you guys just told me that there's so much fundraising that goes on by incumbent members of Congress they really spend a lot of time doing it ah even outside of the campaign season two or three week nights every week if you're a member of Congress you're going to some kind of fundraising dinner um because that's that's what you need to do and what I hope you're sort of wondering now is do they really need to do that in some of the other classes that you've taken if you've taken courses in American politics and Congress especially you might have heard about incumbency advantage okay the idea is that once these guys get into office we usually think that they've got an advantage against their Challenger just by being a member of Congress they get more media coverage they get much more exposure to their constituents the fact that they were elected from the district maybe says something about their general fit to the district that they are demographically and politically appropriate for the district do they really need to raise all those funds especially when they say I don't like to do it okay what's up with that well here's a case where if you think about what I was talking about at the end of class on Tuesday about economic versus psychological explanations this is a case where the economic the game theoretic explanation of why incumbents are such insatiable fundraisers is going to be different from the psychological one a psychologist might look at that situation and invoke ideas like have it you know they start fundraising early on and they just get in the groove and they can't get out link to that idea you might have run across terms like learned helplessness okay you just kind of you learn to do it one way and then you're helpless to switch and as I indicated on Tuesday I think there's something to it that might be the answer it's not the only and what I want to show you today is just the existence of more fundraising than would seem to be necessary the existence of that by itself should not be evidence for the psychological model okay so we don't have to immediately conclude that the only explanation for all the fundraising we see by incumbents who don't seem to need to do it that the only explanation for that is blinders habit the inability to see a better way I'm not saying it's not that but it doesn't have to be okay the way I'm going to make that argument now is by making some specific assumptions on these three points what are the Preferences what are the actions and what are the outcomes okay and in doing that I'm going to set up a gate something I want to point out right now is that in my little narrative right here I've already very informally said some things about who are the players well incumbent and Challenger what do they want they want to win office that's what I'm assuming I bet that's what you guys are assuming – what can they do well they can raise funds or not they could raise funds to a greater or lesser extent as well but I think we're going to think about it is yes/no here and what are the things that could happen well you know one person could win the other person good way that's it in a nutshell whenever we talk about politics we are making assumptions like that assumptions about who's in the game what do they want what they can do and how their different actions how the actions they take affect the outcome we always make assumptions I really want to be clear about that what game theory forces us to do though is to be very conscious of our assumptions we always make assumptions but we don't always make a big deal out of it okay why forcing us to make a big deal out of our assumptions game theory keeps us honest about them okay game theory might make us feel embarrassed if we assume that people that agree with me are smart and considerate and thinking about the good of society as the whole and the people who are running against me our evil and stupid that's not an assumption that we maybe necessarily want to make it's an easy kind of assumption I'm caricaturing it a bit it's an easy kind of assumption to make when we're not self conscious of it okay I'm emphasizing this point now because some of the resistance that people have to learning game theory some of the resistance to learning a style of analysis that isn't our natural way of thinking again reminding you of what I was saying on Tuesday about natural versus deductive ways of thinking you guys might find yourself saying all she does is make assumptions all the time okay if you say that here's my response then you say that to yourself self say back oh but don't we all make assumptions all the time and she's just telling us that's what I hope I'm doing and that's what I hope I teach in you to do okay not to make more assumptions than you otherwise would actually I think that the more self-conscious you become about making assumptions the less you'll do it or the more careful you'll be about doing it but again just it's setting up the story my story had a lot of assumptions in it what I'm going to try and do now is organize those assumptions and make them very clear okay so let's let's get right to those assumptions okay actually before I do assumptions what I think I'm going to do is the the game freeform which is going to it's a way of helping us organize our assumptions okay so this scenario that I've been talking about I'm going to argue fits a sequential game and everything that we're going to do for the first two or three weeks is going to have a sequence to it first one player does something then the other player sees what they do and reacts to it okay so this part of the class is going to be about strategy in the sense of anticipating how the other player is going to react to your own choice okay when there's a sequence the player that moves first should think not just about what they want to do okay do I like this choice or do I want that choice but also to think about how the other player will react to it okay to not think about the other player's action is to fail to be strategic okay when there's a sequence when there is a sequence to the real-world situation we're trying to understand the way that we depict strategy the picture that will help us organize all of our assumptions and figure out what they imply is a game tree time so what I'm going to do is I'm just going to draw one the reason why this scenario has a sequence is that incumbents move first in this story if you're incumbent you become an incumbent early November the election year okay in most cases you don't know who your Challenger two years or if you're a senator six years in the future is going to be okay you can start doing a lot of your fundraising and indeed incumbents do start doing a lot of their fundraising before your challenger is even chosen okay so incumbents make their decision about how much fundraising to do before the Challenger is even in the game okay that means that I'm going to set up a tree that looks like this okay I'm going to draw the whole thing and then I'll talk about it the incumbent has a choice to raise funds not okay this we're going to do colors on this so the green is the game tree the red is commentary on the game tree this is a decision node nodes are these things that branches come out up okay so the incumbent decides to raise funds or not then the Challenger decides whether they are going to raise funds I'm abbreviating here RL raise funds here same thing okay I'm about half way to setting up a game tree here I've 90% of a way of setting up a game tree in terms of the amount of board space I'm going to cover paper space for you guys barely short of halfway in terms of the amount of work I need to do okay but let's talk about what I've done so far what I've done is I've just represented each player with a decision node okay the decision that happens first comes higher in the tree the decision that comes second happens later the incumbent has a choice of raising funds or not and then the Challenger has that same choice the fact that the Challenger has two decision nodes means that the Challenger gets to decide her fundraising decision regardless of which choice the incumbent makes here okay so the way the game gets played out where the Challenger is either going to make her decision about fundraising or not knowing that the incumbent has raised funds or she's going to make that same decision knowing that the incumbent has not okay the Challenger won't end up at both of these places just one and where she ends up depends on the choice of the incumbent we will later see games where the choice that the first-mover makes early on is going to affect whether the second mover has a choice okay and what sometimes what kind of choices they have but in this game it does not okay so again let me emphasize the Challenger has two nodes in the game but the Challenger will only make one choice and again that's the idea that when we want to understand what happens we have to think about what doesn't happen okay so if what does happen is that the incumbent raises funds we have to think about what would happen if she didn't okay what would happen what the Challenger would do if it worked out the other way and the incumbent did not raise funds we would have to still be thinking about what would happen if the incumbent had gone the other way so so far what we've got is a game tree that has nodes that represent players and branches that represent the actions the actions that they can take I am going to throughout this class white game trees top to bottom that's just the easiest way for me to do it um how I learned to do it the book does it from left to right okay same idea either way the natural way that you would read and what is at the top of the diagram or to the left in the book is what happens first in the sequential game you'll see that it's very important in many games to know who moves first and who moves second all right so some of our assumptions as I said are actually in the game tree who are the players is in the game tree next Tuesday I'll be giving you a homework problem it's going to be a little story it's not going to be about incumbents and challengers but it's going to be a political story and you're going to have to figure out who the players are and one of the things you want to ask yourself when you've got your game tree maybe written out just to this extent not finished yeah is do I have all the players if a player can make a choice that affects the outcome they have to have at least one decision node okay so if you're some of you may find it helpful to compile a little checklist of things to ask yourself to know whether you set up a game right and one thing that should be prominent on that list is does every player have at least one decision node if the player doesn't have a decision node and they can affect the outcome you've left something out of the game conversely if somebody's got a decision node and their choice doesn't really affect the outcome that would be a problem too all right so now what I need to do all I need to do is to add what we call terminal nodes in here the other phrase that we'll use for what today are going to be numbers soon they'll be variables that I place at the bottom of the tree is there going to be payoffs okay what we want to do is we want to add something at the bottom of each path in the tree there's 1 2 3 4 possible paths that this interaction could take right the incumbent could raise funds and the Challenger could the incumbent could raise funds the Challenger could not the incumbent could not the Challenger could the incumbent could not in the Challenger could not four possible paths we could go down what we want to get to is our assumptions about the preferences of each player over each of these 4 possible outcomes ok so another checklist thing that you can remember is that when you're thinking about all the possible things that could happen the number of possibilities that you're considering ok the breadth of your analysis if you will is going to be given by the number of terminal nodes the number of different ways to get from the top of a tree to this bottom row ok so we've got four here it's four possible outcomes that could happen the funny thing about outcomes and setting up a game is that figuring out our assumptions about outcomes is the most work ok it's where we really do have to think about counterfactuals and we have to think about them in a lot of detail now again this is something that we do in ordinary life too we do it sort of intuitively but we usually don't do it completely and we're often not really explicit even with ourselves about what we're assuming about the counterfactuals what we observe you know we can observe that we can pay a lot of attention to it what we think might have happened harder to think about harder to communicate with somebody else about and game theory can be very very useful in communicating to somebody else what your ideas about a counterfactual are versus their ideas so outcomes and setting up the game thinking about the possible outcomes that would be associated with each set of choices by the players is where most of the hard thinking comes in the funny thing about it is the outcomes themselves never show up in the game okay I'm going to make some notes about it here but what we want to get to here are numbers that will indicate not the outcomes not necessarily who wins but how the players feel about it okay do they like the outcome or do they not like the outcome okay so that's where we're going to get there let's start thinking about what assumptions we might make about outcomes okay now here's a place where I'm going to tell you guys a set of assumptions that I have I think they're pretty realistic you might not think that they're all realistic whether they all apply everywhere and that's good hold those thoughts today is not a good day for us to talk about whether these assumptions are the right ones or not Tuesday will be okay so well well I'm doing this and you're thinking no I don't think that's right write it down bring it up Tuesday but right now it's Professor blonds game she's going to make her assumptions all right so the assumptions that I have I think I'm going to do them over here give me my assumptions about outcomes okay assumption number one if the incumbent raises funds she wins the incumbent raises funds she wins whether the challenger raises funds or not okay and I'm making this assumption based on what I'd said a few minutes ago based on what people who've studied congressional elections for years think that a well-funded incumbent encumbered who has enough money to mount a serious campaign has a large advantage and is probably going to beat a well-funded challenger or a poorly funded in challenger okay so getting one more color up on the board if the incumbent raises funds I'm writing this in blue because as I said a minute ago it's not really part of the game tree okay but right now it's going to help us get to what is at the bottom of the game tree which is the preferences over these different outcomes so if I'm the incumbent I raise funds you the challenger raise funds I beat you I'm the incumbent I raise funds you don't I still beat you comma wins both places to incumbent does not back to my abbreviating raise funds the Challenger wins oh excuse me not not necessarily that's I wouldn't want to defend that assumption the incumbent does not raise funds and and and and Challenger does then the Challenger wins okay third assumption in this category is um neither candidate raises funds then we're back to the incumbent winning okay so I said it's my game I get to make my assumptions but if I'm using this game to persuade you that I know something about congressional elections I should persuade you that these assumptions are reasonable to make the idea here is that if you have an incumbent that doesn't raise funds an incumbent that's not in a position to send out mailers um have people go door-to-door put on TV ads if it's going to be that kind of campaign is going to be at a disadvantage if the Challenger is doing those things okay if I'm an incumbent who has failed to raise money I can't put my TV out there and you're the Challenger you got my two years of votes to go over and spend the worst possible way and get them out there and the mailers and on the airwaves and that's gonna be real good for you okay a disadvantage that incumbents may have even though we don't really see it a lot is that incumbents have records to run on and records are an easy thing for a challenger to use against incumbent choices that if the incumbent could carefully explain to their constituents why they voted one way why they fail to show up for a vote why their office works this way and not a different way they won't be able to make that explanation if they don't have the money to get their message out if the Challenger has money to get out the message about what's wrong with the way the incumbent has behaved and the incumbent does not have the money to respond this is a case where challenger could overcome the advantages of name recognition fit to the district whatever talents the incumbent used to originally get elected if neither candidate has money though neither candidate has the funds to really mount a big campaign then we're probably back in the situation of name recognition mattering it's probably going to be a low turnout vote and the incumbency advantage will probably prevail he'll okay so let me put those outcomes again in blue part of my assumptions part of the process of setting up a game even though they would not normally appear in the game tree itself okay so this path here the one where the incumbent does not raise funds and the Challenger does this is the one where the Challenger wins over here um maybe another way to justify that assumption this is another path where the two candidates are sort of uneven ground here we have two candidates who maybe are not funded to the same degree but both have the funds to mount a campaign here again there's sort of uneven grounds neither of them having any money when the money situation is even we think that incumbency advantage is what's going to matter okay so over here we've got the incumbent winning okay so let's go through our checklist here um we've got sort of half of what we need for preferences we know who the players are check what can each player do we've got our assumptions about that check that can raise funds or not this is just kind of a restatement what are all the possible things that can happen well we've got them income it's going to enter the Challenger is going to win what is left is going to be the Preferences of the players and you might be thinking that that's kind of a big term for me to be using a big set of assumptions to make it would be except that when we're analyzing a game all we need to know is what are the preferences over the outcomes in the game okay so not in general what do these guys want you know peace on earth a new car also winning the election and no more fundraising but just what are their preferences over these four possible outcomes that makes it a little bit simpler it makes it a more realistic thing that we can actually talk about okay and we'll put those assumptions I think over on the other board no I think actually I do have board space here is this is this good board space for everybody is everybody able to to see this okay good good all right so I'm still an assumption land over here I've got my assumptions about what the outcomes are and again let me emphasize what I mean by that what I really mean is how the different combinations of choices by the players affect the thing that they both care about okay so when I say what are the outcomes really what I want to know is what outcomes result from each combination of actions by the players okay yeah that's good check on that next is going to be assumptions about the preferences of candidate ah that's the incumbent the Honorable representative I okay and what I'm going to do is I'm going to represent incumbents preferences with numbers and in the most natural way to do it a higher number means you like it better okay this is the idea behind payoffs when people use game theory to set up experiments in laboratories and this is increasingly done in political science economics psychology it's done throughout social science um laboratory methods are kind of permeating out from psychology where they initially were when people try to set up little lab versions of these strategic situations in a laboratory the payoffs are often real money okay occasionally their M&Ms or something like that but uh every experiment that I've seen done in recent years the payoff numbers are represented by money have any of you guys participated in any of the experiments out on CASL you even know what that is I'm mentoring around what Castle stands for CAS SEL so social science experimental laboratory I don't know what the CA is it's in the public policy building it's kind of a famous um laboratory set up to do game theoretic experiments and you know they do pay you they're always looking for subjects so if you find yourself in public policy you might see if there's a an experiment to sign up for and get some payoffs of that that's but that's a digression all I'm saying here is that when we analyze a game we represent preferences with numbers high number is good low number is bad so let's represent my assumptions about the incumbents payoffs with some numbers here alright so again I just have to represent assumptions about these four possible outcomes okay and here's how they're going to go okay these are the incumbents preferences and they emphasize that again whenever we talk about preferences we are always talking about the players separately okay always always always game theory has this kind of divide and conquer approach that's how we understand a complicated situation we break it down into components and one way we break it down is we think about preferences being those of the individual not of the players as a group okay so incumbents preferences over the following possible things I I'm the incumbent wins I does not raise funds that's the best I get to be in office I get to make the world a better place and give all those big speeches and no more chicken dinners for me that's going to be the biggest number here next thing incumbent wins incumbent raises funds well that's not so good like I'm telling those political scientists call me up and give me the answer surveys I really don't like raising funds but this number eight this payoff number eight that I'm assigning to it is better than the next outcome which is that the Challenger wins and the incumbent does not raise funds okay these two numbers are key in saying that I have a high being the incumbent I have a higher payoff for winning and fundraising than I do for losing okay challenger wins means incumbent loses and not fund raising says that if I have to raise funds to win I'll do it I'll bite the bullet okay the worst outcome here is gosh what could be worse Challenger wins and I raised funds oh my god hate that okay so those are my assumptions about the Preferences of the incumbent and I'm going to assume that the challengers preferences are symmetric okay but from the challengers point of view the Challenger winning without fund raising is the best the Challenger winning so it does not raise funds here's the best not-not-not challenger wins with raising funds challenger raises funds that's eight okay so it's just symmetric here I'm replacing the role of the incumbent with the Challenger because these are my assumptions about the challengers preferences best cases I win without raising funds next best as I win I have to raise funds next best case is my opponent wins okay I lose incumbent wins at least I didn't raise funds okay free and the worst case would be incumbent wins I spent all that time fund raising that's my lowest payoff okay so where do these numbers come from I made them up I made them up to reflect my assumptions about what these guys want to reflect the story that I was telling when I first developed this idea okay that these guys both want to win the election that nobody likes fundraising the assumption that I didn't really make what I was really not very clear about when I was just talking informally about the problem over here was I didn't really commit myself to what I thought about the differences between these two intermediate outcomes okay do they prefer to win the election and raise funds do they prefer one thing that they like and another thing they don't like to losing the election and not raising funds okay I think this is a reasonable order to rank these two outcomes that if people are in politics certainly in politics is a certain level they probably are people who have decided that winning elections is worth the trouble to raise funds okay but the point I want to emphasize here is this was a part of the story that it was very easy for me to just completely glide over when I was talking about an ordinary language for me to not be clear about um and this aspect of my assumptions about preferences is going to turn out to be critical to understanding what's going to happen okay so now I'm actually getting pretty close to being done setting up the games okay I know what the outcomes are they are in here in blue to indicate they're not really part of the game I setup those outcomes first because that knowing what the outcomes are is necessary for me to know how many things I have to assign a payoff number to so do this every time you get your homework when you're setting up the game name get yourself a list of the outcomes while you're doing it it's kind of natural to think about what combinations of actions they would come from but figure out what the set of outcomes are before you sit down and organize your thoughts about what the preferences are you need to do the outcomes first because you need to know what are the set of things you need to assign numbers to okay so now I've got my numbers let's put them in the game but first let's get this question the Challenger winning and the incumbent good point that's right that's not that's not in there okay that's very good that that's exactly right now the parallel one is going to be in here okay so that's right and it was a good point for you to bring that up let me just remember that I'm supposed to for the webcasting repeat your comments and time on the mic and you're not so what's your name Britney points out that Britney is thinking about where am I going to put these numbers okay these are numbers that are going to go with different outcomes there is no outcome in this tree where the Challenger wins and the incumbent raised funds okay the only place where the Challenger wins is in this branch of the tree where the incumbent didn't fundraise okay and that's actually right so if this had been kind of a hard part of the preferences for me to think about my assumptions and get clear about I could have just skipped it I'm not going to need this okay now it turns out that the parallel part of this is going to be in the tree and an aspect of this set of assumptions that I like is the fact that I'm treating the incumbent and the Challenger the same and not presuming that there's anything special about one or the other okay so it wasn't too hard for me to put this extra preference in here but you're right I did a little more work than I need to here I was there another yes you're also what's your name Veronica is also seeing another impossible case here that the Challenger winning without fundraising is not possible that's exactly right that's exactly right okay and well you know what you guys are you guys chomping at the bit to put the numbers in here so let's put them put them in here what you'll see is I've got all of the assumptions I need I've got a little bit more I've got a little bit more than what I need for the game I've got quite a bit less than what you might be thinking about is going on in the real world and that's going to be a tension that we're going to have throughout the class our games are much simpler than reality okay that's the way we understand reality reality is such a you know hubbub of so many things going on that we have to focus our attention on some parts of it and game theory again just makes us conscious of what parts we're focusing on okay so my main game has quite a narrow focus I got out of my focus a little bit over here as really as part of the process of explaining the logic behind the the assumptions but now I that I've got my numbers let's put them in and um well I'm doing that let me tap them all on to you something that is true about this class this time and this professor which is it's 8:00 till 12:00 now and I'm um my IQ starts to decay rapidly about 50 minutes into a class period okay and this is devastatingly embarrassing to me but it can be fun for you and good for you in terms of learning so I would actually snake at the 11:45 rule start being very sensitive to me putting numbers in the wrong places saying things backwards keep me on a very short leash make sure that I put the numbers in the right places here you'll have lots of rewarding chances to catch me doing something different okay alright so I've got a number I've got to pay off for each player for each possible outcome and a few other outcomes they actually aren't in the game and what I need to do is put those in the terminal note the convention what everybody does and you guys do it to please is to put the first mover's pay off first okay everybody does it this way okay so the incumbent has the first move I'm going to put the incumbents payoff numbers first the challengers second if there's three players they just go in sequential order okay it's you could do it the other way just like you could drive on the other side of the road but do it the way everybody else does you'll be glad ok so what can happen here the incumbent can raise funds the Challenger can raise funds the incumbent wins I'm switching back to green here because the payoffs are part of a game tree until you have payoffs you haven't finished setting up the game okay so what is the incumbents payoff from going down this path in the tree eight okay this is the case where the incumbent wins the little blue note not really part of the game tree but part of our assumptions tells us that the incumbent wins with that combination of actions and the incumbent had to raise funds okay what's the challengers payoff here one poor challenger did all that work didn't get to win the election terrible payoff all right what's the incumbents payoff here I'd say it's no different that's going to happen in many games right you think we're at a different node the incumbent preference might be different but being come into the care about whether the Challenger raises funds or not now you could make a game where they cared but in this game I'm incumbent that challengers out there whatever he's doing this thing as long as he doesn't beat me I don't really care what he's doing okay you might think that's unrealistic you might think if the incumbent is from another party I don't want him marshaling support for that party I don't want him gaining a good reputation and maybe beating me next time but in this particular model I'm assuming that the incumbent only cares about their own fundraising and whether they win the election now a more realistic way to think about that would be to assume that the incumbents feelings about their own fundraising and winning the election are just much much more intense maybe they care a little bit about the Challenger but it's not enough put in the game okay what's the challengers payoff here three okay think about that three is sort of being the challengers payoff for keeping their day job okay I didn't win the election but I didn't turn myself inside out with all that fundraising it's a three it's fine all right over here what's the incumbent three because here the roles are switched now the incumbent is the one that didn't really work very hard didn't win the election either challengers pay off eight very good over here and come and pay off ten oh man I'm the incumbent this is the world I want to be in challengers pay off okay so just to re-emphasize the points that Britney and Veronica made um this payoff never occurs we never in this game see a challenger winning when the income raised funds so the the worst thing that can happen in this game is just not going to happen to the incumbent that seems kind of unfair welcome to the world of politics and the best thing that could happen never happens to the Challenger okay the best possible outcome is a possibility in the game now we're not going to predict that happens at least not with these numbers but it is allowed as a possibility and the worst possible thing for the Challenger as well okay so here's another thing that people when they first started doing game theory find odd one thing that people find odd is what I talked about a few minutes ago that the outcome assumptions don't really appear in the game but that's actually where you have to do the most thinking about what really are my assumptions how do I think the world works it's the knowledge of the world the judgment the creative thinking that you bring to game theory okay are those assumptions about um about outcomes yet they don't really appear in the game at the beginning of the class I'll keep writing in some of the outcomes like this and I encourage you to do it on your homework too it's not like it's wrong but you'll very quickly find yourself and when you're reading the book you'll find the book not having the blue stuff in here okay so when people say write down a game tree what they want you to write down is what's in green their decision nodes at least one for each player in the game actions represented by branches coming out of the node branches coming out of the node controlled by the player who's making the choice there and payoffs in the terminal so that's one surprising thing the other surprising thing is with a lot of games the work really goes in setting them up once you set them up solving them is not that hard this game is not a hard game to solve I don't know if I can have time to say everything we have does say about solving the game but I can do the solution for you right now the most important thing to remember about solving games so important that I'm going to I can't erase that that stuff's so important but I'll switch colors I'll foot colors and put it over here by solving game trees put it in a nice blue box is you always always solve a game tree from the bottom to the top you set up a game from top to bottom or left to right you set up a game in the order that the things happen you solve the game backwards so you'll even hear people using that particular phrase and solving it backwards give us gives us this result Dixit and skeet in your book call the process rolling back so the image they have is like it there's a screen I'm going to solve this game just by kind of rolling up the branches figuring out what would happen if we got to the very last set of nodes in the game thinking about that using what we learn about what would happen at the last set of nodes to figure out what will happen at the next to the last set of nodes and in a more complicated game just marching back up through so just as with setting up the game there's this kind of divide and conquer approach take this complicated situation and break it down into components break it down into preferences actions and outcomes solving the game also has this just divide it up go step by step think about it one node at a time okay that's going to be a feature of especially the early part of the course none of the individual steps are that hard okay that's something that I think sometimes uh sneaks past students I'm talking away up here and every staff which we start seeing me ask you whether eight is a bigger number than three or don you're going to think wow this glad I did that math pretest for this every little step' is really easy okay the trick is and the power of game theory comes from organizing all those steps breaking down a hard problem into a medium-sized number of really easy problems okay so that's what we're going to do we are going to roll back the game we are going to solve it from the bottom up we're going to start with the things that could happen at the last stage of the game okay so say we're here don't think about how we got here or whether we got here but say we're here I'm the Challenger okay everything you need to know about what I want is captured by these two numbers what am I going to do what am I going to do if we get to this stage you're not going to raise funds if we get to this point I have a choice between three and one okay I have a choice between fighting a losing battle or keeping my life I'm not going to raise funds okay now watch how I'm doing this this is actually something there are some aspects of setting up games and solving them where everybody does it the same way like putting the first mover's playoff first there'll be other ones that come up when we get to simultaneous games here people do things differently what I'm doing is trying to do it the way the book does and it's also the way I learned it I'm highlighting what I think will happen okay so I'm making this big and a new color and I'm highlighting the optimal choice optimal for who the Challenger optimal for whoever controls the decision node okay so at any game you solve a node it's a part of the process of saw the game by picking the branch that leads to the higher payoff for the player that controls the node okay so in this case it's not I'm highlighting it I'm making a big deal of my style of highlighting because sometimes you will see people crossing out the path not taken okay that's another way to do it as long as you're consistent okay what about over here if we get to this point I'm the Challenger what am I going to do I'm going to raise funds again we're looking at the challenge and we have to look at the challengers payoffs I'm an ambitious politician do I prefer to win an election by raising plans hard to my ordinary life as an intern in some other politicians office you bet I do that's what I choose all right the way that you indicate the choice taken or the choice not taken is less important than how you take this information and transfer it to the next stage the way that you do that is by replacing the decision node with what's called its strategic equivalent okay so let me just do it first and then I'll talk about the logic of it what I'm going to do is I'm going to replace this node with the payoffs okay I'm going to replace the challengers decision with eight three and three eight okay the reason why I'm doing this is because if the incumbent is strategic what the incumbent is going to think is that if I raise funds you don't like raising funds believe me I don't but if I do it it's the equivalent of getting a payoff of eight if I don't do it it's the equivalent of getting a payoff of three okay so this is what we mean by rolling back okay it's like we've just we no longer need to look at anything below the first branch here with the strategic equivalent if I am the incumbent all I need to think about is do I want to pay off of eight or do I want to pay off of free what do I want eight okay what do I do I raise funds okay so one of you guys want to summarize what the economic story is here remember I said the psychological explanation of why an incumbent would raise funds could be they just do it out of habit okay so even though they don't like do it they just incapable of stopping what well what's the alternative that we've got here yeah for fear that if the incumbent doesn't raise funds the Challenger will win that's what's your name Chiara that was Chiara put it in a nutshell and that is exactly what the game periodic political economic story yes okay let's put it here why raise funds because of how the opponent would react yeah I didn't okay why do we raise funds because of the counterfactual okay it's not because we like raising funds okay it's clear in these assumptions that holding the election outcome constant the incumbents payoff is lower whenever they have to raise funds it's lower when they win and it's lower when they lose okay so it's not that they don't like it but that they understand how their opponent would react we never see the opponent's reaction because the incumbent doesn't want them to do it okay we never observe this happened okay the reason we never observe it is because the incumbent is strategic yes what's your what's your name Alain probably more that I can remember but I might remember a few of them Soleil is wondering what is the logic of the solving backwards and that actually sets me up for another thing I want to say here another phrase that will be helpful I think it's a phrase that I put on the syllabus for this part of the course the phrase is kind of an aphorism for what it means to be strategic and the phrase is look for word reason backward ok so the logic behind solving the game from the bottom up is the logic it's a logic that I bet most of you guys have heard from your parents at one time think ahead think about the consequences if you say that to your brother how is he going to react if you do that what is going to happen ok whenever that that's part of what parents tell their children the world round it's part of what mentors tell their mentees in all sorts of situations and that is the logic of being strategic ok but the way to think about what you should do when you're deciding whether to raise funds or not is not to just look at your choice ok if we start at the top and the incumbent said well look here's my choice I can raise funds or not I don't like fundraising forget about it that wouldn't be strategic whether what the incumbent should do is should think about the long-term consequences of what they're doing what are they what are those consequences how do they like them ok so that idea thinking about all of the consequences and making your choice based on that not just based on the superficial short-run aspects of the choice that would be a feature that I would say would characterize rational decision-making in context that we wouldn't even think of a strategic what makes it a strategic situation is where part of the consequences involve anticipating what another person is going to do ok so another way to think about Elaine's question would be this kind of fits with the logic of game theory to think about what's the alternative if I don't start at the bottom I could start at the top and the way I would get wrong by starting at the top is that I wouldn't be thinking about how my income my opponent would be reacting that wouldn't be being strategic okay this phrase um look for word reason backward was popularized in a book that came out probably the late 80s it's been out for a while now um it's called thinking strategically and um it's a good book I recommend it to you it's in the category of what I would think of as sort of like a popular management book it's written I think mainly for business people people in people thinking about professional decision-making okay it's an it's meant to be advice in that sense that's not a style of book that I'm particularly enamored with I think most of them are are pretty bad but this is a pretty good book and actually one of the authors of this book is the author one of the authors of your textbook Avinash Dixit um so if you are interested in kind of some background reading on game theory how it would apply in a lot of situations that are political in one sense it's about you know power and getting people to do things and conflict being resolved but not in the sense of winning elections voting in legislatures it's more in the sense of getting along in an office that book would be uh something that I'd recommend it's not a hard read okay it's meant to be I think it's meant for people to read on airplanes when they're doing business travel seems to be that kind of book okay so um we've take in the situation translated it into a game solve the game and done a little bit of interpreting the solution so what we've done today is what I'm going to ask you to do on the homework problem that I'm going to hand out on Tuesday I haven't said all I have to say about this example so we're going to start on Tuesday by I'll put the same picture basically back on the board and talk about some of the other things we could have done differently why do we do it this way but on but that's where we are over the weekend read um you might want to read through all of chapter three ok we're not going to we're going to spend some more time on chapter three but just to get a sense of what's going on in these these games

Maurice Vega

29 Responses

  1. You can have all the political science courses in the world and it will never solve humanity's global problems. Humanity is too politically divided to solve the multiplicity of problems. It cannot address human greed, avarice and raw lust for power to control others lives through legislation, slanted to special interest groups, ala, multi-national corporations that care little for anything that does not add to their bottom line. Humanity repeats the same mistakes over and over again, as each new successive generation mistakenly believe that they have the answer to the global problems.

  2. Thank you UCLA and professor! This is amazing resource for people who want to learn in their free time after graduating from school and when busy with work!!

  3. Thanks Professor. I am major in computer science but I just love your lectures and learned a lot from it.

  4. I dont get why people are disrespecting this lady. She is giving free knowledge on the internet and helping people learn this is crazy to me.

  5. thats perfect. penny stocks must be traded with good advices from veterans. dont think its funny, even my mom started making descent money from penny stocks trading using a professional service. i found it here : bit.ly/13jgVgw?=kxhdqs

  6. Alright, I am in grade 11 and am finding this way of thinking really evolving.

    Is it possible to become an unofficial "online student" of this class by just watching the videos and getting involved in the community?

  7. This whole class seems pretty useless to me. I don't know why there's a "game theory" model. In real life, doesn't one simply have to interpret the players, so to speak, based upon their own special circumstances? I can't think of a single profession where game theory plays an active role in decision making. It's all simple logic and deduction. You don't need to make a class about that.

  8. Also, notice the camera guy gave up on trying to keep her in frame? Watch Lec 1 and she paces NON STOP through the whole thing. The camera guy this time is like… fuck it. btw… i sit necessary to have 5 minutes recording of the syllabus/lab class? jeezeus just edit that stuff out.

Leave a Reply

Your email address will not be published. Required fields are marked *

Post comment