Just to keep books in the books section, I'll repeat a decent book I've been reading in game theory: Game Theory Evolving by Gintis. There are only a few situations where GT helps you in sports, but it is the nuts when you are working a mortgage modification.
-
Comment -
That's a terrible book to recommend to people for an introduction to game theory. It is more a reference for relatively advanced applied problems. There are tons of books called 'introductory game theory' and I'm sure they're all better than Gintis'. And if you want the rigorous treatment then you should go to Osborne and Rubinstein's "a course in game theory" (http://theory.economics.utoronto.ca/books/).Comment -
How did you apply game theory to mortgage modification?Comment -
Game theory has no applications to real world problems.Comment -
I would disagree. I am a business consultant (MBA, etc.), and that type of thinking is extremely useful. True, you don't find problems EXACTLY posed like the ones in the texts, but you can apply the principles, just like, for example, you can apply military strategy to life's non-martial problems (Sun Tzu, etc.).Comment -
If you don't see the application then you don't understand GT, it's that simple.
However, if you are trying to modify your loan with a lender, you have several options, as does he. So this is a classic bargaining game of sequential moves. You can evaluate all those options in terms of probability and payoff (cash flow per year), as can the bank. It is fairly straightforward and frankly I've used it for my loans. So far all have been negotiated in my favor as I understood the game.
The Gintis book would help someone who has the math and would like to see worked (cookbook) problems. BTW, it has a better than average tutorial on Bayesian stat.
As you point out there are several decent game theory sites on the net. If you are just a casual person, vaguely interested in GT, and don't mind searching around the odd corners of the internet, then the net is OK. This book wraps up most approaches in one source that I can dog-ear, write notes in the margin, apply sticky notes, etc.Comment -
BTW, if anyone would like to see a worked real-world mortgage modification problem, then PM me. Decision trees are impossible to diagram in this interface and I can fax the diagram to you (or something).Comment -
Sounds fun.Comment -
looks interesting
Comment -
BTW, GT has application in sports.
Take last years last 3 weeks of the season where you had the question of: given that your model says "take Indy against any of the last three opponents" should you do it given that the coach may make the decision to not fully play to win? When you run it through a decision tree with derived knowledge, the answer was no and BTW, the same thing held for the Saints. Where this did not hold up was when the Giants played New England on week 17 a few years back, GT told you that Tom C. would not play to win when in the actual game he did. What you didn't fully know was what other pressures he might have had in an unusual situation with NE going for an undefeated season (OK, so Indy was in this situation too).
This situation occurs in all sports in the final games of play before the playoffs.Comment -
I didn't see Justin7's review on the Amazon site...only three star review was "daringly"--who I thought was somebody else.Comment -
First, I don't have any formal training in GT but I think the way that I select horse wagers may in fact be, at least partially, an application of game theory. I will try to be brief without losing the point.
Having lived in NYC, my earliest gambling was on horse races. I read a lot of books on handicapping and set out to see if I could beat this game. I would handicap a race to determine the most likely "winner" and made my bet. I would win some, lose more but when all was said and done, I would have a net loss after a series of wagers. In continuing my reading I came across the concept of an overlay. Made perfect sense to me, wager on a horse that had a better chance of winning than its odds indicated. The only problem was, how was I to know if the horse I selected was in fact an overlay at the odds offered?
I was already selecting "winners" but they obviously were not overlays because I was not showing a profit. In trying to work this problem out something clicked. If there was an overlay in the race worthy of wagering on then there also must be an underlay in the same race. My plan was now to look for races where I could compare the horses and the odds to try to determine if one horse (or more) was being overbet while another horse was being underbet.
Now finding overbet 50-1 shots doesn't do much good as even if they have no chance to win, they can only increase another's horse chance by less than 2% which is much less than the take at the track. I need to find overbet short price horses. In trying to figure this out, I noticed that there were many races where two horses would get most of the money such as a 9-5 and a 2-1 shot or a 2-1 and a 5-2 shot. This seemed to me to be fertile ground to test my theory. I would look at those two horses and try to determine which horse deserved the low odds and which did not. Some races I couldn't decide, however in most races there appeared to be one horse that was clearly superior (based on my handicapping knowledge) than the other horse. I made my wagers accordingly. After a reasonable sample of wagers I was shocked to find out that I still had a net loss but even worse, my net loss was more than the track take which means I would have been better off selecting horses at random.
I then re-looked at the data out of curiosity to find out how I would have fared if I wagered on the "overbet" horse. More shock when I found out that not only did the "overbet" horse perform better than my "underbet" horse, but it showed a valid profit as well. This didn't make sense to me at first and this is where I think that some application of game theory may apply. My theory was that the one horse looked better than the other horse and therefore if they were at the same or close odds, the better horse is underbet while the other horse is overbet.
However, in thinking about this some more, it occurred to me that the fact that the one horse looked better could not possibly be lost on the public. The public does an amazing job of setting odds. The lower the odds, the greater the win % of a large sample. The public could not do this without the proven ability to analyze the data. My assumption then changed to the thinking that there must be some reason that I am missing that actually makes the horse that doesn't look as good underbet because too many people will be betting on the horse that "obviously" looks superior.
That was the birth of how I make horse race selections. I no longer try to select the "winner". No, I didn't use actual GT but I think GT in some way figures into how I select horse wagers.
Joe.Comment -
I see too many applications. Okay it is probably a bargaining game with alternating offers, and as usual there is a pie to split, but other than that it is very vague. I can even imagine situations where it is not zero-sum.
1] What exactly is the discount rate or cost of negotiation?
2] What is your outside option? What is his? What is the core?
3] What bargaining power (in layman's terms) do you have?Comment -
investing--- bahahahaComment -
how to lose all of your money from gamblingComment -
how to default on credit card by gamblingComment -
how to lose a girlfriend--etc by gamblingComment -
Mathy, instead if looking at all the possibilities out there, you define the game. The game is: do I pay my mortgage given that the Net Present Value (NPV) of the property is negative in relation to the mortgages on it? BTW, none of your questions are related to the problem. As this game has several decision levels and as I cannot diagram them here and as it will take quite a bit of time to write this up in MSWord, I'll get that done this weekend and either try to post it here this weekend, or send it via e-mail to those who PM'd me.
In the environment today, you cannot motivate any bank to modify a loan so then you must forgo making monthly payments in order for them to play the game. So, the first thing you must worry about is the ding you'll take on your credit.
The top level of the decision tree has two decisions: Do I pay the mortgage or not given that I want the bank to modify my loan? Buried within this question is, can I get the bank to modify the loan given that they know that the NPV of the loan on the property is negative? In other words, there is some value to me in keeping the property since the NPV is negative, because if I were only to use the NPV, the obvious choice is strategic default. Given the value of the game, this is the dominant strategy if only money were involved. But given that the wife loves the house, moving costs, damage to credit and a few other things, I had to calc thru all the levels of the game to see it's value if I could get a modification to the loan.
When I work thru all the levels of the decision tree, it turns out that given all the dominant strategies buried within, they must modify or they lose well over 6 figures on this particular game. When I bounce the game value against my other alternatives it turned out that the value of the game said: stop payments, and then when offered, take the modification at a certain level (also found in the game).Last edited by Wrecktangle; 05-27-10, 06:20 AM.Comment -
Re : book
Well, I read the little book in one evening. If you don't know anything about sports gambling it might help. However, I don't think someone with betting experience should get this book. Nothing really new for me in the entire book. Something a bit annoying is the fact the book is trying to sale website services of sportsinsights too many times.
I think a guy getting into sports betting should read Lem Banker book instead !
Rating :
3/5Comment -
Disaster!!!
Just finished the book and I am so dissapointed. It not only doesn't provide good information but is written in horrible style. Is so repetitive and the amount of times it says to subscribe to sportinsights services is ridiculous. I am sure I can write a better book with my 2 year experience. Avoid.Comment -
Rating: 1/10Comment -
Handicappers digest says these guys are on a 1-20 runOriginally posted by greeksportspicks
Comment -
Is there anything in this book not covered by Sharp Sports Betting, Weighing the odds in Sports betting, 2+2 SB stickies/wells/threads and ganchrow's threads?
and mathdotcom continues to own
Comment -
as with most books, you can see the table of contents and a few other pages at amazon:
Comment -
And I was supposed to know all these details of your mortgage, like how your wife feels?
Still a horrible book to recommend to sports bettors. Here is the order of types of books you should recommend to a beginning gambler:
1] A beginner's book on sports betting
2] More advanced books on sports betting
3] Statistics for dummies
4] Ganchrow.com
Someone asked about sports betting books. You replied 'here is a good game theory book that has applications to mortgage modification'. You might as well have replied "here is a great recipe for pumpkin pie; it won't help you with sports betting but it tastes great".Comment -
gamblers book club in vegas has best sports book selctionComment -
I'll check it out.Comment -
Mathy: I sometimes think you and your brother Monkey were dropped on your heads as children.
I was recommending the best book currently on the market in GAME THEORY. In another thread, I described a usage of GT. That book, BTW, is the top book being used in Universities these days.Comment -
This book was of NO help to me. Just my opinionComment -
i have read at lest 30 sports gambling books i get a little here a little there and sometimes nothingComment
