Its not clear to me how you get to this conclusion. I find it much more reasonable to conclude that Geismar reduces his bets after losing and increases his bets after winning, because he has less or more “gambling money” left. This is in fact the optimal strategy to maximize the total amount of fun he has gambling (with certain assumptions about the relationship of his fun with winning, this is a mathematical theorem), and having fun is the justification for gambling with negative monetary expectation.

It’s not very intelligent to bet the same amount of money regardless of how much money you have. In the investing world, that would be like recommending that you should invest $50k in the stockmarket, regardless of whether you have a total of $50k to invest, or a total of $1mm to invest. A sensible investment strategy would be something like “keep 50% of your portfolio in stocks”. The result of that is: if you start with $100k, you put $50k in the stock market. If the market goes down 10%, you have $95k left, so you put $47.5k into the stock market, which means you bet less after losing.

BrPH, I happen to be a mathematician who knows probability theory, and it is not at all true that winning consistently in the market is proof of cheating. That would be true (by the Law of Large Numbers) if any bet in the market had 0 or negative expectation. But, to use your construction, anyone who trades in the market for long enough will tell you that the efficient market hypothesis is just not true – people can and do make bets in the market with positive expectation. There are times when you can buy a basket and sell an ETF for a straight arb, or buy an ADR, crack it, and sell the ord to lock in a win. Those are just a couple of examples of market inefficiencies that can be exploited – such inefficiencies come up every day. Oh, I know, examples of companies who thought they had it figured out and made a mistake abound. JP Morgan just made one, and before them we had the credit crisis, and further back there is LTCM, and on and on. When you’re making bets, sometimes you will lose. When you’re making big bets, sometimes you’ll lose big. There is no question that when that happens, these firms ought not to be backstopped by the government, otherwise it creates a very unfair game that can be exploited. But do all these examples mean that the whole world of professional investing is some kind of shell game, a giant Martingale Strategy? I guess it could look that way from the outside, but if it were true you’d actually see far more blowups than you do.

In the case of QIM in particular, a lot of these comments are way off base. They have had, what, like 15 winning years in a row? Sure, I guess that could be a Martingale-type system at work (although, how many losing months have they even had). But I have pretty good info that it isn’t. I work with a guy who taught machine learning at UVA. That is a pretty serious field and a pretty serious university. He knows QIM’s CEO and tells me that guy used to live in artificial intelligence chatrooms. He says you could tell he was on to something by the questions he was asking. And he has no reason to build up QIM – he is jealous as hell that it wasn’t him. For that matter, I am jealous too – the guy is on the short list of people in the world who are smarter than I am (though he is not better at basketball). If you go to the QIM compound and see the library of books on neural networks, etc. – well, they are putting on quite a show if all they are really doing is pressing on losses. I just don’t believe it.

With regard to BrPH’s comment, the MIT team did count. A little time with google will convince you that “Burning down the house” is far more fiction than fact (as are other titles by the same author), and the only way you can know *how* to ‘cut the deck’ is to do calculations based on the count. As far as I know JP invented the technique – and the simple data logging methods – and worked out the odds and betting strategy – circa 1982. JP called it the ‘non-random shuffle.’

I was there.

- Running the hedge fund he has opportunities to play on insider information. This comes out over and over for consistent winners on Wall Street. Any mathematician who knows probability theory will tell you that consistently winning is near perfect proof of cheating (i.e. insider information.)

- To the extent that your analysis is correct, and he is not playing insider information, he is still optimizing his income. It’s a good strategy for his income to place those kind of large tail bets. If he wins, which will be most of the time, he is a hero and makes huge amounts of money for himself. Odds are that he can keep that going for quite a while. But if he loses? If he’s smart he won’t have all his money in one basket. He will still have his personal money. At worst he will get called before Congress for a flapdoodle session of no consequence. He will be able to raise more money to manage once again and odds are he won’t crash for a while once again.

————-
I will also note that the MIT “kids” didn’t card count. That wasn’t good enough to give them their consistent, large wins. They learned how to cut decks accurately and other things in addition. This is described in “Burning Down the House”.

The double bet strategy is mitigated by the casinos maximum bet limit on blackjack tables. So a $2 min bet table will have a $100 max bet, only allowing $2, $4, $8,$16,$32,$64 – so that’s just 6 bets before that strategy fails not 10. Even if it was $200 then you would have one more bet (7 bets) $128 bet. Generally the max bet is 100 times the size of the min bet which would equate to 7 bets. Casino’s also restrict the strategy with roulette, simply by adopting a max bet limit.

If you take away the bet maximum limit on the tables then you could make small returns for free at no risk at blackjack. The google / Microsoft parallel doesn’t really match the blackjack betting strategy of doubling your bet. They are only comparable as poor investment strategies. The max bet limit is the issue for blackjack players using that strategy. Adopting the double bet strategy without the limit is a superb idea! If you can invest without a restricted (low) limit then you can make lots of money in the investing world. In a world without limits the large hedge funds, can set there own limits. Using the blackjack anaolgy, the amount of bets can increase from 7 to however high they wish by moving the goalpost limit.

What The Whale was doing seems more like poker than BJ – he was playing against other individuals, not against the math. He and Macris seem to have been attempting to “buy the pot” by raising the stakes on a losing hand until the other players were forced to throw in their cards. Didn’t work then, but strangely – it could work now and get JPM out of this whole mess for damn near nothing.

In an ultra-high-stakes game, the counterparties can’t play straight-up against JPM.