Asymmetrical Information

Or watch the very exciting movie called "Rogue Trader." The movie details the true story of a 30 year-old-trader who tried to move the Nikkei Average with his own trading.

The end result: Nick Leeson was sentenced to 6 years in prison. Barings Bank lost 1.3 billion dollars, and the oldest bank in England went bankrupt.

Right. And, of course, the hedge funds themselves have a very strong incentive to play this particular game with other people's money because in the winning years, the managers profit handsomely. In the rare disaster year, the investors are wiped out but not so the fund managers. They simply move on to the next opportunity. And the thing is, investors and financial community don't even tend to hold the disaster against the managers completely--they are still thought of as brilliant financial minds who had a sound and highly profitable strategy but who were hit by an highly improbable set of unfortunate events. Nice work, etc....

Note that another signature of this trading practice (which I dubbed "Askin Alpha" after the once-famous Granite Partners fund) is extremely low apparent volatility of returns. Thus the investment looks not only profitable, but safe.

This is very well known in gambling communities and they call it (or import it from another source?) the Risk of Ruin. Oddly, there is even a lot of analysis where a postive strategy (only making bets when you have a positive exepcted value) can lead to ruin based solely on luck. I have seen some failry techinical articles dealing with bet size/frequency/expected value vs. bank roll size (more in relation to counting cards or poker bank rolling...)

Fooled by Randomness, on which I have blogged before, covers this well. It is the author's investment thesis.

The long run return of holding any security is -100 percent, if we make the long run long enough. Any firm could blow up any day before you can get out, and picking the big companies today that will still be big, say, 200 years from now is pretty surely random. Whatever one thinks of the "equity puzzle," this is surely a part of the answer. The other name for this in the economics literature is the "peso problem."

Re: the gambling example, we can make it even simpler -- take a 15 step martingale, i.e. bet a unit and double whenever you lose, up to 15 times in a row. Such a strategy will return 1 unit all but one out of (1/p)^15 times you play it, where p is the probability of winning a single game, and of course cost you your whole bankroll that one time,... but you're free to play the odds, after all, you might die before the bad luck hits....

"Re: the gambling example, we can make it even simpler -- take a 15 step martingale, i.e. bet a unit and double whenever you lose, up to 15 times in a row. Such a strategy will return 1 unit all but one out of (1/p)^15 times you play it, where p is the probability of winning a single game, and of course cost you your whole bankroll that one time,... but you're free to play the odds, after all, you might die before the bad luck hits...."

Ahh, this isn't what I mean. Betting strategy doesn't mean you have a positive expected value on each bet. I mean, you expect to make money on every single transaction, and yet you can still go broke. It is an odd seeming thing, but without an inifinte bankroll, a .99 chance of making $2 (your $1 + another $1), and a .01 chance of losing your $1 can still get you broke.

It is much more volitle if the EV is near even, or if the chance of losing is higher... And heaven help you if the payoff is negative (.99 chance of making $1 + $1, .01 chance of losing $100, which is a -$0.01 EV, unless my math is fuzzy atm...)

Didn't Niederhoffer blow up because he was selling naked puts (basically Lo's "model" strategy) in the late 90s?

Hmm, it appears from the comments that I'm not the only advantage player who reads AI! This entry quickly brought to mind my first attempt to count cards on a limited bankroll (and the quick trip back to the train station that it earned me).

Video: Charlie Company In Action
http://bareknucklepolitics.com/forum/viewtopic.php?t=137
check it out..

"Hmm, it appears from the comments that I'm not the only advantage player who reads AI! This entry quickly brought to mind my first attempt to count cards on a limited bankroll (and the quick trip back to the train station that it earned me)."

Ah, well, not really. I was a craps/21 dealer in Las Vegas for 4 years after I left the USAF while I was trying to decide what I wanted to be when I grew up.

At anyrate, I saw some...err...interesting systems. Somewhat cynically, I prefer to believe the ones used by securities and equity traders are of the same quality.

Yes, Niederhoffer blew up shorting teeny index puts and lost his entire fund in a day. When he was questioned about his apparent foolishness, he defended his strategy by saying it was "perfectly safe... I've done it hundreds of times".

Kristian: Security/equity traders do have two advantages over gsmblers: The things they trade in are somehow linked to real companies that usually make real money, so the odds are slightly in their favor, versus casino odds which favor the house. Second, there are non-random factors affecting stock and bond price changes, which gives some chance of statistical prediction working more often than it fails, while in an honest casino the games are entirely random and any mathematically correct analysis will just tell you that you are probably going to lose.

Some people do have skills that can tilt the odds in a card game: card-counting in blackjack, and a combination of human-engineering and mathematical skills in poker. However, the casino can frustrate card-counting by simply mixing the used cards back in after each hand, or by using computers instead of physical cards, and I think they're now emphasizing video poker, where the computer neither reveals it's hand with little mannerisms nor tries to read the player.

If you're taking 2% of assets per year in fees, once your fund is big enough, the 20% of profits just become gravy, and you'd rather keep getting bigger than improving your returns. It's not like your own money is at risk if the fund loses money.

First, bets in casinos refer to real money, too. And not through a proxy of ownership of stock/bonds (unless to you want to use the exchanging cash money for checks as the stock sale, but really checks themeselves are fixed value, with no income or appreciation associated with it).

It is certainly possible to make money gambling, but you have to restrict yourself to the games where you have an advantage (Warren Buffet-ish idea) such as some 21 games, sports betting or poker. And further, to tangentially tie this to the previous thread, beating sports and/or poker realies at least as much on psychology of the OTHER bettors. The reason you can beat these games easier than the others is the house doesn't bank those games, they are realy just brokers for population bets on hand/race/game outcomes (quarterly reports?) There are even futures betting before the seasons starts (in sports at any rate).

Of course, gambling has fewer variables, and the market place is more rigidly controlled. This is a simpler system to analyze, and thus the psychology becomes more important. If you watch some of the Poker shows on TV, you'll see a somewhat distorted view of this, as each hand is not played in a vacuum, and many hands are ommitted from the final show. I would not even swear the order of the hands were chronologically correct. But again, I am somewhat cynically about stuff liek that.

Finally, as an throw away thought, remember the 'Trading Places' FCOJ Trading Pit scene? I never really understood it until I worked on a craps game that was dumping money, and had 20 people (2 more than normally allowed) throwing money, yelling bets and the seeming chaos that wasn't.

AT: that is not an accurate picture of how hedge fund operators are motivated. Yes, they could mail it in and collect 2% (or, more likely, less -- most funds charge less than 2% overhead and most are beholden to funds-of-funds, which take their share out of the operator's cut, for a large portion of their capital) on a gradually shrinking pie. But if that's what you want, you would be better off either (a) camping out in some midsize bank's risk management division or (b) retiring to somewhere cheap with better golf. Those who choose to stay in are those who want to become qualitatively richer.

In addition, investors are aware of the trader's option. It is significantly easier to attract investment to a fund if you can demonstrate willingness to commit a large part of your own worth.

When I lived in New Jersey, I used to go down to Atlantic City and count cards at the blackjack tables there. As I tell people now, the same amount of time, effort, and dedication applied to almost anything else will make more money. As far as I can see, the casinos have thoroughly countered (for a long time now) any intrinsic mathematical advantages that might favor the player.

I recently had the opportunity to help teach a class in cost-benefit analysis for public policy graduate students. These are all bright people, and although most don't have a numerate background, they could cope with the time-value of money and various deterministic concepts. Getting them to think in terms of expected values was harder, and to think in terms of the entire distribution of possible outcomes very hard indeed. Teaching people to think realistically about stochastic risk is difficult.

In the investment field, there seems to be a growing body of evidence that the tails of the distributions of possible outcomes are fatter than usually assumed. I have long suspected this to be the case, on the belief that the "system" has multiple operating modes. Most of the time it functions normally -- but from time to time there are feedback loops that get going and things can get very bad very quickly when they kick in.

Comments are Closed.