Curious that both LTCM and his system would both be negative expectation. Could it be the system is a very small part of successful trading? So Bill, if I understand your reply, most traders are actually gamblers who think that they are traders?
This strategy is similar to but probably better than regularly buying lottery tickets. Chances of hitting the jack pot is extremely low, but if you did, you never have to trade again. This may be a worthwhile expense if it is a tiny tiny percentage of your cash flow or net worth. Say you spend 5 dollars each month and continue this for 50 years. That's only $3000 in 50 years! Most of us spend that much in lunch and coffee money in a year.
For further information on this method please google search maclean's retirement 649 http://www2.macleans.ca/2011/03/08/retirement-649/ 'A 58-year-old underground welder in Timmins, Ont., jokes that lottery tickets comprise âthe high-risk portion of my investment portfolio.â' Of course another more common method is to pick among several losers for political office and hope that that works out as well.
It was totally expected and in fact the US was put on "credit watch" some months earlier which is how a credit rating agency gives the world a heads up.
Okay, on a more serious note, may be one way to check the OP's proposition is to simply run the losing algorithm without commission and slippage. If it is still losing big time, then the reverse algorithm might work.
Okay, I forgot to mention: it not only has to be losing big time, it has to be losing more than the expected commission and slippage in the reverse algorithm.
LTCM had a high expectancy system AFAIK but they overtraded it. They traded something like 20x. Volatility killed them. Anyone who trades a negative expectancy system is a gambler. He relies on luck to make money, not on consistent performance. However, in the same blog I quoted before there is another interesting post about Parrondo's paradox. This is possibly cutting edge research at this point. The paradox claims that you may be able to combine two negative expectancy systems into a positive expectancy system. I think this is something worths looking into. Very interesting concept. I Need Two Losing Investment Strategies to Win I am sure some people here like MAESTRO are very interested in this concept and may have looked into it already.
The problem with looking at expectancy with systems such as LTCM's or taleb's... is that expectancy is calculated based on the "average winner & losser" and the probability of each of these 2 outcomes. When dealing with swan events the average becomes irrelevant... sure the average winner and avg lossers of LTCM gave them positive expectancy, and all their Phd Nobel laurette geniuses said that it was impossible for them to go bust... but then the unexpected happened... none of their models could've predicted Russia's default and they took such a loss that it made the average winner and loser irrelevant... they got killed by a tail event, not by the average. Taleb's system is just the opposite... take small lossers and hope to profit from a tail event. At the end of the day, both LTCM and Taleb's empirica fund ended in the same place (bankrupt) the difference is that LTCM nearly took the entire financial system with them... while Taleb just lost on a predictable and controlled pattern... so which one was the real gambler?
I agree with much of what you say about swan events and the statement below is dead on accurate. But the problem with their model is that it was, although constructed by brilliant men, quite naive at its very core. Everyone knows that Russia had been a gangster society for the 70 years of communism but men who run that kind of capital should have also factored in the fact that it was a hard core gangster society for hundreds of years prior to communism. Culture trumps all and they should have known that. To think you can run financial models that are exclusively mathematical in nature against a backdrop of a gangster culture is beyond naive it is insane. For a guy like Meriwether to let the quants run wild with pure (or nearly pure) math and not temper it all with the fact that there was not (and is not) a reliable rule of law to depend on makes the model one dimensional. A reasonable and balanced model could have factored in that bonds from that type of society need a lot more vol factored in. Quants have their place but grown ups are supposed to stay in charge. When they don't craziness takes hold.
Just to be clear it is not that I think LTCM's demise was predictable but I think the possibility of disaster there was much easier to see than many pretend. What is amazing is that the amounts at risk seem quaint today ... lol.