However you could get good indications not only on how to optimize your trading for better profits, but also on how to keep your risks in check. Starting again from Kelly's formula: : F = ((R + 1)*P - 1)/ R where: F is your optimum bet as % of your money, P is the probability of winning, R is the ratio (winning amount / bet) you can rearrange it this way: R = (1 - P)/(P - F) This gives you several good hints: if you reduce F (% of your account you risk) you need a lower R ratio of average win / average loss, to be profitable F reduction has an impact on your R only as long as it is significant in relation to P probability of winning; bellow F < P/5 its reduction doesn't have much effect anymore for a P=33% (i.e. for a trend following system), your F < P/5 means F < 6.6% while for a P=67%, your F < P/5 means F < 13%
I agree, the Kelly formula only uses past trades to calculate risk. Using Kelly the assumption is that the risk in the future is the same as the trades used to make the calculation. An unscheduled event such as 9/11 will find the Kelly user over exposed. One of Taleb's suggestions was to 'use historical data for developing a trading system, but never use historical data to set risk limits' as a 'Black Swan' even can occur at any time. Makes sense to me. Worst case for a trader: combine a historically overfitted trading strategy with a historically overfitted risk management calculation. Good Luck!
I don't see how having 20 positions, each 5% of your account would help you on a day like 9/11. You'd just have 20 positions to repair. The damage would occur no matter how you spread the risk, unless of course you were in synch with the drop down and had shorts on that morning to your eyeballs. If you say that a money management technique can account for 9/11 type disasters i would have to see it to believe it. If you were long; you went for the ride; simple as pie. This is not to discount money management, it is simply my view.
You might find interest in an article I wrote for the magazine. I posted the whole thing (temporarily only). www.stocktrading.com/riskreward2.html BTW, if you would like a free one year subscription to Stocks and Commodities magazine, send me an email with "TASC" in the subject line. send to don@stocktrading.com Don
People quoting 5% talk about the cumulated risk of all their open positions not to exceed it. If they had 20 positions, and they spread their risk evenly, they'd risk 5/20 = 0.25% per each position. Also, this 5% risk usually refers to the stop-loss you decided on. When sizing your position you have to plan for your maximum risk too. Somebody who thinks about trading naked puts may set his stop-loss not to risk more than 5%, but he has to protect himself for a "9/11" type event, i.e. by buying some cheap further out-the-money puts.
Understood. I was referring to a quote from psytrader "for long term positions, no more than 5% of capital is parked in any one stock." On a "black swan" day each of those long positions would be toast. For options, sure you can hedge against such events. But how would an equity trader avoid such events? With money management?
Compared to options and futures equities lose less percentage wise, but you still have to calculate your stop-loss based risk, and pad it for deep gaps. Also, many equity investors besides being well diversified (asset allocation) hedge their positions with options (i.e. puts, collars) and futures, especially on indices.
The only way to hedge against black swan is to always have the most updated information and to always have derivatives that somehow hedge your market exposure. The key is to delete most of your longs while putting on to your short side. This can be done in milliseconds these days but you have to have the information before the exchanges, associated market makers, and specialists. I would forget any notions of getting the info before them. There are ways though to do almost anything.
Unfortunately black swan events have a higher probability of happening when markets are closed. They're open only 5 * 6.5 = 32.5 hours out of a 168 hours week, or ~19%. You can't do anything when facing a gap at opening.
Planning/ optimisation is one thing. However the reality is practically how we are going to deal with the issues such as the risk of ruin due to unexpected straek of losses or else, particularly when running a system automatically.