wilmott.com

I really don't understand why so many trader use stop-loss.There seems to be some widespread belief that you can actually limit you're loss this way. We've all heard catch phrases like "Cut you're losses short and let your winners run". I've heard trading coaches preach about the importance of stop-loss with claims that if you only learn to put a limit on you're trades and use trailing stop loss, you will be ahead of 90% of the market and be able to make money straight away.But this just doesn't make any sense. The only way you would ever limit your loss would be if you stop trading all together after you're trade is stopped out. Otherwise it is the same thing as if you were playing roulette and switched color ever time you lost a hand, and called that a risk management strategy. I understand the you'd want to cancel a trade that does not play out as you had expected, just because that would indicate that you were probably wrong to start with. I also understand that you may need to cut certain positions to keep the risk management team with their VAR figures happy. I also understand that in a crises situations where just everything is going against you, you'd want to cancel all trades and take a timeout to review the models and analyze the situation.But to just say "If this contract plummets to 30 Ill sell" is not a way to manage risk. Why would 30 be a place to sell? Because That means that I've lost 25% of my initial investment and I don't want to risk any more on one single trade..? This is actually what one trader at a fairly large bank told me, and it is probably one of the most stupid things I have ever heard. The only way I would ever sell something that just got cheaper would be if there was any reliable indicator to suggest that the price would keep going down and in that case I wouldn't just sell but turn the position and go short.I'm interested to hear what people at this forum thinks about using stop-loss, and how you are using this weird form of risk control in trading activities.Cheers!

hola, peligroso. a stop-loss strategy is not perfect nor from the academic's (see Hull int the context of a hedging straegy), nor from the practicioner's points of view. nonetheless it makes sense to define a very strict trading framework, doesn't it?

For any quantitative alpha generator strategy I'd say a strict trading framework is generally a good idea, but just the fact that rules are strict does not make them good. Could not find the Hull paper you are referring to.. any link?

In general, I agree that stop-loss orders suck because: 1) stop-loss orders provide false-confidence that one can only lose as much as the stop-loss price (people who backtest with stoplosses often miss this issue and over-estimate the performance of their strategy); 2) stop-loss orders assume autocorrelation or momentum of returns; 3) stop-loss orders destabilize the markets.On the other hand, there's no denying that: 1) future volatility strongly correlates with recent significant price declines; 2) that "unexpected" price changes should reduce one's confidence in one's model; 3) there may be other trades with a higher expectation of returns than a trade that's "going the wrong way".If one is going to use stop-loss orders, one should: 1) have a better model than "cut you're losses short and let your winners run"; 2) carefully estimate stop-loss order performance; 3) have some evidence or secondary indicators that a price drop should produce loss of confidence in the trade or trading model; 4) have a coherent trade opportunity ranking/allocation system that shifts capital from lackluster positions to better trades.

If you put your stop on and allow approx 20% slippage you wont be too far off your max loss. I personally think they are invaluable provided your stop is wide enough not to get stopped out in the more volatile markets before you see your strategy come good