All that matters is the trade expectation characteristics on the current setup or open position. If a trade is a slam dunk, then it should be traded identically, regardless of the prior path of your equity curve, whether you are on a 10 trade winning or losing streak etc. Getting 2:1 odds on a fair coin flip demands identical bet size, regardless of how many heads or tails came previously, whether you won or lost before. The only way that a prior run can influence size rationally is when a bad losing streak calls into question the whole viability of a strategy. You may then think you have a good setup but in fact the market has changed and your setup sucks. But to conclude this, you must have a pretty long losing streak. The typical "push your winning streaks" trader sizes up to the max after a handful of winners, and cuts back after a handful (3 or 4 maybe) losers. This is totally irrational as 2, 3, 4 winners or losers in a row says nothing at all that would lead you to believe a setup has degraded or improved meaningfully. Your equity curve displays almost no useful information about whether the current setup facing you is one that should be bet big, normal, small, or nothing at all. Size based on the current trade expectation from the current market price, and ignore any considerations that don't directly impact that assessment.
I have to respectfully disagree. You're forg+ iftting the psychological factor. It can be very mentally draining to not size down after a losing streak. Sure, you can screw yourself because it takes longer to make back the money, but it can also save your ass. On a winning streak, it's not necessarily random like with coin flipping. The market might be particularly suited to your style of trading that day, and to succeed in this business you really have to kill it when it's there. When there's panic and people are throwing money away left and right, it's your duty to claim it. Streaks don't have to be random, they can have causation which, when you're aware of, you can factor into your position sizing.
Depends on one's style of trading. In discretionary trading or in manually implemented systematic trading being able to discern the change of background conditions for entering and following the trade and taking advantage of it when it continues to be favorable is one of the qualities of a successful trader. Most people on ET seem to think that there is no logic to price movement but there actually is and all traders are, to various degrees- some to a negative degree - are tuned to it. The good traders survive because through their well worked out and practiced routine are more in tune with the market's logic then inexperienced traders. Even with that, since we are human, our efficiency in implementing a successful routine fluctuates. They are times when we are totally alert and other times we are blinded to conditions ahead us. There are times when optimal conditions exist in the market and they are times when we are in peak performance (not dollar wise but in being able to adhere to our plan). These two do not always overlap. So, does it not make sense to make hay while the sun shines, the market is offering choice conditions and we happen to be in peak condition? Cheers, bc
I partially agree. I agree with you that the shape of your equity curve is irrelevant in making investment decisions. Although that is not the case when you factor in risk exposure from the use of leverage, in which case your equity curve necessarily becomes relevant. When you factor out contingencies that relate to what you can trade and on what margin then you have a point. When it comes to making investment decisions you should have a form of hypothesis testing which works regardless of what the equity in your account is doing at time interval T. Ofcourse the same market trend that's eating away your equity can be reason to invalidate your hypothesis, but in principle they are judged independently. If you conduct a thought experiment where all market participants act only on the information provided to them by their equity curve then you're left with a situation in which all of them act completely randomly. So clearly, having no form of hypothesis testing which is independent of your equity curve is irrational.
I disagree. Price is a derivative of human interaction and each situation is not necessarily independent from one another, eg random. The stock market possesses a herd mentality that at certain intervals is strong and at other intervals is weak. If, for example, your long signals are suddenly hitting return pretty good numbers at the turn of the market bottom, then you can be confident that the profile of the market situation have change for the better for your long strategy. Thus, a streak either positive/negative does fundamentally impact your expectancy and must be considered when sizing your trades. However, I do concede that any streak could be simply considered as noise, but this is something that a trader must be able to recognize. It takes skill to be able to separate the wheat from the chaff.
Taking more losses than normal could indicate the market conditions have changed and the strategy could stop working until conditions change back. Although basing size on equity curve didnt succeed for me because it resulted in always taking the losses but missing out on the winners.
Unless the odds of winning your next trade are somehow related to whether or not you won your previous trade, then position size should be the same.
Very good post. Constant sizing is acceptable, but if you want to make really outsized returns without taking too much risk then pressing your bets when winning is a very good strategy. If you are a discretionary trader then there is no doubt that when winning you trade better as your are more confident and are more likely to wait for the best setups and place your stops and targets in the correct places. This principle is one of the key reasons behind my exceptional results over the last couple of years.