The important column to me is the “Cash” column, which shows realised profits and losses. The “Equity” column includes instances where a position went in my favour and then fell back before I took profits. The sharp-eyed will notice that I am ahead YTD:
Cash MTD 0.72%
Cash YTD 0.40%
Equity MTD -0.56%
Equity YTD -2.13%
I have done some thinking this month about when to take profits — in fact I have been thinking about it a lot. Up until now, I have stuck with the same method — i.e. take profits on a position-by-position basis, and I decided this month to keep trying it. My argument against cutting everything when I have hit a certain profit target was that it was inconsistent with a trading strategy that consistently presaged a decent move. It would be daft to cull trades early when they were just starting to work.
However, I wonder whether this approach places too much faith in technical analysis — which I think should be approached with scepticism. At present, my entries are based on strong technical signals, but that means it is impossible for my exits to be (say I buy a new two-month high, and only sell on a new two-month low — each winning trade lasts at least two months and one day; proper winners would probably take several months. Even if this worked, the return distribution would be unacceptably lumpy).
1. At present, therefore, I am selling on weak technical signals — short-term reversals.
1.a. These happen often, curtailing profit potential. Sometimes I have expected them from the market action and newsflow and exited before the reversal happened — the desire to get out when a reversal is due is overwhelming; but there is no reason to think a short-term reversal heralds a change of trend, and this approach will mean I tend to get out too early.
1.b. A small amount of work with my new TradeStation system has suggested that a per-trade profit target is superior to a short-term reversal strategy for taking profits.
2. I started trading hoping get into trends with a good entry point. The obvious way to do this is with a breakout entry system and a trend-reversal exit system (i.e. price falls below moving average (MA), or MA1 falls below MA2, etc.). This is the same idea as exiting on short-term reversals, but is less sensitive to short-term market action. I will test it with TradeStation.
3. Some work in TradeStation has also suggested that my entry signals are not as strong as I had hoped (although of course it’s hard to tell a computer to trade only breakouts from a well-established level and with support from the newsflow; I suspect the signals I am trading are better than the ones in the model).
3.a. Preliminary results show that they make money, but with a lower success rate than I would like.
3.b. Further, in many cases the stop level implied by the breakout (i.e. the level to which it is unlikely that the market will return) only lasts a few days. Perhaps what a breakout gives is a chance to get in and not be stopped out immediately; after that, new factors might come into play — after all, markets are affected by stories and also by reality (although the latter connection is considerably looser). If that is the case, then it would make sense to bring an element of random profit taking into a technical system, because it would mean some positions are cut at a profit but before they reach a profit target. I will test this idea with TradeStation as well.
Psychology of the Current System
It is a weakness of the current system that is rational to take profits when positions move against me given that I have a desire to show positive performance for the month. If I had strong conviction in my profit-taking approach then I would be prepared to forego positive monthly performance for long-term returns; but it would be unreasonable to have that strong conviction because of the various considerations above. Thus when positions move in my favour, I think: “Don’t cut now, keep to your profit-taking rule”. When they move against me, I think: “This could go against you further and give you a losing month.” So the proft-taking approach collapses into “Sell when you feel worried.” As an exit rule, this is deranged (although I suspect it is widely used).
Taking Profits in April
I am going to test profit-taking methods over the coming months with TradeStation. For April, however, I am going to use what feels like it would have been the best system — a 5-risk-unit profit target per trade combined with a willingness to cut everything when I have made a good return (I seriously considered taking a profit of 2.11% this month — but that wasn’t the profit-taking system I was running at the time, so I didn’t). The position-based profit target addresses the short-term-reversal problem discussed above. And a willingness to cut everything when I have made enough introduces the element of randomness discussed in 3.b. As before, I will also cut a trade if the story is changing.
I am also going to introduce the idea of a profit cushion. I will be more inclined to cut trades if I am doing badly for the month; if I have some profits locked in, then I will be more inclined to let them run. This makes a positive monthly return an explicit objective that overrules trade-by-trade profit objectives. The danger is that I will give up too much return by cutting good trades too soon; but the alternative has not served me well so far, and there is always another trade (when I started trading I was worried that I would not get enough strong signals to be able to take the attitude that there is always another trade; but there have been enough over the past few months).
Should I define a good return? I think that it depends on the situation. If four trades are running strongly in my favour, there is the potential for more return; if one has made 5 risk units and two others have made 1 risk unit and then hung around for a few days, there is less potential for more profits. So I am going to play it by ear. A systematic trader would say that this is leaving the outcome to the chance winds of emotion. But a human mind is not just an inferior sort of computer — its outputs may be somewhat inconsistent, but its ability to process diverse information is enormously greater. I can always test deterministic profit-taking rules in TradeStation.