I have decided to do a monthly review of trades, a month after the end of the month. This is partly to give myself a bit of distance between losses and analysis, but mostly because proper analysis of what went wrong is impossible without knowing what subsequently happened in the market and newsflow.

There were four trades closed in January, all losers. They were long sugar, long platinum, long soybeans and short EUR/USD. In the first three cases, a long-standing trend entered a choppy patch, and in the context of my trading process I am inclined to put the losses down to bad luck. Trend-following is not a good strategy at inflexion points. The EUR/USD trade was a little different, as my entry market the exact point at which the trend reversed. I am still trying to work out what to do with currency trades, but certainly I think it is a bad idea to trade an old story at a new low, especially as bearishness on the EUR in the news media was at an extreme. More work is required here, and I am doing it at present. Also, I was too slow to react to the inflation meme which has carried the EUR and GBP upwards in the past two months.

The real story of January, however, is one of missed opportunities. There were several excellent trends which did well in simulated trading: short Eurodollars, long cotton, long coffee, long cocoa, long silver, long corn and long equities. Why did I not trade them?

  • Eurodollars: I preferred to trade Treasury futures, but since rising expectations for rate increases were the driving force here and eurodollars broke out first, I should have traded them. With hindsight, waiting for a breakout in Treasuries was unnecessary (and indeed, by the time they did break out the move was done, and hence I lost money in February on a short Treasury position). 
  • Cotton: I was nervous of how far the trend had run, particularly because cotton is a gappy market with tight daily trading limits. I was afraid of getting locked limit down on a long position. However, the spectacular gains that could have come from trading this market far outweigh any potential losses of getting locked limit down. Such things are a great concern for the highly leveraged, but with my relatively wide stops the danger is much less.
  • Coffee: The story here is simple: breakouts have not worked historically in this market, and they have started working in the current trend. I think I was too slow to react to this — caution prevailed when I should have reacted to the strength of the trend and started going long. 
  • Cocoa: I was too slow to trade, and therefore did not make as much money as I could have on this trend (I took profits in February, so the trade does not technically fall into January’s analysis; subsequently the market broke out again, and I was again too timid — because historically, only the early breakouts have been good in this market — it is prone to sharp reversals).
  • Silver: Gold pulled back in January, taking silver with it, and I didn’t fully understand why. When silver broke upward again, I was still cautious. But with inflation stories circulating and political instability in the Middle East, there was a fair chance that the upward trend would continue, and I should have traded.
  • Corn: I was too cautious about getting into corn, because I was long wheat and soybeans at various times in the month and did not want my trades to be too correlated. This was a mistake — corn was the best market. Fear of correlation has tended not to serve me well. Also, the best breakouts were on relatively low volume. Having done some analysis, I have found that a volume condition actually detracts from my system, and have removed it.
  • S&P 500: I missed the best signal of the month, a double-breakout, because I feared markets were settling into the year and had not yet found direction. In other words, I was doing the kind of wimpish fretting that is common among investors. The trend was up and I had a signal, and I should therefore have traded.
You will notice that all of my reasons for not trading were technical in nature — based on whether there was a simple trend, or my indicators, or correlation analysis. I am still finding the balance between fundamental and technical analysis in deciding whether there is a tradeable trend, and here I think I have let various kinds of technical analysis take too much of a lead. Logically, I should think like this:
  • If there is a clear technical trend and a signal, trade.
  • If the technical trend is unclear but you have a strong fundamental view, and there is a signal, trade.
When it comes to my breakout success indicators, I think I should be able to overrule them if a strong trend seems to be developing. Breakouts work in strong trends, regardless of what has worked in the past in a particular market. And the basic premise of my process is that you can make money by going to where the trend is. Also, I should think less in terms of having to make a judgement call, and more in terms of probabilities. If there is a 50/50 chance that there is a trend, and a 50/50 chance that my signals will work, then that should be good enough.
The various reasons that I had for not trading also give rise to a list of things that I should not be allowed to use as excuses for not trading. In particular:
  • Trades are too highly correlated. So what — the fact that markets are somewhat correlated does not mean that the success of trading signals will be correlated. 
  • Market is too gappy. This doesn’t matter — you still make more in a good trend than you lose if the market gaps through a stop. 
  • The story is old. Analysis last month suggested that old stories offer better returns.
  • I don’t understand why there has been a pullback. This suggests that it is probably random noise, not a reason not to be long.
  • Breakouts have only just started working. If they have started working, there is a reason — a stronger trend than usual in that particular market. Again, this is a reason to trade, not a reason to avoid trading. I may balance the strength of the trend against the past efficacy of breakouts, but not wait for the their present efficacy to be convincingly proven — by then it will be too late.
The last point raises the question of how I should use my breakout success indicators. I think I should ask the following questions:
  • Have breakouts made money in this market?
  • If not, have they made money when the trend has been in the right direction?
  • If not, could the current trend be strong enough to make them work?
If the answer to any of these questions is “yes”, I should trade. And “yes” means there is a fair chance — say 50/50.
I was feeling bad about my process, but having done this analysis I am frustrated with myself. I am sure that the process can work. I can identify tradeable macro trends, using fundamental and technical analysis. You would think that would be the hard bit. But the hard bit is the implementation, and mine remains lousy.
Advertisements