I have spent the morning going over all my trades up until the middle of February. I have classified trades by the type of technical signal, and also looked at whether the stop position meant a profitable trade was stopped out before it had time to run.

I have tried a lot of trade selection signals in the past two and a half months. Most occurred too few times to get any meaningful information. The three major types have been breakouts (where the market moves outside a recent trading range, and one bets that it will not re-enter it), range trades (where the market approaches the limit of a recent range, and one bets the price will stay within the range), and trend trades (where the market is trending and the trader jumps on the bandwagon).

The lines that say “assuming stop not hit” are an exercise in imagining that I had placed a wider stop than the market retracement before a subsequent move in the expected direction. There is some judgement here — if the market broke out and then came back strongly, before effectively breaking out again, then the move was not really tradeable and I have not included it in the “assuming stop not hit” category.

Here are the results:

As you can see, the technical signal with the best success rate has been the breakout, but it is the signal with the worst realised success rate. This means that I have been placing my stops too tight, or not waiting for enough of a retracement before entering on a breakout signal. Attempts to trade a range have had a lower potential success rate, but my execution has been better and I have only been stopped out when the range has convincingly failed to hold. My attempts to jump on trends would have had a poor success rate even with wider stops.

I have also looked at whether it is better to enter a trade in the direction of the expected move (With Trend) or on a reaction in the direction opposite to the expected move (Countertrend). Note that my use of “With Trend” and “Countertrend” means this and with hindsight is not very intuitive.

This table doesn’t say very much about the right way to enter breakout trades. This is consistent with my stops having been too tight, or not waiting long enough for a retracement before entering: my attempts at countertrend entries have not been countertrend enough, so they are not markedly different from my with-trend entries. For ranges, the table shows that I have had much more success trading ranges when the the market has already hit the limit of the range and come back into the range, indicating that the range has held. For trends, I have had no success “going with the trend” and only small success trading against the trend; and even that success represents one trade that was as much a range trade as a trend trade. Jumping into trends has not been a good way of making money.

Finally, I have looked at the ratio of the profit on winners to the loss on losers.

There does not appear to be a significant difference in this ratio between breakout and range trades. The figure for trend trades is highly distorted by the one trade that was a trend within a range — if that is moved to “Range”, then the trend figure is zero and the range figure is 4.

Conclusions:

  • As I expected, breakouts have been the best trading signal.
  • I was surprised at how badly I have capitalised on breakouts. I have been stopped out too many times when I should have been in the trade. I should be placing wider stops, and waiting for better re-tests of breakouts before entering (breakout test has a fairly low success rate, but that is because of a number of bad trades, where either the breakout or the subsequent test was not very convincing, in a small sample).
  • Range trades have been more successful than I expected.
  • The Jesse Livermore approach of waiting to see whether a range holds before going in has served me well so far.
  • Trend trades have been hopeless. I should avoid them.
Update: I have had a look at what my performance since inception would have been had I traded with a constant 0.25% risk per trade, and what it would have been had I done only breakout and range trades (including the ambiguous trend-within-a-range trade, which would not be excluded by such an approach).
Advertisements