Sorry, I haven't made the time to test this idea, but I post it here, in case anybody wants to investigate further. I actually thought of this some time back, but was too embarrassed to post something that I hadn't tested properly, in case it turned out to be a dismal failure. And it seemed so simple that I figured it must have already been tested, and discarded.
My first take was that an uptrend is defined by progressively higher BAT buy signals, and higher sell signals; a downtrend by progressively lower ones.
Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.
I had a quick look at GBPUSD,H1 and it seemed to filter out some losing Trade#1 trades. Obviously it would need to be tested more exhaustively, to see whether the winners filtered out (especially Trades #2, #3, #4) offset the losing Trade #1s. As far as I can tell, the situations where winning Trade #1s get filtered out are where:
My first take was that an uptrend is defined by progressively higher BAT buy signals, and higher sell signals; a downtrend by progressively lower ones.
Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.
I had a quick look at GBPUSD,H1 and it seemed to filter out some losing Trade#1 trades. Obviously it would need to be tested more exhaustively, to see whether the winners filtered out (especially Trades #2, #3, #4) offset the losing Trade #1s. As far as I can tell, the situations where winning Trade #1s get filtered out are where:
- there is a significant enough retracement following a huge winning trade (then the retracement is missed);
- there is a deep enough retracement in the prevailing trend (then the next impulse move is missed);
- prices are moving in some kind of widening wedge.
An obvious variation would be to take smaller size positions (possibly for Trade #1only?) when trading contrary to the above rule. The trend determination could also allow re-weighting the sizes across trades #1, and #2-4, if the trend is particularly strong (e.g. three successive alternating signals are progressively higher, or lower).
David