The answer is very simple - if you have one signal that provides a significantly higher expectancy than another, then clearly you want to have a bigger position when you have the high-edge signal than when you have the low-edge signal. Agreed? So, if you have a small position on due to the low-edge signal triggering, and are showing a profit, and then whilst this position is still profitable you get the big-edge signal, then you should increase your size up to at least what you would normally use for the big-edge signal. Thus you are increasing your position when the market has moved in your favour - i.e. you have pyramided.
There is a second case where pyramiding is justified - that is where you have two different signals of similar expectancy, but your research has shown that when they both occur together, the overall expectancy is significantly higher. In this case, because the odds have significantly improved, you should increase your position.
One example - your counter-trend system indicates that the market is oversold and due for a large bounce, however the trend is still down so you only take a small risk; then, your short-term momentum system shows that the market collapse has stopped and the price is bouncing - with momentum now in your favour rather than against you, your odds improve so you increase your position (pyramiding); then, Maria Bartiromo comes on CNBC and starts saying that maybe now is the time for retail investors to go short stocks - you immediately double your position as the expectancy has gone into the stratosphere with this sure-fire contrarian sentiment signal. Now although you are adding to your position at higher prices, you are doing so because the odds are better at those higher prices than they were lower down. The market is *more* of a buy at those higher prices.
Now, if you compare this to someone who put on a double size position as the market was in free fall, or someone who entered small and didn't add to their position, I hope you can see why pyramiding would produce superior risk-adjusted results.
P.S. you say that you give your signals equal weight. Assuming that you have researched them all and found that they have similar expectancy, then that is fair enough, and you will not pyramid for the first reason I gave. However, have you researched what happens when you get 2, 3 or more of your signals flashing together? You may find that the historical odds of your trades would have been much better - in which case you should research whether a pyramiding strategy would have provided superior results.
