It has been pointed out to me that using both issues and volume together comes close to being duplicative, and I think that's true to a large extent, as these charts show:
Even though issues and volume come out quite similar, I think they are different enough so that having them both represented makes sense, but perhaps they should not each be one-third of the total data set. Thus I am leaning toward using the following weighting:
1 x issues 1 x volume 2 x new highs/lows
Any comments?
have you ever thought about using a ROC indicator of one of your breadth indicators to spot turns before they actually occur? I've been think about this lately but haven't worked on anything yet.
I hadn't thought of it before, but your question prompted me to try a couple things today. I first tried a plain ROC with 10 and then 20 periods, but it simply doesn't plot a line at all.
Next I tried various forms of MACD, with much better success. Here's what I've settled on for now; it's a 13-34 MACD using standard EMAs. It's not exactly perfect, and I'll have to figure out some way to distinguish between a turn and the beginning of a negative divergence between the MACD and the indicator itself, but I think it shows promise. What do you think?
Note: In the following charts, the spaces between the vertical dashed lines represent 10 trading days each.