best scan for angle of crossing lines

splash88

What is the best way to scan for limiting a crossing line scan to a particular angle range?
For example if you scan for a MACD ( or ppo or pmo, tsi, etc) cross of its signal line; you probably don't want things where the cross seems either to flat or to steep. Maybe the question should be how to scan for a limited angle of ascent of those individual lines rather than worrying about the cross. What I am looking for is a way to scan for crosses that ignores things that are actually flat and without any momentum and also those that are so steep they look unsustainable or may actually pullback. Thanks for any suggestions.

• 
  

markd

One approach might be to look at histogram bars for those indicators that have them. The longer the histogram bars, the farther the short term ema has moved from the long term ema, and that happens when price moves more sharply (covers more territory in less time).

So, I think you would want some minimum hist bar height within some maximum number of bars just prior to a crossover. You would have to look at bunch of charts to determine the most useful set of parameters. I think you would use PPO instead of MACD because the scale is constant on PPO.

So for bearish crossovers where the Hist has been above zero, you might look for the Line-Signal crossover and then something like max PPO Hist in the last three weeks (or whatever) > 2 (or whatever). You would do the same for bullish crossovers where the hist has been below zero, but you would look for Hist < -2 within some timeframe.

Alternatively, with a lot more effort, on non-percent based indicators, you could figure out some minimum per cent difference between the indicator line value at its recent peak or trough within some time limit (e.g. last 10 bars, or whatever) and the line value at crossover and code for that. Probably you would have to consider how to handle both positive and negative indicator value situations.

splash88

Thanks I will try that.