Stochastic Oscillator

[DaveSkender]

Community discussion about Stochastic Oscillator and KDJ Index

[DaveSkender]

I've just added the KDJ Index %J value to Stochastic Oscillator. See #315 for the background and assumptions.

[vebollu]

[vebollu]

Can you please add Fast Stochastic property too?

[DaveSkender]

A fast variant can be done with parameter argument smoothingPeriod=1

[vebollu]

ty, it got worked.

[netprok]

Hi, thank you for your efforts and help.

I am using your np and trying to get KDJ results, but the PercentJ and Signal are missing.

[DaveSkender]

Most indicators have a warmup period where results cannot be calculated on the front end of your historical quotes, so they are null. With the parameters shown, I'd expect the first 18-19 periods to have null %J results.

If you want the library to automatically exclude the recommended amount of warmup periods, use the .RemoveWarmupPeriods() extension:

// usage if you want to remove the 
// recommended warmup periods
IEnumerable<StochResult> results =  
  quotes.GetStoch(14, 3, 3)
    .RemoveWarmupPeriods(); 

See the Utilities Guide for more information.

[DaveSkender]

r.PercentJ = (3 * r.Oscillator) - (2 * r.Signal);

Given how %J uses oscillator and signal values, I suspect you're not using enough historical quotes. The bare minimum you'd need to get all three values is lookbackPeriods + smoothPeriods + signalPeriods, so you'd have to update your NeededCandlesCount if you are using that to limit the size of your quotes.

With that said, I think most people will use larger datasets in general, not just the bare minimum.

[DaveSkender]

Another discussion on KDJ, compared to other charting platforms - see #612.

If anyone can find source materials from the original author, George Lane, let me know. It would help resolve the debate over which recipe to use.

There are a few opinionated variants out there and they all use a slightly different approach; none of them being more "accurate" than another, just slightly different.