Comparing to other charting sites (TradingView, Yahoo, StockCharts,etc), values appear to be off by about 10% with the latest fix.

Reverting back to the previous version fixes most of it, but needs more testing.

export function rsi($close: Array<number>, window: number) {
  let gains = [0], loss = [1e-14];
  for (let i = 1, len = $close.length; i < len; i++) {
    let diff = $close[i] - $close[i - 1];
    gains.push(diff >= 0 ? diff : 0);
    loss.push(diff < 0 ? -diff : 0);
  }
  return pointwise((a: number, b: number) => 100 - 100 / (1 + a / b), ema(gains, 2 * window - 1), ema(loss, 2 * window - 1));
}

1. why use ema here?
2. why should the window be double?

refference : https://www.macroption.com/rsi-calculation/
i don't think you should double the window.

Hi,

I'm interested in this library.
Can i use it with Angular 10 ? If yes, how ?

Thanks for your time !

Current example/index.html doesn't work well for real trading data. It creates gap in the plot.

Following fix will take out weekends and non-trading hours. If including pre-post trading hours in the data then this shouldn't be an issue, but most of the users may only be working with trading data.

margin: { r: 5, t: 5, b: 50, l: 5 }, xaxis: { autorange: true, type:'date', rangebreaks:[{bounds:[13, 6.25], pattern: 'hour'}, {bounds:['sat', 'mon']}]}, yaxis: { autorange: true, type: 'linear' },

I've worked with few market data apis and their t (time) data is usually provided in seconds. I was trying to understand what format ta-math uses, but am not sure, is it also seconds or do I need to multiply them by 1000 to get milliseconds?

I think the current implement of Stochastic Oscillator (Full) is not right. Instead, it's the implement of Stochastic Oscillator (Fast)

current algorithm:

export function stoch($high: Array<number>, $low: Array<number>,
                      $close: Array<number>, window: number, signal: number, smooth: number) {
  let lowest = rolling((s: Array<number>) => Math.min(...s), $low, window);
  let highest = rolling((s: Array<number>) => Math.max(...s), $high, window);
  let K = pointwise((h: number, l: number, c: number) => 100 * (c - l) / (h - l), highest, lowest, $close); 
  if (smooth > 1) { K = sma(K, smooth) }
  return { line: K, signal: sma(K, signal) };
}

Below is the description about difference between fast, slow and full.

Fast Stochastic Oscillator:

Fast %K = %K basic calculation
Fast %D = 3-period SMA of Fast %K
Slow Stochastic Oscillator:

Slow %K = Fast %K smoothed with 3-period SMA
Slow %D = 3-period SMA of Slow %K
The Full Stochastic Oscillator is a fully customizable version of the Slow Stochastic Oscillator. Users can set the look-back period, the number of periods for slow %K and the number of periods for the %D moving average. The default parameters were used in these examples: Fast Stochastic Oscillator (14,3), Slow Stochastic Oscillator (14,3) and Full Stochastic Oscillator (14,3,3).

Full Stochastic Oscillator:

Full %K = Fast %K smoothed with X-period SMA
Full %D = X-period SMA of Full %K

compare with current algorithm, current implement is the fast one.

I recommend modify current description, change Stochastic Oscillator (Full) to Stochastic Oscillator (Fast). Meanwhile, imply another two indicator.

Reference: here

I got an error in vscode saying that types are missing. After glancing at package.json they seem to be defined, however it appears that actual dist/ta-math.d.ts file does not exist. Never worked with rollup, but I'm pretty sure it is due to 'declaration': false in rollup config? https://github.com/munrocket/ta-math/blob/master/rollup.config.js#L15

If that's the case, I can make quick PR to change it to true 👍

Hey 👋

I came across this lib today and after using it for few calculations, I think it has great potential (I found it very hard to find good ta libs for typescript env, that are also well tested).

It was a bit tricky to get started due to little documentation and I found that example a bit overwhelming as it was doing a lot of stuff related to drawing charts. My use case was simpler, I just want to use this lib to perform ta calculations in NodeJS env and get indicators.

This lead me to idea of how I might be able to help here in terms of contribution, would you be open to following idea?

I want to create a set of small https://codesandbox.io examples for each function that just use small dummy data array and mainly aim to provide good examples of how each function should be used. These examples could than be added to your tables in readme perhaps?

Latter on if I have time I could try to also add charts showcasing each indicator individually, but for time being I just wanted to output indicator values as text on the page.

Let me know what you think, would love to help 👍 will also help me learn the lib better.

There is an error with this repository's Renovate configuration that needs to be fixed. As a precaution, Renovate will stop PRs until it is resolved.

File: .renovaterc.json
Error type: The renovate configuration file contains some invalid settings
Message: Configuration option bumpVersion should be a string

this.

is test Pars fix?
I want to use it at https://softscients.com/stocks
thanks a lot