+1 (307) 227-6611

Free risk tools

Return autocorrelation calculator

Does one bar's move say anything about the next? The autocorrelation of your closes' returns at each lag, against the noise band.

Free, no sign-up On sale: Desk and Desk + Signals. See the plans

  • Free
  • No sign-up
  • Nothing you type is stored

The closes

Oldest first, separated by commas or spaces; a few hundred closes give a narrower band.
How many bars back to compare; 10 is common.

The autocorrelation

Lag 1 autocorrelation -0.09: inside the noise band of ±0.14, so one bar's move has said little about the next one in these closes.

All lags at once
11.2 against 18.3: the first 10 lags together show no more pattern than chance would (Ljung-Box, 5% level).
Lags outside the band
None of the 10 lags is outside the band.
Volatility clustering
The size of the moves has a lag 1 autocorrelation of +0.12: the size of one move has said little about the next one here.
Read it with care
Measured on 199 returns; the band is 1.96 divided by the square root of that. A small autocorrelation can be real and still too small to pay the spread.

The correlation of each bar's % move with the move that many bars before, and the 95% noise band.

LagAutocorrelationBandOutside
1-0.087±0.139No
2+0.039±0.139No
3-0.054±0.139No
4+0.114±0.139No
5-0.078±0.139No
6+0.042±0.139No
7-0.015±0.139No
8+0.032±0.139No
9-0.121±0.139No
10+0.073±0.139No

Education only, not investment advice. The answer comes from the numbers you typed in, not from your account or the market. Simulated results are hypothetical and have limits: real trading has slippage, gaps, changing results and emotions. Trading leveraged products can lose money quickly, including all of it. Risk warning

Want this check on every trade?

Educational only, not financial advice; trading involves risk.

The first test a quant runs

If today's move said something about tomorrow's, the market would have a pattern to trade. Autocorrelation measures that: the correlation of each bar's % move with the move one, two or more bars before. Most liquid markets sit inside the noise band at most lags, which is why simple pattern rules rarely survive costs. The size of the moves is another story: big moves cluster, so volatility is far easier to forecast than direction.

Ask the same question across longer moves with the trend or mean reversion test and the Hurst exponent calculator, count runs of up and down closes with the close streak calculator, and size the trade with the position size calculator.

Not advice. Statistics on the closes you type; a test on past prices does not predict the next ones.

Next tools

All free tools

Pricing

Three ways in. Pick yours.

Baroq never holds your money and never trades for you. Educational only, not financial advice; trading involves risk.

Test your trading idea free

No spam. Unsubscribe anytime.

Educational only, not financial advice; trading involves risk.

or join us on Telegram