Exocharts candle types quick reference

TIME-BASED

Milliseconds, Seconds, Minutes, Hours, Day, Week, Month, Session.

Standard candles.  
Starts UTC day start by default. 
Week starts on Monday.
Month Starts at Month first day.

QUANTITY-BASED (Rotational)

Tick:

Closes after X trades. 

Filters noise during low activity periods.

 Normalizes market attention per bar.

Volume  

Closes after X volume traded. 

Busy market = fast candles, quiet market= slow candles. 

Better than tick because 1 contract ≠ 500 contracts.

Delta 

 Closes when buyers vs sellers imbalance hits target delta. 

Pure aggression signal, great for momentum analysis.

RANGE-BASED

Range (HLR) 

Closes when price range (high minus low) reaches X ticks. 

Pure volatility measure.

RangeY (OCR) 

Closes when price moves X ticks from open. 

Catches directional moves.

Point & Figure

Classic reversal chart with orderflow data. 

Ignores time, shows only significant moves. 

1st param = box size, 2nd = reversal boxes (1-5). 

OFA style = 1 box, 

Classic = 3 box.

Trend-Reverse (TREV)

Asymmetric thresholds. Easy to continue trend, harder to reverse. 

1st param = continuation ticks, 2nd = reversal ticks. 

Typical ratio ~0.68

 (e.g., 10/15 or 20/30).

RENKO

True Orderflow Renko

Fixed-size bricks, new brick when price moves X ticks from previous close. 

2nd param = reversal boxes (Classic = 2 box). 

During high volatility creates many bricks rapidly

Use Normalize option to rate-limit if needed.

IMBALANCE BARS 

From López de Prado's "Advances in Financial Machine Learning" (2018), Chapter 2, pg 29-30.

Core idea: sample MORE during informed trading, LESS during noise.

Tick Imbalance (TIB)

Closes when |buyTrades - sellTrades| hits threshold. 

Random market has buys ≈ sells.

When imbalance builds, thats informed flow → new candle. 

Aggressive buying = many bars, sideways chop = almost none. 

Detects institutional accumulation/distribution.

Exocharts deploys Constant Imbalance Bars - a variant, which is actually preferred by many practitioners. 

The official definition by López de Prado states that the expected candle size should be calculated as an EWMA of T values of previous bars. 

However, in my experience and like other people, the sizes of the bars end up exploding (very big sizes of thousands of ticks) after a few itterations. 

The dynamic EWMA version has a known feedback loop problem:

threshold grows → more ticks needed → threshold grows more → infinity. 

Two advantages of the Exocharts version: it use actual trade direction (isBuy) not inferred tick rule - more accurate threshold avoids the explosion problem.
VIB is the same logic - just weighted by volume instead of tick count. 

So Exocharts implementation follows the core concept.

ENTROPY BARS (Exotica)

Based on Shannon's information theory (1948). 

Closes when "enough new information" arrives.

All volume concentrated at one price = low entropy = no new information = candle stays open.

Volume spread across many prices = high entropy = new information arrived = sample new bar.

Measures actual information content, not just activity.

Setting is Entropy × 100:
- 150 = 1.5 bits (~3 active price levels)
- 250 = 2.5 bits (~6 price levels)
- 350 = 3.5 bits (~12 price levels)

No other platform has this. Unique feature of Exocharts orderflow platform.

Ref: Shternshis et al. (2022) "Measuring market efficiency: The Shannon entropy of high-frequency financial time series"

Normalization Option

Available for non-time candles. Sets minimum candle lifetime in seconds.

Prevents flash crashes from creating 50 candles in 1 second. Even if threshold hit instantly, candle wont close until X seconds pass.

Example: Delta 500 + Normalize 3s = requires both delta threshold AND 3 seconds minimum.

Quick Summary

"Time bars should not be used because markets do not process information at a constant rate." 

- López de Prado, AFML