Skip to content

Why NQ imbalances do not read like ES imbalances

By Author of ORDER FLOW — Market Mechanics

The 3:1 imbalance convention was not designed for one contract in particular. Applied to NQ and to ES, it measures two different things.

Most traders first notice this as a hunch: NQ throws more signals, the readings are noisier, what works on ES needs adjusting. The reasons are structural and specific, and knowing them precisely beats compensating by instinct.

The same tick is not the same increment

Both contracts move in 0.25 index points. That sounds like a shared unit, but NQ has traded at a multiple of the ES index level for years, so the same 0.25 increment is a much smaller fraction of price on NQ.

The same percentage move · 0.25 point tick on both ES lower index level ~5 levels NQ higher index level ~18 levels Same tick increment, different price. Aggression that crosses one ES level is spread across several on NQ.
Schematic of how far the same percentage move travels in ticks on each contract. Tick counts are illustrative.

The practical consequence is that a move of equivalent economic size crosses a handful of price levels on ES and several times as many on NQ. The ladder is finer, so the same behaviour gets chopped into more rows.

Fewer contracts, spread across more levels

The second factor compounds the first. ES trades on the order of two to three times the daily contract volume of NQ, and its resting book is correspondingly deeper. So NQ starts with less volume and then distributes it across more price levels.

The same aggressive volume, distributed ES · 5 levels ≈ 1,000 per level NQ · 18 levels ≈ 280 per level Imbalance ratios are computed per level. On NQ they are computed on a much smaller sample.
Schematic of the same aggressive volume distributed across the levels each contract spans. Values are illustrative.

The imbalance ratio is computed per level. It compares the ask volume at one level with the bid volume at the level below. When the per-level figures are large, that comparison is built on a meaningful sample. When they are small, the same arithmetic is being performed on numbers that can swing on a handful of contracts.

A 400-against-120 imbalance and a 12-against-3 imbalance both clear 3:1. They are not comparable evidence, and on NQ the second kind is far more common than most people reading a ratio realise.

Stacking happens mechanically on NQ

Stacked imbalances get their authority from being rare. Three consecutive levels imbalanced in the same direction is supposed to indicate pressure that sustained itself while price moved through a range.

On a thin book a stack can mean much less. When resting liquidity is sparse, price traverses several levels without much trading at any of them, and each of those levels records a small number on one side and an even smaller number on the other.

Thin stack Bid Ask 3 11 2 9 2 18 3 5 all three diagonals clear 3:1 45 contracts across the three levels Stack with size Bid Ask 38 160 25 124 41 242 30 60 all three diagonals clear 3:1 630 contracts across the three levels The ratio test passes in both. On NQ the first kind is common and says little about sustained pressure.
Schematic of two three-level stacks that both satisfy the ratio test. Values are illustrative.

Both stacks pass the test. Counting every cell across the three stacked levels, one represents 630 contracts traded and the other 45. The pattern-matching that makes stacked imbalances useful on ES will happily fire on the second one, and on NQ the second one occurs many times a session.

NQ's book also thins asymmetrically through the day. The stack that means something during the cash open and the stack that means nothing during the midday lull look identical in the record, because the record does not include the depth that was available.

Volume is only the first filter. A stack with real size behind it still says nothing on its own about whether it starts a move or ends one.

What to change, and what not to

The common advice is to raise the ratio on NQ: use 4:1 or 5:1 instead of 3:1. That helps, but it treats a volume problem with a ratio adjustment, and a ratio cannot distinguish 12-against-3 from 400-against-120 no matter where the threshold sits. Raising it to 5:1 simply means accepting 15-against-3 instead.

The change that addresses the volume difference is an absolute volume floor underneath the ratio: a minimum number of contracts at a level before the ratio is evaluated at all. That filter does on NQ what the deeper book does automatically on ES.

Two things do not change. The comparison is still diagonal rather than horizontal, for the same microstructural reason on both contracts. And the threshold, whatever value you pick and whatever floor you pair it with, remains a setting, like the swing threshold that decides whether a delta divergence exists at all. Knowing why NQ differs stops you treating the number as if it came from the exchange.

One further caution specific to index futures. A meaningful share of NQ flow is hedging and arbitrage against ES, against options positions, or against the underlying basket. That flow shows up as directional in the order flow record while expressing no view on price, which is one more reason to treat an isolated imbalance on NQ as a prompt to look.