A full-size contract is ten micros. The results are not.

A full-size futures contract is worth exactly ten of its micro. ES is ten MES, NQ is ten MNQ, GC is ten MGC. Identical instrument, identical hours, identical tick behaviour, ten times the money per point.

So it is tempting — and common — to run one backtest on the micro and publish the full-size figure as that result times ten. It is wrong, and we can show you by how much.

The engine does not see the same market

Two things differ between the contracts even though the underlying does not:

Order flow and cumulative delta are not shared. The micro book and the full-size book are different books with different participants. A strategy reading volume or delta is reading different data.

The price histories carry a persistent offset. Not noise — a systematic gap of a point or two, sometimes more, which steps as contracts roll. A level broken by one and a half points on one contract is a level held on the other.

How much it matters depends on what triggers your strategy

This is the part we did not expect, and it is the useful finding.

The logic is clean once you see it. A pattern — a shape in the bars — survives a small price offset. A level does not: the whole trigger is a threshold being crossed, and shifting every price by a point and a half changes which crossings happen at all.

So if your strategy fires on breaks of a level, a session range, a prior day's high — expect real divergence between contracts. If it fires on a pattern or an indicator relationship, expect very little.

Which way does it go?

Both ways. A full-size run can finish above or below ten times its micro counterpart, and ours have done both. On the trades the two runs genuinely share, the P&L scales at slightly better than ten times — because commission does not scale. Trading one full-size contract costs meaningfully less in fees than trading ten micros.

But that is a statement about shared trades. It says nothing about the trades one run took and the other did not, which is where the divergence actually lives.

What we do instead

We run the backtest again on the real contract and publish what it returns. Every full-size figure on this site is the actual output of the actual backtest on the actual contract named beside it. Neither figure is presented as a proxy for the other, because neither predicts the other.

It would be easier to multiply. It would also be the kind of claim that is impossible for a buyer to check and convenient for us — which is a good working definition of the thing not to do.

Trade-sharing counts are measured on our own published backtests over the published window, comparing each engine's micro run against its full-size run on the equivalent contract.