All systems operational· Status page →
GMJD Docs

Step 2 · Compatibility

Holding rules overnight and over weekends, the stage change between evaluation and funded account, and the noise band.

Nine strategies against sixty models. The question is whether the holding rules overnight and over the weekend fit the strategy's trading hours.

Five answers

The matrix distinguishes five states: works, haircut, better without, turns negative and does not work. The line between "haircut" and "does not work" sits at a chosen share of remaining return. That is a setting, not a measurement.

The stage change costs money

On some models, holding overnight is permitted during evaluation and not on the funded account. Anyone who passes with such a strategy has paid and then cannot trade it.

Those models therefore sit before the matrix and not as a footnote below it: in the table's shorthand the case is invisible.

Models with no statement from the source

Some models say nothing about holding overnight. That is shown as "not verifiable" and explicitly not as "no restriction". It is a statement about the source, not about the product.

Distinct from that is a model whose source contradicts itself. That is a finding about the product and not a gap in the collection, and it is therefore not hidden along with the others.

Distinguishable answers

Above the rows sits the number of distinguishable answers for your selection. That number is a property of your selection and not of the rules: models differ only as far as the chosen strategies separate them. With a pure intraday selection everything collapses into a single answer.

The noise band

The matrix says how much return a holding ban costs. It does not say whether that share depends on multi-day holding at all.

Draw the same number of trades at random from the same strategy and you also get a return share, and it varies. The band in the chart is the range produced by 20,000 such draws.

  • Point inside the band. The haircut is indistinguishable from a random selection of the same size. It depends on the number of trades, not on them running overnight.
  • Point outside. The haircut really does depend on multi-day holding.
  • No band. Every trade of the strategy is multi-day. A random selection of the same size is then the same set, and the question does not arise. The point is drawn unfilled.

The third case looks like the second in a two-valued display and means the opposite.