A market call and a winning trade are different outcomes

What would a repeatable intraday process need from its market judgments? How much of a correct direction survives the trade's stop, target, and costs?

This study measured the conversion between two different outcomes. A call was correct if price finished on its side after eight hourly bars. A trade succeeded according to the ordered path through its stop, target, or expiry.

Hypothetical entries opened during specified UTC daytime hours. Distances used the prior completed daily ATR. We varied trade geometry, charged spread relative to initial risk, and closed unresolved trades at the end of a 24-bar holding window.

For each entry we measured the result on the hindsight-correct side and the other side. We then mixed those two groups at different assumed call accuracies. That produced a conditional illustration of what accuracy could be worth under this sampling scheme.

We also measured outcome correlations, resolution times, and the maximum number of trades the simulated schedule could accommodate while respecting its group constraints.

A correct terminal direction could lose because the stop came first. An incorrect terminal direction could win because the target was reached before the later reversal. Costs then moved the break-even point further.

Trade geometry changed that conversion. Larger distances reduced spread as a share of the original risk and changed the chance of a call surviving to its outcome. They also generally occupied a trade slot for longer. This was a trade-off between cost, path, and time, rather than a universal best distance.

Trades sharing an underlying currency often had related outcomes. Counting their tickets separately overstated how many different decisions the process was making.

The accuracy ladder is not a forecast of what any trader's accuracy will earn. A real trader may be correct on a different subset of paths from the mixture used here. Their trade selection, timing, costs, and payout distribution can differ too.

The simulated trade counts are capacity ceilings. They assume a suitable fresh read whenever a slot becomes free. An available slot does not establish an opportunity or an advantage. Similarly, sample-size calculations assume sufficiently separate observations; many calls on one theme supply less independent evidence.

The original study used a limited cost model and had shorter hourly histories for gold and NZDUSD. Later research refines both the range unit and the way shared exposures are understood. Its numerical floors should not become trading targets.

The useful accounting decision is to retain the judgment and the trade as separate records. That makes it possible to ask whether the read, expression, execution, or management helped. A single win-rate number cannot answer all four questions.