Defaults to interrogate: do market orders fill at the bar close or at the far side of the spread; do limit orders fill whenever the price merely touches the level; is the full size always available; do stops fill at the stop price or at the next traded price after it.
The limit-order assumption is the worst offender. Being touched is not being filled, because you sit in a queue behind everyone who was already resting there. A realistic rule is to require the price to trade through your level, not merely to it, and even then to assume partial fills in thin names.
Worked example: a strategy assuming limit fills on touch shows 2,400 trades and a 1.4 profit factor. Requiring a trade-through drops it to 1,500 trades and 1.05. Nothing about the idea changed; one unexamined assumption held the result up.
Related: slippage-model, next-bar-execution, quote-data, event-driven-backtest