The blank screen does not make the change hard to see. It removes the motion transient — a brief luminance signal at the exact location of the change — that would otherwise fire automatically in your peripheral visual system and redirect your attention. Without it, you need attention to do the work the transient was doing. The blank screen exposes the ordinary mechanism by temporarily removing it.
The simulation can show that detection slows dramatically in blank-flash mode. It cannot show what's absent: the visual system does not maintain a detailed internal record of the scene. The stable, coherent world you experience is not a buffered copy of the image — it's a model built on demand, refreshed by attention, and filled in by assumption. The blank flash reveals this by creating a gap where the assumption fails. In the direct-cut mode, you never encounter the gap; the motion transient closes it before you know it was there.
The distractor task approximates what Rensink called marginal interest — change blindness is strongest for objects outside your current focus. The task moves your focus elsewhere. But the simulation can't replicate the attentional resource constraint that makes this structural rather than a performance trade-off: you cannot choose to attend everywhere at once, even without a task.
The underlying debate (O'Regan & Noë 2001): change blindness suggests the visual world isn't internally represented in detail, that vision is for acting, not for storing. Against this: Lamme (2003) argues that vivid phenomenal experience exists without reportability — the representation might be there, just inaccessible to the reporting system. The experiment doesn't decide.