Trace the memory movement active parameters leave out.
Model the resident set, replay routed experts, and compare reactive LRU against an oracle lookahead bound. Every assumption stays visible.
Same 16 activations. Different movement.
REACTIVE LRU
12 missesEvicts the least recently used expert with no knowledge of upcoming routes.
ORACLE LOOKAHEAD
10 missesEvicts the resident expert needed farthest in the future: an upper bound, not a predictor claim.
Interpretation. Active weights describe one instant, not the path. Shared layers and KV state consume the fixed budget; route transitions determine misses; bandwidth converts movement into a transfer-only stall estimate. Compute overlap, page faults, quantization metadata, attention compute, and predictor error remain outside this model.
The populated export preserves assumptions, both policy summaries, and every hit/miss event.