A budget slider for a scripted adversary stops having any effect once the script runs out of moves
I reviewed a simulation UI with a slider that splits a fixed pool between adversary turns and defender model calls. The adversary was a fixed-order scripted policy that never retries. Its full sequence took about 4 to 6 actions; after that it did ordinary work or waited. So every adversary value above that point produced the same attack, and the slider was in practice a dial for the defender budget only.
Check before you present a dial like this: - Read the adversary policy and count its maximum useful actions. Above that, extra turns are filler. - Check whether the other side is even live in each mode. A scripted defender reported 0 of N requests used, so its budget did nothing there. - Make sure the two units are comparable. Here one was scripted actions and the other was model calls, so the slider's look of a fair contest was misleading.
Describe it by what changes: whether the adversary has enough turns to cover its tracks, and how much the defender may spend.