Fig. 8: Results of simulations with alternate assumptions indicate that optimal difficulty threshold may vary.
From: Optimizing practice scheduling requires quantitative tracking of individual item performance
The original simulation (in black) as well as two additional simulations in which the efficiency of failures is increased. An alternative simulation in red shows performance if feedback time cost for failures was set to 0 s (instead of 4). Another simulation in blue shows performance if failures provided 25% more learning gains than successes (but time cost was same as original). Optimal threshold for paired associate learning remains high for all of them, but the relative benefits over lower OETs is clearly influenced by gains from failures and time cost. S15 (repetitions every 15 trials, trial duration not fixed) was also simulated under the three sets of assumptions and serves as a point of comparison.
