numTrials for getApproxHint, scaled with the SortedTroves size (the Liquity
heuristic ceil(15·√size), clamped to [15, 2500]). A flat 15 oversamples a
tiny list and undersamples a large one; more trials → a closer approximate hint
→ fewer on-chain traversal steps at insert (cheaper gas), at the cost of more
view work. The contract always places correctly regardless, trials only buy gas.
numTrials for
getApproxHint, scaled with the SortedTroves size (the Liquity heuristicceil(15·√size), clamped to[15, 2500]). A flat 15 oversamples a tiny list and undersamples a large one; more trials → a closer approximate hint → fewer on-chain traversal steps at insert (cheaper gas), at the cost of more view work. The contract always places correctly regardless, trials only buy gas.