We assessed accuracy in point fire intervals using a simulation model that sampled four spatially explicit simulated fire histories. These histories varied in fire frequency and size and were simulated on a flat landscape with two forest types (dry versus mesic). We used three sampling designs (random, systematic grids, and stratified). We assessed the sensitivity of estimates of Weibull median probability fire intervals (WMPI) to sampling design and to factors that degrade the fire scar record: failure of a tree to record a fire and loss of fire-scarred trees. Accuracy was affected by all of the factors investigated and generally varied with fire regime type. The maximum error was from degradation of the record, primarily because degradation reduced the number of intervals from which WMPI was estimated. The sampling designs were roughly equal in their ability to capture overall WMPI, regardless of fire regime, but the gridded design yielded more accurate estimates of spatial variation in WMPI. Accuracy in WMPI increased with increasing number of points sampled for all fire regimes and sampling designs, but the number of points needed to obtain accurate estimates was greater for fire regimes with complex spatial patterns of fire intervals than for those with relatively homogeneous patterns.