Analysis of light detection and ranging (lidar) derived topography combined with field data enables measurement of small (<30 m), previously unrecognized dextral offsets beneath dense temperate rainforest along New Zealand’s central Alpine fault. Field offset measurements often have lower uncertainties than lidar measurements. Offsets of 7.5 ± 1 m for the most recent earthquake (ca. A.D. 1717) and cumulative offsets of 12.9 ± 2 m and 22 ± 2.7 m can be averaged into three 7.1 ± 2.1 m increments of repeated dextral slip at a point, and when combined with an offset compilation show a uniform slip distribution of ∼7.5 ± 2.0 m over 300 km in A.D. 1717. Comparing these offsets with the 1.1 ka paleoseismic record and slip rate demonstrates a mismatch between offsets, timing, and slip rate that can be explained in two ways: (1) major (full) ruptures (moment magnitude, Mw ≥7.9) every 270 ± 70 yr (e.g., A.D. 1717) and with moderate to large (partial rupture) Alpine fault earthquakes (Mw ≥6.5; e.g., A.D. 1600) occurring between full ruptures, and (2) some off-fault shaking data may instead reflect paleoseismicity from other faults. If explanation 1 is true, the Alpine fault has two (i.e., bimodal) or more modes of behavior (or the slip rate has not been constant since 1.1 ka) and rupture is perhaps width limited. If explanation 2 is true, perhaps the Alpine fault behavior is characteristic, and other faults are responsible for some shaking records. Ultimately, bimodal behavior is our preferred interpretation, which has implications for our understanding of plate boundary seismic hazards worldwide.