In 1975, E. R. Robinson conducted the hallmark study of the behavior of classical nova light curves before eruption, and this work has now become part of the standard knowledge of novae. He made three points: 5 out of 11 novae showed pre-eruption rises in the years before eruption, one nova (V446 Her) showed drastic changes in the variability across eruptions, and all but one of the novae (excepting BT Mon) have the same quiescent magnitudes before and after the outburst. This work has not been tested since it came out. We have now tested these results by going back to the original archival photographic plates and measuring large numbers of pre-eruption magnitudes for many novae using comparison stars on a modern magnitude scale. We find in particular that four out of five claimed pre-eruption rises are due to simple mistakes in the old literature, that V446 Her has the same amplitude of variations across its 1960 eruption, and that BT Mon has essentially unchanged brightness across its 1939 eruption. Out of 22 nova eruptions, we find two confirmed cases of significant pre-eruption rises (for V533 Her and V1500 Cyg), while T CrB has a deep pre-eruption dip. These events are a challenge to theorists. We find no significant cases of changes in variability across 27 nova eruptions beyond what is expected due to the usual fluctuations seen in novae away from eruptions. For 30 classical novae plus 19 eruptions from 6 recurrent novae, we find that the average change in magnitude from before the eruption to long after the eruption is 0.0 mag. However, we do find five novae (V723 Cas, V1500 Cyg, V1974 Cyg, V4633 Sgr, and RW UMi) that have significantly large changes, in that the post-eruption quiescent brightness level is over ten times brighter than the pre-eruption level. These large post-eruption brightenings are another challenge to theorists.