I did a rollout which stopped after 144 trials as the difference between the plays was lage enough. I realized I wanted to see the results for the full 1296 games and went and unchecked the stop when difference is large enough option. When the roullout resumed it continued from where it left off. The time elapsed started fresh from 0 seconds. The calculation for the time remaining was incorrect however. It was figuring that the time elapsed was the time needed to get through the original 144 games plus the additional ones since then, whereas it should have based the time per game on the elapsed time to get through the additional games.
For example: Suppose when resuming that 6 games took 150 seconds to play. Thus on average 25 seconds per game and the estimate for time remaining should be (25sceonds)(1296-150). The calculation it actually was doing was figuring that the 6 seconds was for all the 150 games. Thus is estimated (1sec)(1296-150).