Following your logic: Until the the pilot decided to climb ("over a layer"), he had two options at that time: 1) find a place and , in the vicinity, land or 2) or revert course and find a place to land. Climbing his options would have increased (as would the work load): 1) Flight to destination , VFR on top a layers, and hope for a hole; 2) or IFR approach to an airport; 3) or Precautionary landing on some ridge top, 3) or Declaration of an emergency. Note: More information is required to determine if fuel would have been an issue and 4) "pop up' check things out and make a determination from there (and continue flight or go back down and make a new decision). As you pointed out, what doesn't make sense is the climb to 2300' and then a sudden turn (especially since that it seems the ROC to get there was about 900 FPM) (10 seconds and then in the 'clear') Presumably, that climb was with poor forward visibility and use of instruments (autopilot?) must have been in use. If there were no other, casues, then, the post about the left turn crashes after VFR takeoff into IMC in the sim make sense, if there was a decision to go back down and in those moments there was a loss of spatial orientation (since the flight path obviously doesn't indicate a controlled descent). As you noted, the weather conditions and options) would have been recognized (presumably) by a seasoned pilot and the best decision would have been not to go in the first place.