Comparing Revisit Puzzles with their Original Puzzles
|Opened on:||Monday, September 29, 2014 - 22:12|
|Last modified:||Wednesday, July 1, 2015 - 17:37|
If a Revisit Puzzle (like #978) is based on a native (known) structure, one way to compare its results with those from its Original Puzzle (like #134) is as below:
For each puzzle, find some measure of similarity (like the alpha-carbon RMSD value) between the solution with the best Foldit score and the native (known) structure. Also, from all solutions for that puzzle, find the solution with the best RMSD value and note this RMSD value as well as this solution's Foldit score. Ideally, the solution with the best RMSD value should also have the best Foldit score.
If the Foldit scoring function and game tools have improved between the Original Puzzle and the Revisit Puzzle, you'd expect the Revisit Puzzle's RMSD value for its best Foldit score to have improved. You'd also expect the Revisit Puzzle's best RMSD value to have improved. You'd also expect the solution with the best RMSD value to be more like the solution with the best Foldit score, and this similarity could be measured for both puzzles as well.
I would be interested to see results like these posted as comments under each Revisit Puzzle after the Puzzle finishes.