As shared (just now) with scientists on puzzle 1373, a solution with a score of 10690 shows a recentbest score of 10805, but recentbest.Restore does not change the score.
The solution has 100 residues, incurring a penalty of 5 * -23 = -115 points. So the recentbest score is incorrectly ignoring this penalty.
I've also noticed this problem on other design puzzles, including the current 1383, but I haven't been able to reproduce it reliably.
Changed to bug.
Shared a similar case for 1383.
Solution score 10412, residue count penalty 345, recentbest shows as 10757.
The problem seems to go away after a recentbest.Save. Reloading a solution also reloads the incorrect recentbest score.
The same problem can also be seen in recentbest and absolutebest autosaves. These autosaves scores don't reflect the residue count penalty. They do include residue IE and other filters.
The recent best scoring glitch is still happening with puzzle 1404.
Contrary to what I previously said, recentbest.Save does not correct the score of the recent best pose. Recipes which rely on recentbest may not produce the expected results.
Old issue, seems to be gone in 20180725-d14dae34c3-win_x86-devprev, closing.