It’d be great if you could propose a means by which two people can be asked to “come to an agreement” about the score, that is not susceptible to one person in bad faith disagreeing with what the actual score should be. It would be interesting to see if this is “just better UX” or a whole different process (for example, we just set the score to be what the AI says it is after both players pass, removing the need for the players to actually assess their position at the end).
" One person in bad faith disagreeing with what the actual score should be" is what “score cheating” is, and it happens a lot (10-15 reports per day).
The cases like the OPs where “something seems to have gone wrong in the system” are rare, which is perhaps why up till now it’s been rare that folk receiving the unduly harsh message have not complained much. Because many folk receiving that message were cheating and they know it.
If we knew a better way to handle it, it’d at least be “on the list” to implement. Right now I don’t think there is a proposal with legs…