It's worse than that. The historical scores for what are now FL are probably wrong at present. They should be correct soon. Correct in this case means old value up to 1/2/2017, new value after. I may rewrite history and make them the new bigger value back in time. I don't know yet. This points error arose because of a "send three and fourpence, we're going to a dance" comms failure.
Yes, it changes the past. Yes, some people may have fewer chases in a specific region that changes the status of some awards. But we wont be removing awards, they were issue against valid data at the time. As for uniques, you should not have lost or gained any. The summit reference may have changed from F/VO-001 to FL/VO-001 and the points may have changed. But the database knows that it is the same summit it always was.
Splitting was done to better accommodate the French geography. It's nigh on impossible to get a decent points banding when you have some lovely wee summits and Mont Blanc in the same country. The split allows better banding for the places with plenty of lower summits. At the same time of the split there has been some determined validation of that the summits were really the right prominence and these summits have been deleted. New summits have been found and added. I'm not sure of the overall change in numbers at this point.
Splitting will throw up corner-cases. Deleting all the summits from one association and creating new summits creates bogus uniques. If you have climbed up F/VO-001 you have climbed it whether its reference changes or not. Swings and roundabouts I think between bogus uniques or some issues on displaying awards.
There is no bulk way to collect the old values for points as the summits list shows today's values. However, every summit is in the database still, basically so you can submit a log for a QSO in 2005 say for someone on a summit deleted in 2008. You can use the find summit feature to show every summit there has ever been and if the score has changed you should see the a table showing the score for various dates.
I should say that I try to mail yourself and Rob DM1CM when there are new associations coming so you get a heads up in case it breaks your code for the gpx files you make available or Rob's web site but it was missed off my list this time.
I'll have a think about making a version of the summits list available including all the points.