(Yeah, I was wondering why you replied to newer message lol)(I'm guessing this got stuck in moderation queue or something because this wasn't here when I replied to another response yesterday. Anyway.)
I take your point and get what you're trying to achieve - but a few caveats. To me, both tools are equally wrong if they're 8M out, regardless of positive or negative. If tool2 states 56M as a MAXIMUM value for data, and the explorer decides that's good enough for 50M, well, the explorer is in the wrong as it's clearly a maximum value. Ideally both tools say "at least x, and up to y".
I'm not even sure if the various tools account for things like first discovered / first mapped because it's hard (or sometimes impossible) to infer from Journals, making it kinda moot anyway (I might be wrong on this, I don't use any 3rd party tools myself (except my own))
For your purposes, I'd suggest going back earlier in the thread and finding all the data you can for expected and actual sold values for all terraformables (a few people provided impressive amounts of data), then working out the percentage and using that as your percentage to apply. I'd also suggest separating out HMCs and WWs - I suspect the average percentage for HMCs if higher than that of water worlds, which seem much more volatile.
Yes, they would be both wrong, the difference is that a decision coming from looking at a negative difference is way more impactful than one coming from a positive one.
Giving the "Maximum value" alone could be deceptive in a scenario like this.
For what concerns First Discovery I think that pulling system info from EDSM/EDDB could make it easier to determine it (even if not sure), probably the best way is the other option you said "your scan data worths at least ###cr, maximum projected value: ###cr".
Regarding HMC and WW, the percentage you are talking about is the probability on which the bonus is calculated (i.e. HMC get the bonus more often than WW) or the bonus increment value?