I imported a game a bit earlier, but I wasn’t ready to review it yet. I went over now and it looks like this
I’ve imported it from the same client as another 5 games, with the same way and consistent file names, but all the other games are OK.
It’s not that dead stones appear faint, it looks random.
It seems that the client used “territory white” and “territory black” tags to mark territory, but it appears that not all of it was marked properly. Since not all of the territory is marked properly with these tags (which may be a problem with the client or a mistaken scoring issue), I think this might be confusing OGS somehow.
I “cleaned” the SGF file from your game by downloading the SGF and reprocessing it through another editor. Here is the “cleaned” file with the territory tags removed:
cleaned.sgf (2.9 KB)
I also reuploaded it here:
Now, things seem to be working normally.
I only posted the screenshot and not the link 
Thanks for looking into this.
It’s weird because it went to scoring, I don’t know if sgf stores score as well. 
At least I marked it correctly in the original (I mean the actual game), B+110,5.
SE gives 113,50 but I don’t see why and I’m going to ignore it.
Yes, the result B+110.50 was stored in your original SGF file and preserved in the clean version that I uploaded as well.
Actually, it’s possible to set that arbitrarily and OGS will just display it
It seems that this setting causes the OGS scoring computation to get confused
This came to stay.
The score estimator (not KataGo, but the button on the side menu) uses Chinese rules, whereas your game was scored with Japanese rules. Since there were some extraneous passes by white, that will increase the discrepancy between Japanese and Chinese rules.
Note that filling dame at B6 should affects the score under both Japanese and Chinese rules, since Black playing there would require White to protect at A5. Hence, KataGo also disagrees with the final scoring.
Of course, I understand that you may have just wanted to avoid dragging it out any longer.
