no, it's all the games when he was able to close out the set. 2:0 up in a bo5, he had to lose 3 matchpoints. it is NOT decider matches.Match point in that post is clearly defined. It is the decider match of a series when both players are tied and the series has went to the last game.
rest of your post is correct anyway, but to be scientifically correct, you'd need to do much more calculations because the 16 (or rather 14) games are cherry-picked. you could arguably exclude the Jordan games and would be down to 12. then you may include more games from the past when he was for example able to beat Viper on a decider in Clown Cup. there also might be some contributing factors such as fatique or civ picking approach. still, it is a pretty crazy number which probably will hold to be significant under all realistic assumptions.
the spectacular thing here was that in the non-matchpoint games, he even won comfortably over 50% against the same opponentsWell, if I understand correctly, the opponents he faced in those 16 match points are Viper (x4), Liereyy (x3), Hera (x3), Jordan (x2), Yo (x2), Saymyname (x1) and Villese (x1).
I understand your argument that Tatoh got to a match point against them in each case, but it still seems really difficult to give anyone a 50% chance on average of winning those 16 matches.
In all of his 25 non-matchpoint-games, he won 15:10 games.
In his 14 matchpoint-games he lost 14 times. 0:14.