Because we already know what causes most statistical variance - good ol' luck, and positive or negative regression to the mean.
Let's take an example.
Kilmarnock 2-1 Rangers, first league game after the winter break. The narrative was that Rangers bottled it because we couldn't handle the pressure of being top and we didn't have the bottle to play consistently.
This accusation followed Rangers 0-1 Aberdeen last December, when we were also top but then slipped to second following this defeat.
BUT... in actual fact, we played pretty well in both games. It wasn't a mentality issue - we turned up and got to work and did the same things we do in every game. The only difference was the outcome, and that's down to statistical variance.
From ModernFitba:
Rangers - Aberdeen:
Goals:
0 - 1
Shots:
14 - 2
Shots on Target:
5 - 2
Shots inside box:
9 - 1
Possession:
71 - 29
Our xG that day was about 2 - in other words, given the chances we created we could typically expect to score 2 goals in that game. It just so happened that we got unlucky and it never fell right for us. If you replayed that game 1000 times, we would take 2.60 points from it on average.
Killie was even unluckier. Again, from MF:
Kilmarnock - Rangers:
Goals:
2 - 1
Shots:
8 - 19
Shots on Target:
3 - 8
Shots inside box:
4 - 12
Possession:
37 - 63
For this match, we had an xG of 3.91, so we could expect to score four goals in this game on average. Again, replay this game 1000 times, we'd take 2.91 points from it on average, so the fact we got none shows that it was an extreme outlier.
So what's changed this season?
Last year, we underperformed our xG - scored less goals than we would expect to. Therefore, we got some slightly unlucky results.
This year, we're overperforming our xG - so we're scoring more goals than you might expect to given our chances.
Has bottle caused this change? No - our luck has simply evened out, and we're positively regressing towards the mean.
At our level, bottle pretty much isn't a thing.
EDIT: source -
https://www.modernfitba.com/blogs/2019/4/14/expected-points