Being not useful means detrimental too or just irrelevant? Because... I would always accept the slightest increase of odds.
What are the numbers for mulligan?
It is better than nothing, of course. However, the best increase is ~3.45%, which is less than 4 chances in 100,
with the best possible conditions. It's not impossible, it's just terribly underwhelming. Mulligan is not easily mathematized, but with some assumptions...
Supposing a certain chance (0.40) of understanding that a match is lost by starting hand, mulligan kicks in when you:
1) Have a game when you mulligan, two games when you win and two games when your opponent wins, in any order.
2) Win the last game.
With these numbers, mulligan has a max effect of ~18.75%, and beats blitzkrieg starting from ~8% winrate onwards. If you deck has 1 chance in 10 to win and the person playing has 4 chances in 10 to understand when a game is lost by the starting hand only, then blitzkrieg is useless.
I can run the spreadsheet with a lower chance to understanding when to mulligan if you wish...
