It's accepted because you get as many smurfs on your team as the enemy team if you play a large amount of games
Edit for the 4/9 and 5/9 chance people that might not see my answer to this point:
Given you are not the smurf and there is exactly one smurf in the game, that's absolutely correct
I think you'd have to play an unplayable amount of games to show there's a significant difference, though, because the probability of a smurf filling one of those slots is already low
Say on average one in 5 games that you don't play has exactly one smurf in, that's 1/50 slots, 2% chance of having a smurf in a slot
Chance of a smurf in x of your slots is
4Cx×(0.02)^x×(0.98)^4-x
x=0, p=92.237%
x=1, p=7.529%
x=2, p=0.230%
x=3, p=0.00314%
x=4, p=0.000016%
Chance of smurf in y of enemy slots is
5Cy×(0.02)^y×(0.98)^5-y
y=0, p=90.392%
y=1, p=9.224%
y=2, p=0.376%
y=3, p=0.00768%
y=4, p=0.0000784%
y=5, p=0.00000032%
Say you play 300 games, expect 23 games have one smurf on your team, 28 have one smurf on enemy team
5 games in 300, net, lost to smurfs, is close enough to 0 for me