BK is talking nonsense.

The only point he may have (but didnt make) is that a higher winrate reduces your downswing risk. Note, this isnt the same as lower variance. Put one way, an 8 ptbb winner will have fewer and smaller slumps than a 1 ptbb winner if their variance numbers are comparable because the intervals established by his variance/std deviation do not enter the negative zone as prominently.

In this particular argument though, i think there's a consensus that the more aggro tags achieve higher winrates but with correspondingly higher variance. You could do a survey and the math to see if the higher risk/reward is worth it.