my study contradicted a few of the items above, but im not sure of the sample size used to compile the above stats.
The thing to remember about stats is that they are relative. The above article sort of indicates that there is a fixed relationship between the actual number, and the outcome, which is a contrary to correlation coefficients, which is what I produced.
Correlation is the relationship between two numbers that represent different things.
X represents Box Score Assists, Y represents Passing skill.
Over a sample area, the correlation is representation by the relationship between a lower X value and a lower Y value.
HOWEVER, this does NOT mean 1 assist = Atrocious passing.
What it means, is that if you have two players, Player 1 and Player 2, and Player 1 assisted 1 time, and Player 2 assisted 2 times, and the correlation between assists and passing is high, then there is a high likelihood that Player 2 will have higher passing, than player 1.
The correlations do not suggest that Player 1 has Atrocious passing and Player 2 has pitiful Passing. Merely the relative relationship between the two.
In the end, my advice is to use box score as differentiators between two players, but in a 'gambling' context. You are gambling that that box score indicates the skill is higher/lower. But its educated gambling.