UPDATE: I just realized that parts of my criticism are already addressed here: http://fivethirtyeight.com/datalab/steals-are-predictive-but-are-they-that-important/ . In my opinion the metric that Benjamin Morris uses there is the way more important one (predictive value) - but the result would have destroyed the superlative that his main article produces. 'Steals are super predictive' is more interesting than 'Steals are a bit predictive, around as predictive as Rebounds or assists'. But back to my article...
I read the 'The hidden value of the NBA steal' article yesterday and read the same day something about '...Can you really retire as the all-time leader in a statistical category (which is now gaining more favor among stat-heads)...', where they spoke about steals. I am not 100 percent sure what precisely Benjamin Morris did in his analysis, but let me quickly summarize some things that can be fishy about it.
1. Just showing some absolute values
Morris states that in terms of predicting a players value, a steal is worth 9.1 points (and a block for example 1.7 points). He never mentions things like correlation and significance. An absolute value can be very high, but if your data is all over the place, the predictive value is not really given. Morris says 'A player that averages a steal more is worth 9.1 points'. But the second part of the sentence easily could be 'This prediction can be off by plus minus 20 points'.
2. Ignoring that most of those values are correlated to each other
You play more minutes, you produce more everything. For example, assists, steals and turnovers are highly correlated.
3. (the combination of 1 and 2). Ignoring the number of occurrences
So, to elaborate: He just shows some absolute values that basically state something along the line:
'If Player A has 2 steals per game and Player B 1 steal, Player A's team will score 9 points more'. But, as I mentioned before in 2, all those stats are correlated and in 1, he doesn't show the real prediction value of a steal, but the absolute number. When I take the averages of Rebounds, Assists, Blocks, Turnovers and Steals that players produce and multiply them with his numbers, I get
As you can see, the steal is somewhere in the middle of the pack.
So to summarize: I cannot reproduce precisely every step that Morris did in his analysis. But I am pretty sure that there are some values that are far from significant. Players that play more, produce more stats. The lines between correlation and causality become very blurred. Best case scenario for the analysis: Players that produce a lot of steals are in general more valuable. Worst case scenario: The analysis is complete bubkes.
I hope the article is halfway coherent. (It's written at 8 in the morning before I leave for the weekend). As always, let me know what you think, my dear imaginary reader.
Have a nice weekend,
P.S.: I hope that nobody believes that a steal is REALLY worth 9.1 points...