So, I did some quick data sleuthing with relation to Harden's turnovers.
One of the primary causes of turnovers isn't necessarily passing, but rather, passing
to the rim. The best assists are assists at the rim, because a dunk/layup is the highest expected payoff out of any shot, right? This occurs because they're far more likely to go in than jump shots (an open layup/dunk is practically a sure thing, and this doesn't occur for jump shots).
However, the expected turnovers for an "interior" passer are generally higher. Interior assists are more likely to be contested (whether it's by intentionally defence, or simply by other players being in the way), meaning that:
a) the ball is less likely to hit the intended target
b) there are more likely to be interior assists that don't happen because of fouls, even if they result in team points
So, I created a list of "interior assist" percentages (i.e. assists at the rim/total assists):
Many of the players that rank notably lower (e.g. DeRozan, Irving, Walker, Butler) are players that have traditionally had strong assist/turnover ratios, but are seldom known as being great playmakers, and the interior ratio is a reason why - they're generally making safe passes which won't turn the ball over, but aren't quite as likely to concoct elite scoring opportunities.
Now, this isn't a proxy for being a good/bad passer - to an extent, the assist opportunities that arise stem from your teammates, which is why some players can vary greatly from year to year in this ratio. For example, LeBron in Cleveland (surrounded by 3 point shooters) had a fairly low ratio, but he has a very high ratio in LA, notable for their lack of spacing. But it does also explain a bit of the assist/turnover ratio, and how it sometimes differentiates from our perceptions of passing ability, and how the same player can vary from year to year. For example, Garnett's
interior assist/turnover ratio was very similar throughout most of the 00s, but his raw assist numbers dropped off quite a bit, because he was no longer providing as many "risk-free, but low-leverage" passes to the jump shooters on his teams. Jason Kidd's A/T ratio increased in his later years of the Nets, because he was servicing jump shooters more, rather than the Kenyon Martins of the world.
Anyway...
Out of anybody I've acquired the numbers for, Harden ranks 1st in the league. This doesn't mean that he's not turnover prone, but it shows that his passes that
aren't turnovers are creating high-leverage opportunities that raw assist and turnover counts aren't really capturing. The "at the rim" frequency without Harden on the court is 27.82% for Houston, and it jumps up to 33.48% with Harden on the court, even though Harden only shoots 28.23% of his shots at the rim (and he is something like 40% of Houston's usage...)
However, with Harden on the court, the percentage of 3 point attempts by his team also stays at roughly 51-52%, so he's replacing midrange shots with "at the rim" shots. And I don't think I need to get into Harden's foul drawing.
In other words, Harden is not only a master of the 'Morey index' (i.e. at the rim/3 point shots), but he's also terrific at generating these shots for his teammates too. The principle behind spacing is that occupying certain zones of the court helps open up other areas of the court - become a 3 point threat, and it's easier to generate shots underneath the hoop. Harden is a player who can not only occupy both the rim AND the 3 point line, but he actually creates MORE of these high-leverage shots for his teammates with him on the court!
And that's why I find his turnovers forgivable - sure, he turns the ball over a lot, but that's a byproduct of being arguably
the best generator of the Morey index principle in the league. And that's precisely why, despite his turnovers, he seems to constantly lead strong offences.