Presumably, the idea behind fastbreak is to beat the defense down the floor. It seems this idea would also apply to double-teams. Any insight on how double teams work against a fastbreak team? Are they only able to be executed if the offense doesn't get a fastbreak shot off and has to set up on offense?

A double team functions mostly the same against FB as with the other offenses, I think. In my opinion, the primary function is to decrease the effectiveness of the player being doubled, and the secondary function is to somewhat limit their looks. A few folks will tell you I have this backward, and I think this is based on some old dev chats and such, but I think the way most folks have their distributions set up most of the time, counting on a double to really limit a primary scorer‘s looks (if that is who you’re doubling) is kind of wishful thinking. And anyway, you can’t control how your opponent sets their distribution. What you can do is try to disrupt what your opponent is doing.

I will say double teaming is a little less effective against balanced scoring teams. Many FB teams are set up to be fairly balanced (as are motion teams, quite often). So if there is a difference in outcome, that’s probably what we’re looking at, IMO.

Ok, maybe my conception is flawed but I am thinking that the simulation engine, at the start of each possession by some metric, determines whether the possession is a fastbreak possession or a normal one. If it is a fastbreak possession, do you think that the double team comes into play?

Or is it a different mechanism, where at the start of the game (if set on 'always') a player is assigned a double team modifier that affects them throughout the game, somewhat independent of possessions?

I think it’s closer to the latter. The double team doesn’t work on every possession, sometimes it just whiffs, but I don’t think the engine is intelligent enough to differentiate between a true breakaway possession in a FB context and a set FB possession (for instance after an offensive rebound, in which it essentially operates like a motion set), any more than it would off a breakaway steal in any other set (hypothetically some of those should happen, too).

That is a really interesting point. I feel like I do see a steal lead straight to a lay-up in the play-by-play quite frequently, almost always 1-2 seconds after the steal happens and always either scored or assisted by the player who stole the ball. Do you think that is cosmetic and not actually indicative of what the engine did?

Second, 1. Do we know when in a possession it is determined if the double team whiffs? 2. Do we know what determines if a double team whiffs (presumably some of the two players doing the double team's ratings/IQ vs some of the player being doubleteamed's ratings/IQ)?

Thank you shoe for answering all these questions.

Posted by bpielcmc on 5/4/2021 5:12:00 PM (view original):
That is a really interesting point. I feel like I do see a steal lead straight to a lay-up in the play-by-play quite frequently, almost always 1-2 seconds after the steal happens and always either scored or assisted by the player who stole the ball. Do you think that is cosmetic and not actually indicative of what the engine did?

Second, 1. Do we know when in a possession it is determined if the double team whiffs? 2. Do we know what determines if a double team whiffs (presumably some of the two players doing the double team's ratings/IQ vs some of the player being doubleteamed's ratings/IQ)?

Thank you shoe for answering all these questions.

Yeah, I think we see those results with all offenses, though. That’s what I mean when I say the engine doesn’t really differentiate between FB and other offenses in that regard. The big discrepancy in how FB works is 1) how many possessions you naturally see without forcing the tempo, and 2) how much stamina is required. So it is what the engine did, it’s just not a FB offense specific outcome, if that makes sense. I mean maybe it takes 4-5 seconds instead of 1-2 when it’s flex instead of FB, but those transition baskets off turnovers are still a thing.

To my knowledge, the effectiveness of a doubleteam is mostly the result of how the players involved match up, ratings and IQ, and then it’s just a probability thing. I assume the determination of effectiveness is made toward the start of the possession, but I don’t know for sure. And we don’t know the “help” member of the doubleteam, but we can usually assume the primary guy is the one lined up across from the target, at least with a man base. My impression is also that doubling with zone works a bit differently than it does with man. I have very little experience with man, so I won’t say anything concrete, except to say it feels to me that doubling with a zone base tends to be a little more effective than with man, all other things equal.

Ok I think my double team questions are cleared up.

On the transition basket thing, my point was that the engine clearly has a way of implementing a "fastbreak-style" possession for all offenses, but when running a FB set, does it attempt to implement this same sort of mechanism every time?