Wingstop are top 5 among active players in postseason TS%

[Chinook]

You missed the point. I understand the calculation fine. What I said was that you can't look at it in a vacuum. Since the ultimate point of stats is always to compare one player to another, it's probably a good idea to understand the times when a stat doesn't do that very well... don't you think? If you look at Dwight Howard's TS%, and it's higher than Steph Curry's, does that meanDwight is a better player, or would help a team more? I can't help but look at the "why", when I see something like that.

So you say the first thing, then you say the second thing. If you think the point of TS% is to say who's better or who's more helpful, then you aren't interpreting it correctly. Is it possible for a dude who mainly dunks to be more efficient than a good three-point shooter? yes. Tyson Chandler in his prime was arguably the most efficient player of the modern era. That doesn't make him a better player or even a better helpful player than Steph, but it doesn't have to.

For the record, sample size isn't the cure-all you say it is. Position and playing style are a big factor, and they have an effect over the whole season, and whole careers. For his career, Howard has something like 11 FGA to 9 FTA per game. And a LOT of those FTA's are scoring opportunities without any FGA's associated. I just pulled a random year for Shaq. He had 336 two-shot FTA opportunities that year. That's the equivalent of 168 2P shots going up, without any FGA's to go with them. He made a ty 47% of his FT's, but that gave him a .534 TS% for those 168 scoring opportunities. That's a big enough sample, isn't it?

There is not a "big enough sample" it can always be bigger. So as you include other players and seasons, the distribution will become normalized, and the average will approach the mean. Therefore, you will eventually get to the .44 coefficient, which will make TS% near-perfect. Some players may be aberrations, but that won't be a big enough number to skew the usefulness of the stat. You can make it more accurate for a player by moving the .44 coefficient toward .5 if they had mostly shooting fouls and toward 0 if they have more and-1s and Ts, but you don't really have to worry about that in the grand scheme. It evens out.

Since you've established yourself as the bull-goose authority on everything, you should also understand that TS% rewards the out of high-variance 3P shooters. A guy who makes 2-10 3-pointers one game, and 6-10 3-pointers the next game, has a scorching .600 TS%. But there's a good chance that his shooting lost his team just as many games as it won. That's not an aberration of sample size. It's something that happens over and over again through a season, and the effect gets magnified in the playoffs when more games are grind-it-out contests in the last few minutes. 3P shots score a lot of points, but they also result in more dry possessions, and more points being scored on the other end. But that doesn't show up at all in TS%, does it? That's always been my knock on Manu - he plays high variance ball even in close, close games.

Since when is it the burden of TS% to show consistency? Or to show how defenses respond to the misses? There can be other stats for that. It would be like blaming blocks for not showing how far away from the basket a player was when they are blocked. Like who cares? That's for other stats to determine. I'm not saying you aren't asking for good data, but I am saying you shouldn't judge a basic stat like TS% for not being an advanced stat.

Yes, though, sample size does limit variance. That's a fundamental law of statistics, not my opinion.

The point is, TS% tells you something at times, just like any other stat. But the inferences the OP was making? Not so much.

It tells you exactly what it's supposed to tell you, and it doesn't tell you what it's not supposed to tell you. If you want gravel, beat a stone. If you want water, go to the river. Don't be mad at the stone because you're thirsty.

[SpursFan86]

TS% is supposed to give you an idea of how efficient of a scorer a player is...that's it. It looks at how many FGA and FTA a player has, and how many points a player scored off those attempts, and then spits out a number. It's not designed to tell you which player is better. All it aims to show is shooting efficiency. Obviously there's more to basketball to that. But criticizing the stat itself because some people misuse it seems strange.

[Seventyniner]

TS% is supposed to give you an idea of how efficient of a scorer a player is...that's it. It looks at how many FGA and FTA a player has, and how many points a player scored off those attempts, and then spits out a number. It's not designed to tell you which player is better. All it aims to show is shooting efficiency. Obviously there's more to basketball to that. But criticizing the stat itself because some people misuse it seems strange.

Even then, two players can have the same TS% but one could have a much higher TO%, making them less efficient on offense overall. You're right, it's just one stat.

[tholdren]

So you say the first thing, then you say the second thing. If you think the point of TS% is to say who's better or who's more helpful, then you aren't interpreting it correctly. Is it possible for a dude who mainly dunks to be more efficient than a good three-point shooter? yes. Tyson Chandler in his prime was arguably the most efficient player of the modern era. That doesn't make him a better player or even a better helpful player than Steph, but it doesn't have to.



There is not a "big enough sample" it can always be bigger. So as you include other players and seasons, the distribution will become normalized, and the average will approach the mean. Therefore, you will eventually get to the .44 coefficient, which will make TS% near-perfect. Some players may be aberrations, but that won't be a big enough number to skew the usefulness of the stat. You can make it more accurate for a player by moving the .44 coefficient toward .5 if they had mostly shooting fouls and toward 0 if they have more and-1s and Ts, but you don't really have to worry about that in the grand scheme. It evens out.



Since when is it the burden of TS% to show consistency? Or to show how defenses respond to the misses? There can be other stats for that. It would be like blaming blocks for not showing how far away from the basket a player was when they are blocked. Like who cares? That's for other stats to determine. I'm not saying you aren't asking for good data, but I am saying you shouldn't judge a basic stat like TS% for not being an advanced stat.

Yes, though, sample size does limit variance. That's a fundamental law of statistics, not my opinion.



It tells you exactly what it's supposed to tell you, and it doesn't tell you what it's not supposed to tell you. If you want gravel, beat a stone. If you want water, go to the river. Don't be mad at the stone because you're thirsty.

LOL chinook always goes to this argument. No one has every "interpreted" a stat correctly except chinook. Yet he is unable to come to terms with statistics being something that happened in the past to ESTIMATE what COULD happen in the future. It's sad really. Chinook is a smart dude, he just can't grasp the foundation. One day