How accurate is bbref's Head-to-Head stat?

[turk3d]

http://www.basketball-reference.com/pla ... finder.cgi

Anyone know? If accurate, I find this to be very a very useful stat in comparing players (especially those who play the same positions).

[Bobbcats]

I don't see why it wouldn't be completely accurate.

[turk3d]

I don't see why it wouldn't be completely accurate.

Depends on how they do the data collection.

[Bobbcats]

I don't see why it wouldn't be completely accurate.

Depends on how they do the data collection.

? Unless I'm missing something they just look at box scores, and include the game as "head to head" if both players played.

[turk3d]

I don't see why it wouldn't be completely accurate.

Depends on how they do the data collection.

? Unless I'm missing something they just look at box scores, and include the game as "head to head" if both players played.

Then probably not to accurate since I'd be more interested in it if they were directly matched up. Wouldn't take into account substitutions or when adjustments made and said player is playing a different position on the court. May not be that reflective of how one player plays against the other in that case. Boxscores don't really tell it all.

[Gideon]

Depends on how they do the data collection.

? Unless I'm missing something they just look at box scores, and include the game as "head to head" if both players played.

Then probably not to accurate since I'd be more interested in it if they were directly matched up. Wouldn't take into account substitutions or when adjustments made and said player is playing a different position on the court. May not be that reflective of how one player plays against the other in that case. Boxscores don't really tell it all.

They're not looking at some game from 1988 and figuring out exactly how many minutes Pippen guarded Bird or whatever. These are just games/averages from when two players on different teams were both in the lineup.

That said, they still have usefulness. For example, I've linked to the head-to-head matchup on bbref to debunk the "KJ dominated Stockton" myth, since they obviously played the same position and were matched up against each other most of the time when their teams played (and Stockton usually did better).

I think that's the best way to use the tool -- when we know players played the same position and were matched up against each other at least most of the time (so, especially with PGs and Cs). It might not be super-precise, but it can give us a good idea of how Stockton and KJ (or Hakeem and Shaq, or Bird and Pippen, or Duncan and Garnett, etc.) actually did against each other. It can also show how players did against certain teams -- for example if we enter Bird vs. Magic, they obviously aren't matched up with each other most of the game, but it will show us how Bird did against the Lakers and how Magic did against the Celtics.

[turk3d]

? Unless I'm missing something they just look at box scores, and include the game as "head to head" if both players played.

Then probably not to accurate since I'd be more interested in it if they were directly matched up. Wouldn't take into account substitutions or when adjustments made and said player is playing a different position on the court. May not be that reflective of how one player plays against the other in that case. Boxscores don't really tell it all.

They're not looking at some game from 1988 and figuring out exactly how many minutes Pippen guarded Bird or whatever. These are just games/averages from when two players on different teams were both in the lineup.

That said, they still have usefulness. For example, I've linked to the head-to-head matchup on bbref to debunk the "KJ dominated Stockton" myth, since they obviously played the same position and were matched up against each other most of the time when their teams played (and Stockton usually did better).

I think that's the best way to use the tool -- when we know players played the same position and were matched up against each other at least most of the time (so, especially with PGs and Cs). It might not be super-precise, but it can give us a good idea of how Stockton and KJ (or Hakeem and Shaq, or Bird and Pippen, or Duncan and Garnett, etc.) actually did against each other. It can also show how players did against certain teams -- for example if we enter Bird vs. Magic, they obviously aren't matched up with each other most of the game, but it will show us how Bird did against the Lakers and how Magic did against the Celtics.

You make a good point. Back in the day mostly those guys did match up at the same positions unlike today where you get all kinds of "screwy" matchups going on. And what you say regarding it's usefulness makes sense too. However, what I am most interested in is how certain players match up with each other, as I think this would really be the best way to determine who' better between two players, how they match up against each other.

Kinda like the Bird/Magic debate which used to go on as to who's better. Really difficult to settle those types of comparisons. However if they each went up against each other head-to-head I think you can make a more worthwhile assessment. This removes all the other factors (to some degree) regarding teammates, coaches, team style, etc. Thank you very much for your imput, it's appreciated.

[Gideon]

Thanks! -- Glad it helped