1175Re: [APBR_analysis] Euclidian comparison of current and past players
- Aug 2, 2002On Fri, 2 Aug 2002, mikel_ind wrote:
> In APBR room, I touched on the Euclidian Distance method of comparing[...]
> statistical profiles. This got me to thinking.
> Currently, there are a number of players whose career totals bring up
> Hall of Fame mention. Others have just been around forever, and so
> have amassed big numbers.
> I decided to use career totals, rather than per-game/t*nd*x-type
> stuff. Playoff games are included, but weighted no than regular-
> season games.
> In no particular order, current players and those from the past with
> most-similar career totals:
> The Euclidian Distance has 6 factors: pts, reb, ast, stl, blk, and
> total of those first 5.
Interesting stuff, but I think the lack of rate and ratio statistics
weakens the formula. If nothing else, I think you should put career total
minutes in there, to distinguish players who had similar totals, but who
may've amassed their stats at different rates. E.g.:
> Mookie BlaylockIt's believable that KJ's career totals may be fairly similar to
> .41 Cheeks
> .54 Richie Guerin
> .63 KJ
Blaylock's, but KJ piled up the points and assists (per game and per
minute) faster than Mookie. Also the lack of a FG% measure (corrected for
era if you want to get fancy) means that one of the biggest differences
between Cheeks and Blaylock isn't getting measured.
> Chris WebberTwo more examples of the pitfalls of looking only at totals and not rates;
> .28 Mickey Johnson
> Tim Duncan
> .23 Thurl Bailey
of course Webber and Duncan are both young and so their career totals are
unknown, so getting good matches is much more complex. Interestingly, I
think of Mickey Johnson and Thurl Bailey as being sort of similar to each
other: tall thin guys who were too skinny for PF but much taller than
typical SFs of the era, who could score in a variety of ways (not just
- << Previous post in topic Next post in topic >>