An arbitrary statistical ranking of career accomplishments

[theMoMA]

In the NBA elimination thread, I made a spreadsheet assigning arbitrary point values to certain career accomplishments to help people compare basketball careers. I decided to do something similar for quizbowl.

I compiled some information on players who came to prominence after the start of the "player poll era." I was interested in quantifying the careers of the best players whose careers began in that era because that's when the most complete information exists, and because the poll itself gave me an easy heuristic for inclusion. For now, I included only those players whose careers I believed to contain no accomplishments relevant to this exercise before the 2006-07 season. So the generation including Seth, Jerry, Sorice, etc. doesn't show up here. (And members of the preceding generations, such as Zeke, Andrew, Subash, etc., are obviously not here either.)

I will eventually go back and compile a full accounting of careers from 2000 to present based on the data that exists.

Methodology: Anyone who appeared on at least 2 player polls from 2007 to 2014 was included (I reconstructed the lost 2007 poll from my emails, although no one from the relevant generation was ranked).

The accomplishments I took into consideration are, from left to right: player poll appearances (PPX-Y; X and Y denote ranges), ACF Nationals top-12 scoring (NatsT12), ICT top-12 scoring (ICTT12), Chicago Open top-12 scoring (COT12), ACF Nationals victories (NatsW), ACF Nationals runner-up finishes (NatsF), ACF Nationals top-4 finishes (NatsT4), ACF Nationals top brackets (NatsTB), ICT wins (ICTW), ICT runner-up finsihes (ICTF), ICT top-4 finishes (ICTT4), ICT top brackets (ICTTB), Chicago Open wins (COW), Chicago Open runner-up finishes (COF), ACF Nationals DII titles (NatsDII), ICT DII titles (ICTDII), ACF Nationals undergrad titles (NatsUG), ICT undergrad titles (ICTUG).

I assigned the following arbitrary point values to the accomplishments:

12 points: Wins at Nationals or ICT
10 points: Wins at Chicago Open
5 points: runner-up finishes at Nationals or ICT; top-5 player poll appearances
4 points: runner-up finishes at Chicago Open; 6-10 in player poll
3 points: 11-15 in player poll
2 points: top-4 finishes at Nationals and ICT; all undergraduate and DII titles; 16-20 in player poll
1 point: top-12 scoring at Nationals, ICT, or Chicago open; top-bracket finishes at ICT or Nationals; 21-25 in player poll

The columns at the end represent the total number of points (Total), the total number of unweighted accomplishments (Grey ink), the total number of top brackets at ICT and Nationals (TBs), and the total number of wins at ICT and Nationals (Titles).

Without further ado, here is the table. I apologize for constructing a ranking that places me first; since this is the second time I've come out on top of a stat I made, perhaps the Dwight Wynne Award For Self-Aggrandizement In Quizbowl Statistics can be renamed in my honor!

If you notice an error in your own numbers or someone else's, let me know. If you've got any suggestions for tweaking the point values or other accomplishments that might be worth adding, discuss below.

[Nine-Tenths Ideas]

I'm really intrigued by this idea, but I wanted to tweak your metrics a little bit just to see what would happen.

So for this build of the table, I assigned some number of points to players for the following accomplishments:
38th place finishes at PACE NSC
13th place finishes at ACF Nationals
Number of Superpowers at Triacontakaipentagon
Stand-up Comedy Competitions Won
Number of PACE NSC All-Star Games Read

Using these metrics, I created a table. It can be viewed here. The results may surprise you!

[vinteuil]

In all seriousness, Andrew, you surely need to account for who was playing each thing, otherwise someone who's been playing at a middling level since 2000 sweeps etc.

[Skepticism and Animal Feed]

I was expecting to see David Seal and Michael Arnold ranked very high on this list, ahead of many objectively better players, and my critique was going to be that this rewarded players for how good their teammates are (determined by luck and university admission committees). But I see now that the player poll inclusion has mostly screened coattail riders out.

I suppose PPG/individual performance can tell you a lot about what tier players fall into, and then the inclusion of championships/near-championships gives you an additional tool to judge the top tier against itself. So this may be useful for ranking the upper end of players, which I suppose is the one we typically argue about.

[theMoMA]

In all seriousness, Andrew, you surely need to account for who was playing each thing, otherwise someone who's been playing at a middling level since 2000 sweeps etc.

I'm not sure what this is supposed to mean. My goal (as in the NBA thread) is nothing except a quick-and-easy ranking of gross accomplishments. Anyone who looks at this chart and sees that Matt Bollinger and Matt Jackson have achieved 90% of what I have in my career despite earning those accomplishments in 4 instead of 6 years (and racking up more titles and higher individual honors in the process) and nonetheless concludes that my career was "better" is dumb. This is only supposed to tell you what everyone has objectively done so people can better conceptualize and compare entire careers. (I think people have a very good intuitive sense of how good people were at their peak, with a few exceptions like forgetting how good Brendan was in 2010, but are less good at visualizing sustained success over many seasons.)

As I told John Lawrence via PM, I plan to make a web app that has all this data and more going back to about 2000; ideally, it will also have some data on tournaments played and years active, so that average stats can be constructed alongside bulk counting numbers. I also hope to allow people to create their own weights so they can sort by whatever metrics they value. The arbitrary values that I've chosen (with some help from people in the IRC a few weeks ago) reflect my own intuitions on how people feel about the relative values of these accomplishments. Obviously you can make other value judgments that would make things shake out slightly differently (downplaying DII and undergrad titles would vault Eric into the lead, for example; valuing stand-up competitions won and NSC all-star games read results in Isaac's ascendance), but the general order is pretty tough to subvert.

[Demonic Leftovers]

I was expecting to see David Seal and Michael Arnold ranked very high on this list, ahead of many objectively better players, and my critique was going to be that this rewarded players for how good their teammates are (determined by luck and university admission committees). But I see now that the player poll inclusion has mostly screened coattail riders out.

Thank god we dodged that bullet!

[vinteuil]

In all seriousness, Andrew, you surely need to account for who was playing each thing, otherwise someone who's been playing at a middling level since 2000 sweeps etc.

I'm not sure what this is supposed to mean. My goal (as in the NBA thread) is nothing except a quick-and-easy ranking of gross accomplishments. Anyone who looks at this chart and sees that Matt Bollinger and Matt Jackson have achieved 90% of what I have in my career despite earning those accomplishments in 4 instead of 6 years (and racking up more titles and higher individual honors in the process) and nonetheless concludes that my career was "better" is dumb. This is only supposed to tell you what everyone has objectively done so people can better conceptualize and compare entire careers. (I think people have a very good intuitive sense of how good people were at their peak, with a few exceptions like forgetting how good Brendan was in 2010, but are less good at visualizing sustained success over many seasons.)

Ah, that makes more sense; I'm still curious though: if you're just trying to show exactly what people did, why not just make the chart and forego the statistic?