Comments on: Is It Better To Be Peaky or Consistent?

By: Jon

Jon — Mon, 11 Jan 2010 15:20:43 +0000

I don't like peaky as an adjective. Meteoric is apt and is already a word. Nonetheless, I enjoyed the article.

By: Dave

Dave — Sat, 09 Jan 2010 07:47:00 +0000

Neat study, and I hardly think you need to run it again to see that your results tend to match up with James's. The group that has the two highest single season win shares totals--which is to say the group with the higher peaks--won more championships. Do you have a season-by-season break down? How many championships did the "peaky" group win in season 1? And how many did the "consistent" group? etc. I would think you'd see a non-linear relationship: the years where a player has 7 win shares would win more than twice as many championships as the years with a player with 3.5 win shares. It's been a long time since I read Politics of Glory, but I remember James's point as basically being that you have to be exceptionally good to win a championship, and that--in terms of championships--it's no big difference if you're a little above average or if you suck: in both cases you don't win the ring. That's kind of what you say in the last paragraph about needing a player playing at an insanely high level to win a ring.

By: pageup

pageup — Fri, 08 Jan 2010 23:00:32 +0000

it'd be interesting to have an actual player comparison to make the point, a peaky player who made the hall against a consistent player that didn't... I'm trying to think of two...

By: izzy

izzy — Fri, 08 Jan 2010 22:29:14 +0000

hmmm maybe i should run the BBR blog.

*kidding* Neil you're the man eff tha haterz

By: Neil Paine

Neil Paine — Fri, 08 Jan 2010 21:57:46 +0000

I was thinking the same thing, too. I'll do a sequel to this post next week where I create an exaggerated version of each player type and then re-run the simulation. I would imagine we'll see much better results with realistic but fictional career arcs instead of this overly-averaged version.

By: Johnny Twisto

Johnny Twisto — Fri, 08 Jan 2010 21:33:28 +0000

I was going to note the same thing. I think by using large groups of real players, you've lost their unique characteristics. If the peaky player has a lower peak than the consistent player, it seems to defeat the purpose.

It might be worth trying to experiment with invented numbers. How many titles is a player who gets 10 WS a year for 10 years likely to win in comparison with one who gets 5 WS a year for 20 years. Or different variations of that.

By: Downpuppy

Downpuppy — Fri, 08 Jan 2010 20:49:59 +0000

And yes, it looks like in combining peaky players, who peak at different points in their careers, you've created a model consistent player.

By: Downpuppy

Downpuppy — Fri, 08 Jan 2010 17:00:01 +0000

Is this about Reggie Miller?

By: Neil Paine

Neil Paine — Fri, 08 Jan 2010 15:50:25 +0000

No, I noticed that, too. It's almost like, from looking at the numbers, it's hard to tell which was peaky and which was consistent. When I first ran the numbers, the peaky players had the higher, well, peak, but I had to drop their total down to make sure both groups had the same # of WS.

By: izzy

izzy — Fri, 08 Jan 2010 15:37:50 +0000

Sorry, I meant to say 2 seasons of 8+ WS for the "consistent" player as opposed to 0 seasons of 8+ WS for the "peaky" players