Who has the stats?

redlizard · Post by **redlizard** » Wed Jan 22, 2003 6:10 pm

Has anyone who is running a league with a LOT of players collected enough statistics to create a table that shows the percentage chance of the "stronger" team winning at different levels of TR difference? I'm imaging something like this:

TR Diff Strong Win
0-10 50%
11-25 55%
26-50 58%
......

Coaching ability and a lot of other factors make a difference in addition to TR, that's why a the data has to come from a lot of coaches with a lot of games. Anyone have this? Just curious to see the results.

Dangerous Dave · Post by **Dangerous Dave** » Wed Jan 22, 2003 6:30 pm

I've got a lot of this type of info.... will post it later.

Dave

christer · Post by **christer** » Wed Jan 22, 2003 9:25 pm

I could probably make a view like that for the FUMBBL data...

Does ~1500 matches qualify?

-- Christer

grep-v · Post by **grep-v** » Thu Jan 23, 2003 3:09 pm

redlizard wrote:Has anyone who is running a league with a LOT of players collected enough statistics to create a table that shows the percentage chance of the "stronger" team winning at different levels of TR difference?

I suppose just the difference in TR wouldn't be that meaningful. IMHO it would at least be necessary to view the results as a multinomial distribution in Race1, Race2, TR1, TR2.

For example the difference between TR 100 and 150 is not at all like 350 and 400 and winning with TR 100 undead vs. TR 100 chaos is not at all like TR 300 on both teams.

christer · Post by **christer** » Thu Jan 23, 2003 4:59 pm

Check the FUMBBL statistics

This message has been brought to you by the SCWM (StatCrazyWebMasters) group.

-- Christer

christer · Post by **christer** » Thu Jan 23, 2003 5:16 pm

I suppose just the difference in TR wouldn't be that meaningful.

I think it might be useful if you want to come up with a new handicap table...

-- Christer

The UniGoblin · Post by **The UniGoblin** » Fri Jan 24, 2003 2:48 am

christer wrote:Check the FUMBBL statistics

Wow! Those rated over 90 are guaranteed wins?! Quite a nice bit of bribery goin' on there, eh?

neoliminal · Post by **neoliminal** » Fri Jan 24, 2003 7:00 am

Wow. These number are actually pretty close to what we had predicted (give or take the ties...)

Dangerous Dave · Post by **Dangerous Dave** » Fri Jan 24, 2003 9:16 am

I tried posting my info but it didn't work - I'll post an html later.

Dave

Post by **DoubleSkulls** » Fri Jan 24, 2003 9:41 am

christer wrote:Check the FUMBBL statistics

This message has been brought to you by the SCWM (StatCrazyWebMasters) group.

-- Christer

That looks pretty good - but I don't think there are enough results for 90+ differences to say that they are guaranteed a win.

Grumbledook · Post by **Grumbledook** » Fri Jan 24, 2003 9:42 am

Well as we don't have use of the handicap table most people avoid playing tr differences of 30 let alone 90.

Post by **GalakStarscraper** » Fri Jan 24, 2003 10:44 am

neoliminal wrote:Wow. These number are actually pretty close to what we had predicted (give or take the ties...)

JKL .... keep in mind one thing ... since its not automated yet in the program, the FUMBBL league doesn't use the handicap table or any aging rules (I think this is true on the aging, but I'm not 100% sure).

Christer as soon as you can get Ski to automate this items ... you will have one hell of a stats machine for the game.

Footnote: the stats it creates are already great, but I'd love to see it model the game with the 2 long term balancers in play (ie handicap and aging).

Galak

Grumbledook · Post by **Grumbledook** » Fri Jan 24, 2003 10:47 am

We do use aging, its only the handicap table we don't use, in fact christers star ghoul aged a couple of hours ago, hahahahaha.

The aging is done by hte website, along with all the player purchases etc, rather than in javabowl.

Post by **GalakStarscraper** » Fri Jan 24, 2003 11:03 am

Grumbledook wrote:We do use aging, its only the handicap table we don't use, in fact christers star ghoul aged a couple of hours ago, hahahahaha.

The aging is done by hte website, along with all the player purchases etc, rather than in javabowl.

Ah good ... okay I just thought I had saw some very skilled players that had not aged. Well then this is me saying to Christer ... thank you, thank you, thank you ... basically your top star page kinda proves out on the BBRC beliefs. Skill based aging doesn't really work. I've long believed this ... and now I really believe it ... okay time for EXP to get its turn...

I think it be experimental in the BB Annual so I look forward to it coming.

Oh and as long as the aging is in play ... Ski might want to wait on going gun ho automating the handicap table. Every proposal I've seen from a BBRC member is a much smaller better table. No reason to automate all that junk if 1/2 of it will be trimmed out in October.

One thing you might want to consider if you want to encourage high rated team playing lower rated ones. Pariah's old suggestion of a free team reroll for the game for every 25 points of TR difference would be a very very simple add to the program by Ski as a handicap option. Just a thought.

Galak

Anyway, Christer, although we've agreed to disagree on what makes a good Necro team ... I love your league (even though I don't play in it) and the raw power of numbers that it provides. Thanks for building such a great engine.

Grumbledook · Post by **Grumbledook** » Fri Jan 24, 2003 11:07 am

Yer ive played against some of those monster players who didn't age, its not nice. Id much rather give exp a try in the league.

Talk Fantasy Football

Who has the stats?

Who has the stats?

Re: Who has the stats?