52 coaches online • Server time: 15:43
Forum Chat
Log in
Recent Forum Topics goto Post Creating a custom to...goto Post Secret League Americ...goto Post DOTP Season 4
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:06 Reply with quote Back to top

humm yea i could do that, so games against those races don't count

in the lowest bracket 960-1210 there were 36414 games played, 1629 involving Ogres, Flings and Gobbos (4%)
so its not that big a deal, i doubt it would change the relative orders around anyway

biggest difference is wood elves, yea. and my results show strong leaning toward fast agile teams

i think maybe WE win almost all the time against certain races, but against most races they lose just below average?
that way, their win rate overall would be high (my approach) but their 'number races they win against' would be low (i think that was VM's approach?)
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:12 Reply with quote Back to top

what i'd really like is a fresh scrape of data. this set includes 107,000 games, but it was early in the adoption of the ruleset

more dataz would show out averages better, in some categories there's very few games played

newer dataz would show reaction to the ruleset stabilising, metagame stabilising, etc.

i've done all this in excel and it would be fairly easy to just drop new data in from the html style that hito first produced and update my charts


Last edited by Sp00keh on Dec 23, 2011 - 13:28; edited 1 time in total
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:16 Reply with quote Back to top

is there a way of actually having images so they're shown on the site? or can i only link to them
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 23, 2011 - 13:20 Reply with quote Back to top

I think there is an interesting trend on show here which quite clearly (to me) points out the reason for all the CPOMB hate.

win percentage drop
Lizardmen -8.26
undead -10.44
Dwarf -9.67
Norse -11.36
Orc -9.78

Because this counts draws as losses you can see that these teams above really suffer at a high TV in terms of winning. I would suggest that these teams are possibly drawing alot of games or losing games against the CPOMB teams that seem to be very common at high TV. Its not that the CPOMB teams themselves win too often, it may be that they cause all but the ag4 teams to lose and draw more that is the big issue for people.

Sp00keh wrote:
is there a way of actually having images so they're shown on the site? or can i only link to them


You need to upload them to your gallery, then put them in a thread like this-
Code:
[img]file path here[/img]
koadah



Joined: Mar 30, 2005

Post   Posted: Dec 23, 2011 - 13:32 Reply with quote Back to top

Sp00keh wrote:
...


Utter nonsense!

Any team with no easy access to C-POMB stinks. It's a well known fact.

_________________
Image
O[L]C 2016 Swiss! - April ---- All Stars - Anniversary Bowl - Teams of Stars - 13th March
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:34 Reply with quote Back to top

Garion wrote:
I don't think Ogres, Flings and Gobbos were counted anywhere other than the first Tv bracket, they dont show on his reading, and in that TV bracket flings and goblins are okay really. Ogres are still pointless mind. Smile


i counted them in all the brackets. i didn't show them when doing the 3x bracketed rankings which were a comparison to your lists


win rates in depth for the 3 teams:

960-1210 TV
Halfling 20.9% 461 played
Goblin 20.3% 543 played
Ogre 17.1% 625 played

1210-1460 TV
Ogre 17.7% 269 played
Goblin 17.2% 257 played
Halfling 13.9% 176 played

1460-1710 TV
Ogre 15.9% 225 played
Halfling 10.4% 104 played
Goblin 10.1% 58 played

1710-1960 TV
Halfling 25.0% 4 played
Ogre 10.5% 120 played
Goblin 0.0% 5 played

1960-2210 TV
Ogre 16.7% 18 played
Goblin 0.0% 0 played
Halfling 0.0% 0 played

2210-2460 TV
Ogre 24.9% 8 played
Necromantic 22.2% 9 played
Lizardman 20.0% 5 played
Undead 14.3% 7 played
Slann 14.1% 7 played
Goblin 0.0% 1 played
Halfling 0.0% 0 played


at the end there, it shows the data really running out. the highest bracket, even the most active team (chaos) has only played 204 games

also its basically impossible to get gobs/flings over 2000TV
also at very high end, ogres have some results but this is probably just due to 1 coach, so not worth drawing any analysis from
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:42 Reply with quote Back to top

Garion wrote:
I think there is an interesting trend on show here which quite clearly (to me) points out the reason for all the CPOMB hate.

win percentage drop
Lizardmen -8.26
undead -10.44
Dwarf -9.67
Norse -11.36
Orc -9.78


that's what i was looking for as well

but on the other hand it could be due to these teams having a strong natural starting point.
maybe their str4, good basic skills, or whatever

and then by the time the other teams have also developed, this early advantage has faded, in other words loads of block for norse isn't amazing once everyone else has block.

the cpomb teams don't drastically climb once they reach the middle-high TV, where their players would have 3-4 skills like i was expecting to see as proof of cpomb's dominance:

mutation access teams - change as TV climbs:
Skaven -1.26%
Chaos Dwarf -3.65%
Chaos 4.83%
Nurgle 4.52%
Chaos Pact -8.52%
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 13:56 Reply with quote Back to top

Image

still isn't working :/
file is in my gallery
Code:
[img]http://fumbbl.com/i/353014[/img]


my previous links i'd used IMG tags on the uploads i did to tinypic and imageshack hosting sites, any ideas?
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 23, 2011 - 13:59 Reply with quote Back to top

While i agree in terms of undead and Norse who theoretically should be doing better at a low TV and possibly worse later on - Orcs, Dwarves and Lizardmen are all Av9 teams, who have traditionally done very well at all TVs, the only reason I can think of for them not doing so well anymore at a high TV is CPOMB.

Pact and Chaos Dwarves will struggle a bit more than usuall at a high TV because they have a lower average St than Nurgle and Chaos and will in all probabilty be getting battered by those other CPOMB teams, and there is also a good chance these CPOMB teams are messing up each others stats somewhat by drawing a lot of their games. While the Ag 4 teams are better at getting wins than any team. No suprise there really.

_________________
Image
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 14:25 Reply with quote Back to top

yea, thinking about it a bit more,

claw players are generally slower than average, so although they can smash other team about (which will stick in people's minds), they still may not get the win and could often be held to a draw

in other words, you can crush your enemies, drive them before you, and hear the lamentations of the women - but unless you score more touchdowns than they did, it still goes down in hito's results as a loss
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 23, 2011 - 14:33 Reply with quote Back to top

Its working fine now, you just need to wait for an admin to approve the image before people can see it.

what would be interesting to me is if you get rid of all the ag4 teams and mirror matches and see how teams win ratio changes then, I would expect all the CPOMB teams to do a hell of a lot better then.

There is no doubt this rule set has without doubt benefited the Elf teams the greatest because of journeymen, wizards as inducements and the introduction of wrestle. I think if you did that test above it would show how much all the other teams suffer at the hands of cpomb more, which is a shame as it has limited diversity a lot and also stopped the teams that can hit elf teams the hardest from being more popular, by this i mean teams like humans, norse and undead etc... which i personally think are the hardest teams for elves to face.
dode74



Joined: Aug 14, 2009

Post   Posted: Dec 23, 2011 - 14:36 Reply with quote Back to top

Just to be clear, this counts draws as losses? If so it's a new metric and I don't think particularly helpful.
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 23, 2011 - 14:43 Reply with quote Back to top

It is useful at seeing which teams win the most, as is its intention. You can also draw some other conclusions as well as discussed on this page, from how certain teams win alot less at higher TVs than lower TV. It is just as useful as any other stats really.
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Dec 23, 2011 - 14:50 Reply with quote Back to top

yea, draws as losses. thats what hito's data says

it'd be possible to reconstruct draw rate because he shows both sides of the matchup

eg his format is of the style:
amazon vs chaos- games played 136 win rate 40%
chaos vs amazon- games played 136 win rate 45%

hence the draw rate would be 15%

i've not included, not sure how i'd add it in as my excel is all structured around just 'win rate'

27.7% of the games in the sample are draws, at most i'd expect it to vary only between 25-30% with the elf style teams probably draw less and the slow teams to draw more
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 23, 2011 - 14:52 Reply with quote Back to top

ahh, are these tables taken from hitos data because his blog says they are now updated so draws count a 0.5% http://fumbbl.com/FUMBBL.php?page=blog&coach=30979&op=viewcomments&id=10620
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic