Improvement table awarding a given value of improvements.

Got a great idea and/or proposal for BloodBowl?

Moderator: TFF Mods

Post Reply
dode74
Ex-Cyanide/Focus toadie
Posts: 2565
Joined: Fri Jul 24, 2009 4:55 pm
Location: Near Reading, UK

Re: Improvement table awarding a given value of improvements

Post by dode74 »

First of all the top 2 teams cause the most cas per game on average, they also have the highest winning percentages. Then if you remove the elves and skaven from the equiation the table makes pretty sorry reading for everyone else who seems to be suffering round about 3 cas a game and all have a win percentage lower than 50%. The only exception to that is amazons who will be doing quite well because their whole team will be blodged up will have at least 4 gaurd players and a little bit of punch as well. Pretty much the perfect team for facing lots of cpomb.
That is not an indication of a correlation between causing casualties and winning games though. You're applying a thumb-in-the-air to what is actually a statistical method and claiming it is valid. Read the link I gave you.
win with any consistency.
Win% is fine though :P

Reason: ''
User avatar
garion
Ex-Mega Star, now just a Super Star
Ex-Mega Star, now just a Super Star
Posts: 1687
Joined: Mon Jun 07, 2010 12:59 pm

Re: Improvement table awarding a given value of improvements

Post by garion »

dode74 wrote:
First of all the top 2 teams cause the most cas per game on average, they also have the highest winning percentages. Then if you remove the elves and skaven from the equiation the table makes pretty sorry reading for everyone else who seems to be suffering round about 3 cas a game and all have a win percentage lower than 50%. The only exception to that is amazons who will be doing quite well because their whole team will be blodged up will have at least 4 gaurd players and a little bit of punch as well. Pretty much the perfect team for facing lots of cpomb.
That is not an indication of a correlation between causing casualties and winning games though. You're applying a thumb-in-the-air to what is actually a statistical method and claiming it is valid. Read the link I gave you.
win with any consistency.
Win% is fine though :P

You have to be blind to think there isnt a corrolation between those teams high Cas caused and win percentage. It is what those teams are now all about.

Depends what your view of fine is. Look at orcs for example - for me Orcs should never have an average win percentage lower than 50% they should be one of the top performing teams such is the strength of their roster. On that link there percentage is only 41% :( madness!

Reason: ''
dode74
Ex-Cyanide/Focus toadie
Posts: 2565
Joined: Fri Jul 24, 2009 4:55 pm
Location: Near Reading, UK

Re: Improvement table awarding a given value of improvements

Post by dode74 »

Would that be the link with the artificial filter of TV180+ applied? Because if so you are cherrypicking.

As for whether there is a correlation for those teams, without running the full stats I couldn't say. You, however, are leaping to conclusions if you say there is. I'm not saying you are necessarily wrong, just that the data you have provided does not support what you are saying because you haven't actually looked for a correlation.

Reason: ''
User avatar
garion
Ex-Mega Star, now just a Super Star
Ex-Mega Star, now just a Super Star
Posts: 1687
Joined: Mon Jun 07, 2010 12:59 pm

Re: Improvement table awarding a given value of improvements

Post by garion »

yes 180+ because this is where the problem lies and too a lesser extent 1600-1800. This is what we are all saying?

As I said before see if you can filter out the elves and see if it changes much, it should be interesting?

Reason: ''
User avatar
mattgslater
King of Comedy
Posts: 7758
Joined: Thu Dec 15, 2005 5:18 pm
Location: Far to the west, across the great desert, in the fabled Land of Comedy

Re: Improvement table awarding a given value of improvements

Post by mattgslater »

Dode, what we don't have going on in the online format is a long history of teams at similar experience levels but very different TVs. Fortunately, my experience in leagues is that a lot of coaches try to balance games by TV rather than matches played or records, so this data will be valuable.

1800 is a good line, given that we're talking about high-level problems and trying to compare apples to apples as much as possible. All the teams at 1800+ have played long enough to get some good players. Most of them have stars. The difference between a team that has played 20 games and one with 100 games is notable, in that the 100-game team will have a lot more variance in player ability, with legends and rookies, while the 20-game team will, if successful, have a lot of players in the 1-4 skill range. But by contrast, 1500 teams in those environments will feature brand new outfits in with the minmaxed Norse/CP/Zon predators, which doesn't happen much in league environments.

If you want to factor in mismatches, you can intersperse a cocktail of tournament results. You may have to wait another year for the new [R] rules to mature before you get the best possible picture.

Reason: ''
What is Nuffle's view? Through a window, two-by-three. He peers through snake eyes.
What is Nuffle's lawn? Inches, squares, and tackle zones: Reddened blades of grass.
What is Nuffle's tree? Risk its trunk, space the branches. Touchdowns are its fruit.
dode74
Ex-Cyanide/Focus toadie
Posts: 2565
Joined: Fri Jul 24, 2009 4:55 pm
Location: Near Reading, UK

Re: Improvement table awarding a given value of improvements

Post by dode74 »

ANY line by TV is a bad line. It fails to take into account the lows as well as the highs and, as such, rewards those teams which are better suited to staying above that line. A far better metric is number of games played - that metric is equal for all teams and makes no distinction for ability to maintain a certain level of anything, because once you reach the level you are past it and can never go back.
But by contrast, 1500 teams in those environments will feature brand new outfits in with the minmaxed Norse/CP/Zon predators, which doesn't happen much in league environments.
You also get teams which have taken a beating and are on their way back up, or with a couple of 200TV stars MNG or dead.

If you look at games played beyond 30 then you have a fair metric. TV as a metric is reliant on ability to keep players alive; games played is not.

Reason: ''
dode74
Ex-Cyanide/Focus toadie
Posts: 2565
Joined: Fri Jul 24, 2009 4:55 pm
Location: Near Reading, UK

Re: Improvement table awarding a given value of improvements

Post by dode74 »

Fortunately, my experience in leagues is that a lot of coaches try to balance games by TV rather than matches played or records, so this data will be valuable.
I just noticed this. Why do they do that? We're back to the "just because everyone does it doesn't make it right" argument. The game was not meant to be matched by TV alone.

Reason: ''
User avatar
Darkson
Da Spammer
Posts: 24047
Joined: Mon Aug 12, 2002 9:04 pm
Location: The frozen ruins of Felstad
Contact:

Re: Improvement table awarding a given value of improvements

Post by Darkson »

viewtopic.php?f=16&t=35407

If anyone's actually bothered about the idea, rather than arguing about an exact wording of what the issue is, whether ClawPOMB is OP, etc etc etc, then please you that thread.

If you still want to argue about this crap, then please continue to do so here.

Reason: ''
Currently an ex-Blood Bowl coach, most likely to be found dying to Armoured Skeletons in the frozen ruins of Felstad, or bleeding into the arena sands of Rome or burning rubber for Mars' entertainment.
Post Reply