Was: Claw Poll - Now: Dice vs RNG

[Mango]

Not The Craw, The Craw!

hahahahaha

[GalakStarscraper]

Uh? FUMBBL discusion at Cyanide forums?

I read multiple forums ... thought some there would as well to see the cross argument.

Tom

[DoubleSkulls]

Regardless of whether its the ultimate game breaking skill or not (and it isn't) the rules aren't changing so what's the point?

[DoubleSkulls]

Uh? FUMBBL discusion at Cyanide forums?

the problem is the same. where you have random match ups in an pure open format how do you do it?

The simplest way is just do it on TV. So Cyanide you'll play within 430k - with, as far as I can tell, a preference towards similar TVs.

That has an impact of making people optimise their team to play very similar TVs and also it means there is a real incentive to keep your TV in the team's perceived sweet spot. In cyanide you'll see a lot of teams with 11/12 players and lots of skills on all of them to reduce TV bloat.

Generally speaking in Blood Bowl, the most enjoyable games are where the teams and coaches are evenly matched. So, IMO, if you want to maximise the enjoyment for every coach in every game then the best thing to do is work out the probability of winning for each coach and assign match ups with a preference towards even odds.

If you accept this then there are three factors that really contribute
1) TV differential (higher TVs are more likely to win)
2) Racial factors (dwarves have a very good chance of beating halflings)
3) Coaching skill (better coaches win more games!)

Chess only needs to deal with the 3rd factor - skill of the players - and the Elo formula was developed to help factor this in to determine the world chess rankings. Basically beating a coach ranked higher than you gets you more points than a coach who's worse ranked, and vice versa.

This gets more complicated for Blood Bowl because its common to see coaches who have "good teams" which out perform their TV - +AG on a thrower is generally a lot better than +AG on a blocker - but adds the same to TV. So you might want to get a ranking based on teams.

Coaches often have a bias to be better or worse with different teams. I think I'm better with Orcs than I am with High Elves, for example. The NAF rankings are purely based on coach record per race.

Finally you might believe some coaches are just better at the game than other coaches, so you might want to track their ability with all races.

That has led me to believe that the most accurate way to assess win probability is to account for the following

1) TV difference
2) Racial difference
(it would be best to combine them into a single factor, but that needs a lot of data to be more reliable than treating each alone)
3) Coach's Team Elo
4) Coach's Race Elo
5) Coach's Individual Elo

Obviously there is then a temptation to use 3,4 & 5 to rank coaches, but that's a separate discussion I think.

[Steam Ball]

430K means you can pick the most expensive Star Player, or a combination of lesser inducements; even if there is a preference to not give so much, at least it is possible to happen. Blackbox is a lot less difference, normally no inducements avaliable.

Such systems that go for pretty much the same TV will be a pain for teams that rely on the inducements. In other words, probably those teams should had get the things in the roster if they were done "to play with this and that inducement by design".

[GalakStarscraper]

Problem there Ian is we didn't design it to work that way. Not sure if you are suffering from short term memory ... but the rules I remember writing with you included the concepts that specific inducements would be part of the balance of the game.

Cyanide is missing a lot of these so it has problems that might be fixed if it had all the inducements in.

FUMBBL's 150k is so tight that it removes some the effects the BBRC delibrately put in.

Ah well ... I'm glad I don't have to go around and around on this now.

Tom

[DoubleSkulls]

Tom - I'm not really sure what you think I'm talking about. I'm agreeing with you...

What I'm saying is that by only matching on TV they are missing a lot of the fun and flexibility that go with BB. They should look for even match ups on coaching skill, TV and race because then more of the games will be even and probably more fun for everyone.

By using factors other than just TV the match ups will have a higher variance of TV than at present and that will encourage teams to be a bit less one dimensional.

430k isn't a lot of inducements - its only enough for Morg. I've seen very entertaining games at much higher gaps.

[Darkson]

Hah - you two know nothing - you obviously forget that Cyanide/Focus know much more about how the games and rules are meant to work than you.
Just cos you wrote the things.

[GalakStarscraper]

Tom - I'm not really sure what you think I'm talking about. I'm agreeing with you...

What I'm saying is that by only matching on TV they are missing a lot of the fun and flexibility that go with BB. They should look for even match ups on coaching skill, TV and race because then more of the games will be even and probably more fun for everyone.

By using factors other than just TV the match ups will have a higher variance of TV than at present and that will encourage teams to be a bit less one dimensional.

430k isn't a lot of inducements - its only enough for Morg. I've seen very entertaining games at much higher gaps.

Okay then we do agree. I really misread your last post. This one is much better put.

Yes we designed the game to work with higher variance of TV and in fact there are teams that are designed to count on it.

All good.

[Smeborg]

The only comment I would make is that, in my experience, the higher the TV difference, the higher the coaching skill requirement for the underdog. A rookie coach is more likely to pick the wrong inducements, or use them badly. (I don't offer a solution to this, other than an unwritten obligation on the part of experienced coaches to advise the rookie underdog.)

All the best.

[funnyfingers]

Smeborg - What does that say for me winning this match: http://www.stuntyleeg.com/index.php?sec ... t&mid=3454

[Smeborg]

Smeborg - What does that say for me winning this match: http://www.stuntyleeg.com/index.php?sec ... t&mid=3454;)

In my experience, it is also possible for an underdog by a large margin to win handsomely (I have done this myself). But it probably implies they are not a rookie coach.

All the best

[DoubleSkulls]

The only comment I would make is that, in my experience, the higher the TV difference, the higher the coaching skill requirement for the underdog. A rookie coach is more likely to pick the wrong inducements, or use them badly. (I don't offer a solution to this, other than an unwritten obligation on the part of experienced coaches to advise the rookie underdog.)

All the best.

There seems to be a lot of qualitative evidence suggesting the opposite. Inducements are a bit too good at higher gaps.

Anyway they aren't meant to make it even, just competitive.

[Smeborg]

It's not a complaint, just an observation. And I agree that high TV differences can favour the underdog.

All the best

[gandresch]

Hi,

i haven't read all the posts (there are quite a lot of them) on Claw.
I think the Claw skill is good as it is except one thing. It reduces all AV to 7 even the AV received by a skill roll (AV+). Therefor I think, it would be good, that the AV+ from skills rolls are added to the 7 afterwards, so that an AV10 Black Orc would have 8 and a Human Lineman with AV+ would have 8, too, after the Claw skill is used. A usual Treeman or Morg would have 7, because they have no AV+ as a skill.

Greets,
gan