FUMBBL :: Online Blood Bowl League

treborius

Joined: Apr 05, 2008

Posted: Jan 14, 2009 - 12:01

DukeTyrion wrote:

Squishy teams have a problem in leagues and KOs, because they don't have time to recover from a bad game before the next match, where a bashy team would have got stronger. So in a VET smack they may be 150/150 vs 150/148 for game 1, but often it's 144/131 vs 158/156 for the second game.

However, in Blackbox, every game is a recovery match, since you are always playing against someone of equal TS. Therefore whether they are good at growing or not, is not very relevant.

i agree it's not relevant for the chances to win next game.

personally, though, i would still like to see black-box to account for
the durability of a team (a bit, at least) as well, because i think team-building
is still important to quite a few blood-bowlers (me included) and it would
probably even out racial imbalancies a bit better (especially at higher
TR/TS-levels) than it is currently the case.

whether or not BBR does that already i can't really say, but i'd like to learn
how others perceive this (BBR)?

sk8bcn

Joined: Apr 13, 2004

Posted: Jan 14, 2009 - 12:14

I don't think that low av teams should have a boost, however, a thing should be done in order to have a better selection of opponents (playing bashers everytime is angering).

I play elf at the moment, but how long before the cas sound I usually get becomes boring?

_________________
Join NL Raises from the Ashes

Frankenstein

Joined: Aug 02, 2003

Posted: Jan 14, 2009 - 13:13

There definitely is a significant problem regarding TR-calculation of agility modifiers:

Code:

- pro elf lino (vanilla) : 50k ok
- pro elf lino (+DP) : 80k ok
- pro elf lino (+Guard) : 80k ok
- pro elf lino (+Block) : 70k ok
- pro elf lino (+AG) : 110k fail

- beastman (vanilla) : 60k ok
- beastman (+AG) : 105k ok

- skeleton (vanilla) : 30k ok
- skeleton (DP) : 65k ok
- skeleton (DP, -AG) : 45k fail

1) No way a pro elf lino with AG+1 is worth 2.17 odinary lineelfs. His agility upgrade doesn't really add much to the team (I'd say with just AG+1 he is even weaker than his counterparts with Block or DP), whereas a beastman with AG+1 fundamentally changes the options of a chaos side.

2) AG-1 on skeleton is 5k at most. A 20k discount makes AG-1 extraordinarily valuable as it essentially grants a 20k TS-drop for free. This makes AG-1 skeletons highly desirable, which is simply wrong and entirely against the spirit of both the rules and blackbox.

The latter is one of the reasons, why I am convinced it would be better for blackbox if there were no injury-discounts at all, after all the rules want injuries to result in player-turnover. To prevent any thisregarding abuse, all injury-discounts should perhaps be entirely removed (the risk of keeping injured players should exclusively affect the responsible coach, not his innocent opponent). Anyway, I doubt that blackbox-games are supposed to become pre-game niggle-gamblings (which punish innocent coaches 5/6 times).

Personally, instead of a complicated TR-formula (which will always have its flaws and be reason for tons of discussions), I'd would suggest to go with a straightforward approach:

Just use LRB5's costs, bad team development choices should be up to the coach exclusively and do not need to be part of the formula at all (this would also end the discussion about 0 RR exploits).

westerner

Joined: Jul 02, 2008

Posted: Jan 14, 2009 - 13:42

I think agility teams should get a slight boost to encourage greater diversity.

Protestations of "grow a pair" notwithstanding, the empirical data suggest pretty clearly that a majority of coaches prefer bashing to being bashed, all else being equal.

The suggestion of modifying the scheduler to match (non)bashers vs (non)bashers might cause segregation of the box where nonbashers rarely play bashers, reducing classic bash v. agility matchups.

Logic suggests that there is a TS tipping point where most coaches would be equally happy playing either team, as measured by actual matches played. I think this is what the Box should aim for.

_________________
\x/es

CircularLogic

Joined: Aug 22, 2003

Posted: Jan 14, 2009 - 15:12

DukeTyrion wrote:

However, in Blackbox, every game is a recovery match, since you are always playing against someone of equal TS. Therefore whether they are good at growing or not, is not very relevant.

The problem is, that the TS formula becomes increasingly inaccurate for teams that start with 10players or less. Usually 7-10 player teams will lose against most 11-12 player teams at the same TS - simply due to the prevalence of DP.
So lightly armored teams are alot more likely to slip into situations, where they are overrated and don`t have an even remote shot at winning. So looking at 2 games, bashy teams will win more if equal TS equals equal winning chances.

Even if this problem was fixed by a perfect TS:
If you can choose between a 50% winning chance while anihilating your opponent or a 50% winning chance while watching your team being destroyed, what would you pick?

I know I asked this before, but maybe this time even Johan will answer this simple question.

Ash

Joined: Feb 03, 2004

Posted: Jan 14, 2009 - 16:17

I agree there is an issue about diversity (14 last games of my DE where equaly against khem, chaos, orc, dwarves and ogres... nothing else!).
I disagree we should change the actual "elf" in any formula for encouraging more diversity. Formula is good as it is for that matter.
For more diversity you got to grow some b---s on coaches. Or categorize each race into 2 or 3 categories and then force coaches to have the same quantity of team in each category.

my 2 cents...

_________________
Ash

pac

Joined: Oct 03, 2005

Posted: Jan 14, 2009 - 16:21

I'm not at all sure that any notional 'happy medium' actually exists.

westerner talks about a tipping point where most coaches would be equally happy to play either side. That point would be hard to define, but probably does exist - but there's a problem with it. The tipping point will exist where the coach looks at the match-up and says, 'I'd be happy to play the elves because, even though they're probably going to get beaten up, they are better equipped to win' and 'I'd be happy to play the bashers because, even though they're more likely to lose, they should get a few CAS SPPs out of it.'

In other words, to aim for such a position is to aim towards building in a bias in favour of elves winning more matches - something which conflicts with the overall objective most coaches have for the box that matches should be 50-50 in terms of win chance (whether between the two teams or the two combinations of coach+team - that's a separate issue).

Individual details of the TS system can be improved, but the overall picture will remain the same: if you want a system where the draw typically gives you a 50-50 odds match-up, that will also be a low-elf environment (given free roster selection for coaches). How low? That remains to be seen, but enough time has passed that we must have useful data by now. Too low? That comes down to personal taste, but it has to be kept in mind that measures to balance the roster mix will tend either to jeopardise the 50-50 balance in match-ups or to be unwieldy and/or irritating to coaches who want the freedom to make and run the team of their choice.

I think that [B] is successfully hitting, or is on its way to doing so, enough targets (make any team you like; get a roughly even match fast; be sure that your opponent built his team in match-ups resulting from the same system as yours) without overstretching by trying to hit them all.

Unstoffe

Joined: Aug 22, 2004

Posted: Jan 14, 2009 - 16:27

Looked at the title, thought this would be a thread about how the poor bashy races are disadvantaged, due to their lack of agility and ball handling skills. How disappointing...

Re. diversity in [B], I'm firmly in the 'grow some' camp. Let's have a look at the top 30 [B] teams on TR (it being the easiest [B] list to get to Smile

)
On that page, the first and second best teams in terms of winning % are... Dark Elves and Skaven. While it would be nice to see some more such teams (so many orcs.. so many), I don't see that they need any help.

_________________
British or thereabouts? Check out the White Isle League

DukeTyrion

Joined: Feb 18, 2004

Posted: Jan 14, 2009 - 16:36

CircularLogic wrote:

DukeTyrion wrote:

However, in Blackbox, every game is a recovery match, since you are always playing against someone of equal TS. Therefore whether they are good at growing or not, is not very relevant.

The problem is, that the TS formula becomes increasingly inaccurate for teams that start with 10players or less. Usually 7-10 player teams will lose against most 11-12 player teams at the same TS - simply due to the prevalence of DP.
So lightly armored teams are alot more likely to slip into situations, where they are overrated and don`t have an even remote shot at winning. So looking at 2 games, bashy teams will win more if equal TS equals equal winning chances.

Even if this problem was fixed by a perfect TS:
If you can choose between a 50% winning chance while anihilating your opponent or a 50% winning chance while watching your team being destroyed, what would you pick?

I know I asked this before, but maybe this time even Johan will answer this simple question.

Not sure that is exactly accurate.

I have played many games with less than 11 players, and often the additional skill available (and on the pitch) mean more than the other teams 13 players, with 2 off the pitch.

I will agree however that if does change at differing TR levels. 9 Elves at 100 TR, will often win without trouble against 12 rookies 100 TR. 9 Elves at 170 TR though, will be in some trouble against a 12 player team of 170.

At the end of the day, if you want to have some adjustment to the bash vs score debate, the number is already in the formula. There is a BBR rating (factored at 0.3), if you really want to change things, make this number 0.5, 0.8 or even 1.0.

I still lean towards the TS only view, but I understand why these formula are in place, all that I am saying is, if you want to change things, look at the BBR, so that the squishy teams end up playing less games giving away 15 TS to bashers ... that'w where it gets really frstrating.

sk8bcn

Joined: Apr 13, 2004

Posted: Jan 14, 2009 - 16:46

pac wrote:

I'm not at all sure that any notional 'happy medium' actually exists.

westerner talks about a tipping point where most coaches would be equally happy to play either side. That point would be hard to define, but probably does exist - but there's a problem with it. The tipping point will exist where the coach looks at the match-up and says, 'I'd be happy to play the elves because, even though they're probably going to get beaten up, they are better equipped to win' and 'I'd be happy to play the bashers because, even though they're more likely to lose, they should get a few CAS SPPs out of it.'

In other words, to aim for such a position is to aim towards building in a bias in favour of elves winning more matches - something which conflicts with the overall objective most coaches have for the box that matches should be 50-50 in terms of win chance (whether between the two teams or the two combinations of coach+team - that's a separate issue).

Individual details of the TS system can be improved, but the overall picture will remain the same: if you want a system where the draw typically gives you a 50-50 odds match-up, that will also be a low-elf environment (given free roster selection for coaches). How low? That remains to be seen, but enough time has passed that we must have useful data by now. Too low? That comes down to personal taste, but it has to be kept in mind that measures to balance the roster mix will tend either to jeopardise the 50-50 balance in match-ups or to be unwieldy and/or irritating to coaches who want the freedom to make and run the team of their choice.

I think that [B] is successfully hitting, or is on its way to doing so, enough targets (make any team you like; get a roughly even match fast; be sure that your opponent built his team in match-ups resulting from the same system as yours) without overstretching by trying to hit them all.

This is interesting, and actually, you are right.

Seen as this, if the goal is to have 50% win proba, then, of course, low AV teams are screwed.

Actually, thinking about myself, I lost quite a few games now (ok I play pro elves in a basher environment...) + plus got quite my part of bashings. And more and more, I am thinking about switching to a team that has strength.

It's not about having balls, it's about fun.

It's not to be constantly battling hard to win, lose players you'd like to build and start anew. If that was not an issue, people wouldn't be avoiding multi-dps and C+RSC teams in R.

(By the way I have no solution to offer.)

_________________
Join NL Raises from the Ashes

CircularLogic

Joined: Aug 22, 2003

Posted: Jan 14, 2009 - 16:53

BBR isn`t the solution, at least in the way it`s incorporated at the moment. For one, it`s per coach and not per team. While per coach measuring is arguably an option for coaching skill, it`s not for bashing success, as the races differ more extremely.
Next point is that BBR is meshed together with BR and TS into one figure. But I have rambled enough on this in other threads...

I`m with you on the TS only matchmaking - though it wouldn`t help too much on the racial diversity thingy, which is discussed here. BBR will do nothing to encourage racial diversity, giving incentives might do the trick.

SillySod

Joined: Oct 10, 2006

Posted: Jan 14, 2009 - 17:09

Quote:

In other words, to aim for such a position is to aim towards building in a bias in favour of elves winning more matches - something which conflicts with the overall objective most coaches have for the box that matches should be 50-50 in terms of win chance

Actually there might be a neat way around this one...
- set the win balance to 52-48 or something which is generally slightly in favour of fragile races
- use TS to assign the matchups (with or without BR)
- use empirical data combined with TS and BR (and everything else) to calculate each sides win chances for the game, but do so only as part of the BR change calculation

This way bashy teams would generally kill the most and agility teams would generally win a little more but the BR changes would compensate for that. So the elf teams might win more but their BR would go up less for each win such that both bashy and agility teams would be equally good for raising your BR (which is the point of Blackbox).

I dont think that the fragile teams should win very much more than bashy teams but there is probably a neat little balance where neither bashy or fragile teams are too irritating to coach.

_________________
Putting the "eh?" back into Sexeh.

"There are those to whom knowledge is a shield. There are those to whom it is a weapon. Neither view is balanced."

pac

Joined: Oct 03, 2005

Posted: Jan 14, 2009 - 17:26

SillySod wrote:

So the elf teams might win more but their BR would go up less for each win such that both bashy and agility teams would be equally good for raising your BR (which is the point of Blackbox).

I may be missing some announcement somewhere (I do mean this, as I don't keep track quite as well as in the past), but isn't BR still invisible? How can you say that the point of the box is to raise a statistic whose absolute or relative value you will never know? Even if it were visible, I don't think many people would agree with you on this.

Quote:

I dont think that the fragile teams should win very much more than bashy teams but there is probably a neat little balance where neither bashy or fragile teams are too irritating to coach.

Indeed, as I said at the outset of the previous post, I do think there is such a balance. The point is that achieving that balance conflicts with the 50-50 objective - and by much more than 52/48 (if it's statistically significant enough to change coach behaviour, it will also be significant enough to annoy a lot of coaches by making games too hard/easy).

Last edited by pac on %b %14, %2009 - %17:%Jan; edited 1 time in total

westerner

Joined: Jul 02, 2008

Posted: Jan 14, 2009 - 17:31

sk8bcn wrote:

It's not about having balls, it's about fun.

Bravo.

It takes balls to say that, and I support it 100%.

@Pac
Nice to see you, I hadn't seen your posts in a while. Great summary.

@Unstoffe
I understand you are limited by availability of data, but I think any agility teams that make it to the top 30 TR are very likely expert players. Those teams probably don't need a boost, but I think the majority of agility teams are hurting a bit.

@All
In general I don't like the constriction of fixed quotas. I'd rather see some type of incentive-based tweak that pleases the most and irritates the fewest coaches.

_________________
\x/es

pac

Joined: Oct 03, 2005

Posted: Jan 14, 2009 - 17:36

westerner wrote:

I'd rather see some type of incentive-based tweak that pleases the most and irritates the fewest coaches.

I think you're right, but I don't think we necessarily need an officially-backed, compulsory initiative to do this. Groups like ELF and Rat Race are aiming (and will do more in the future to aim) at [B]. I think metagroups with their own internal rewards and kudos can do a lot to influence coach behaviour.