'Fixing' domination of certain teams in cyanide

[dode74]

Ah yeah, forgot about the "fix".

It's more a semi-reset than a full one. Salary is reduced to max 1150 and you can keep up to 3 players of a max value up to 600TV.

[Chris]

Yes the lack of stars in Cyanide is really bad for some teams. I couldn't face using skaven in the OCC without their stars as too many games your roster is busy MNG'ing. Compounded by their journeyman implementation that stops you 'saving' for a star.

I think the OCC multi-league system also stops the kill teams from affecting the top leagues. They don't win as much so don't get promoted as much. I could get a kill chaos team with every beast having clpomb and warrior clmb, but I wouldn't win many games. Worse each team I shred will simply strengthen my opponents in the league as they get a weaker team to beat next matchday.

[dode74]

They don't win as much so don't get promoted as much.

I'm not so sure about that. Win percentages are roughly the same for the T1 teams. I think that the bash teams are less streaky than the agi teams, though, and I also think that while consistency keeps you at a high level it's streaks that win the titles.

[RoterSternHochdahl]

The occasional AG4 Beastman will win Matches for you. Chaos are a very versatile Team when run over Long distance.

Regarding Aging: Cyanide has a great Chance that aging will improve the playing experience in their open leagues. I am pretty sure though that it won't do the same for private leagues without forced resets. Depending on the exact ruleset they use it will just lower the average Maximum TV in those leagues. Depending on which Teams are particularly good at that TV (and keeping in it despite aging) the structure of the league will Change. Yet, what you achieve is not a necessarily balanced league, it is only a different race (or two or three) bullying the others.

[fidius]

Regarding the game's ultimate problem of teams only being good at certain TV ranges (and certain ones taking over at 1800+), my own theory is that the issue is mainly one of skill effectiveness.

There are a handful of "tier 1" skills that are obviously better than the others: Block, Dodge (and by extension Tackle), Guard, Mighty Blow, Piling On, and Claw. Teams that can take those skills on an unlimited basis dominate once they acquire them. Teams that reach a cap on new "tier 1" skills tend to peak at that TV.

So for example Chaos and Nurgle can take as much of those skills as they like except Dodge, and have AV9 to boot. Amazons and Norse start with spammed core skills which is what makes them dominant early, but run out of options quickly with only 4-5 Strength positionals. Meanwhile their competitors acquire Tackle, and/or only need 16 SPPs (POMB) to de-pitch an AV7 player every 3 turns on average (55%+ x 58%) using only the blitz, which is only a third of all blocks in a typical game.

The right answer is to equalize the effectiveness of skills across the board to the extent possible, try to lessen the impact of the rock-paper-scissors effect of Tackle/Dodge, and slow down development of Elf teams who benefit most from easy Dodge. (Tackle's a huge problem imo, as stunty teams are the collateral damage to the ruleset's attempt to slow down Dodge synergy on AG4/AV7.)

But the realistic answer is to do an artificial hack like Cyanide is proposing, without attempting a fix on the ruleset. Which I suppose we should be thankful for.

[Fassbinder75]

The right answer is to equalize the effectiveness of skills across the board to the extent possible, try to lessen the impact of the rock-paper-scissors effect of Tackle/Dodge, and slow down development of Elf teams who benefit most from easy Dodge. (Tackle's a huge problem imo, as stunty teams are the collateral damage to the ruleset's attempt to slow down Dodge synergy on AG4/AV7.)

But the realistic answer is to do an artificial hack like Cyanide is proposing, without attempting a fix on the ruleset. Which I suppose we should be thankful for.

I disagree, your suggestions would make for a bland and unsatisfying experience – aiming for perfect balance ends at best ends up with everything looking the same. Games should aim for perfect imbalance instead, and I think Blood Bowl has it down relatively well – which could explain its continued popularity. Obviously, continued manufacturer support would also give us roster tweaks and 'new' that would change the Blood Bowl meta - but we can't have everything...

In my opinion, changing rosters is far more desirable in terms of avoiding unintended consequences than diluting skills is. Tackle is a problem because it’s spammed out of the box on two teams that already have Block and thus stunty sides get a double-whammy. Lose the Tackle, lose the issue.

[Chris]

Regarding the game's ultimate problem of teams only being good at certain TV ranges (and certain ones taking over at 1800+), my own theory is that the issue is mainly one of skill effectiveness.

There are a handful of "tier 1" skills that are obviously better than the others: Block, Dodge (and by extension Tackle), Guard, Mighty Blow, Piling On, and Claw. Teams that can take those skills on an unlimited basis dominate once they acquire them. Teams that reach a cap on new "tier 1" skills tend to peak at that TV.

Really? I have always thought it is basically starting stats. AG4 or STR4 in large amounts on a team mean they an make far better use of skills. Stat increases are far more TV inefficient so if you imagine an 'ideal build' of stats and skills, those that start with higher stats are cheaper when they reach it.

[dode74]

Really? I have always thought it is basically starting stats. AG4 or STR4 in large amounts on a team mean they an make far better use of skills. Stat increases are far more TV inefficient so if you imagine an 'ideal build' of stats and skills, those that start with higher stats are cheaper when they reach it.

I tend to agree. More useful (read TV-efficient) skills are available to teams with greater skill access and fewer (useful) starting skills, meaning level-ups remain TV-efficient for longer. Since stats tend to "cost" less on initial builds than they do on rolled stats, and are obviously far easier to get (since you already have them!), this increases the TV-efficiency of a high ST or AG team (or MA, to an extent) even further.

[fidius]

AG4 or STR4 in large amounts on a team mean they an make far better use of skills. Stat increases are far more TV inefficient so if you imagine an 'ideal build' of stats and skills, those that start with higher stats are cheaper when they reach it.

Yes, both you and dode are correct, if you prefer to think of it that way that's fine. The "sweet spot determinant" is the TV range where skills are most effective in combination with stats. AG4 and Dodge is a clear synergistic combo, as is ST4 and Guard/MB/PO, AV7 and Block/Dodge, MA7+ and Dodge. The point is that once a team starts losing the arms race in terms of adding these synergistic skills to their statlines, they decline in effectiveness.

I disagree, your suggestions would make for a bland and unsatisfying experience – aiming for perfect balance ends at best ends up with everything looking the same. Games should aim for perfect imbalance instead, and I think Blood Bowl has it down relatively well – which could explain its continued popularity.

Your statement that perfect balance necessarily implies sameness is only true in the extreme, but nobody is suggesting that (see chess). I'm not talking about "perfect" team balance, I'm talking about trying to balance TV:skill_effectiveness ratios a bit better, without adjusting skill values (that strikes me as too complicated for players, and doesn't fit into the advancement system very well). Skills that continue to have vastly different effects and advantages would still exist. Sub-tier skills would then be tried, new combos discovered & used, new strategies would rise to the surface. I think BB needs a refreshing infusion of new abilities too, perhaps quicker player development (new ways to get SPP), incentives for teams to develop in diverse but effective and interesting ways. Nothing about this idea is bland and unsatisfying in my eyes.

The complicating factor (to refer back to Chris/Dode's concerns) is that, yes, statlines make skills more or less useful depending on the player/team. Thus perfect balance is impossible and sweet-spot TV ranges would still exist even if we dumbed down the skill canon. However we can get part way there by means of initial skill allocation on rookie players, so it's not a total loss to try, imo.

[VoodooMike]

Here's how you fix public (matchmaking) leagues:

1) Matches arranged by closest games played, but with no hard limit on difference.
2) Inducements determined via TVPlus, limit on individual inducements removed.
3) "resurrection" format - no injuries or deaths carried past the end of a match.
4) Treasury capped at 1M, no SE. (gold wouldn't mean much anyway)

Ta da, Blood Bowl is completely balanced across all rosters without making any changes to the rosters themselves, and without changing how skills work or other on-pitch rules.

Of course, that doesn't mean everyone would be happy with the change, but we can confidently say that not everyone would be happy with ANY change anyone suggests, and unlike every other suggestion, we know with 100% certainty that this succeed in its goal.

[dode74]

1) Matches arranged by closest games played, but with no hard limit on difference.

I thought you said that was a relatively poor predictor of a fair match?

• Given this, there does not seem to be a strong case for matching by “games played” being a superior method for matchmaking than matching by TV similarity. If anything, it would result in less equal match-ups. - https://voodoomath.wordpress.com/2013/0 ... -analysis/

[MattDakka]

1) Matches arranged by closest games played, but with no hard limit on difference.
2) Inducements determined via TVPlus, limit on individual inducements removed.
3) "resurrection" format - no injuries or deaths carried past the end of a match.
4) Treasury capped at 1M, no SE. (gold wouldn't mean much anyway)

1) It should work, as long as you have a great playerbase, on FUMBBL I doubt it would work due to small userbase, there are some teams with over 1000 games played, these would struggle to find a game.
2) I have no idea on TVPlus + Inducements, maybe it would work;
3) No injuries or deaths? Then I think people would play oneturner teams in this case, I think playing vs teams with one or more oneturners would become boring after a while, the same goes for playing vs stat-freaks team (WD ST 4, AG 5, ST 5 Bull centaurs with AG 3 Blodge etc.);
4) Treasury capped at 1M is something I thought too;
about no SE, well, since there is no long term attrition and so no need to replace often players, why remove them?

[VoodooMike]

I thought you said that was a relatively poor predictor of a fair match?

It absolutely is a poor predictor of match fairness within the context of the existing TV-based system. Under TVPlus you can match any way you choose and the matchings will eventually be very difficult to predetermine outcome for so long as TVPlus is used for inducements. The longer a team has played under TVPlus, the more accurate TVPlus is at rating that team, and at generating the required amount of inducement money to balance the team against any other team of any TVPlus level... so in that respect, games played works just fine - it gives teams time to generate a relatively accurate TVPlus rating, and then lets TVPlus do its thing.

If you were to use TV alone for inducements, or not use inducements at all, then games played would be a terrible method for choosing matches, as the numbers you linked demonstrate. It is specifically because TV is a mediocre predictor, and that inducements fail to bridge the gaps in TV differences, that games played is increasingly inaccurate. Under TVPlus the reverse is true.

TVPlus used for matching would work fine, it would just minimize inducements... and when doing so isn't necessary, why wouldn't we let inducemnets run free? I like the idea of increasing the variation seen in matches so long as it does not create painful imbalances, and that's the primary difference in the matching criteria decision.

It should work, as long as you have a great playerbase, on FUMBBL I doubt it would work due to small userbase, there are some teams with over 1000 games played, these would struggle to find a game.

They would not struggle to find a game. As the first point says, there is no hard limit on difference, so 1000 games played can still be matched with anyone, if there's nobody better available. It's also important to note that these points are not to be taken individually - on their own, none of these points will solve anything... only in conjunction with each other would they be guaranteed to do the job.

I have no idea on TVPlus + Inducements, maybe it would work

You can know if it would work by answering a simple question: If I gave you infinity gold to spend on inducements, could you beat the highest TV team you can think of with a gang of snotlings? If yes, then there exists a point between zero and infinity where you have enough inducements to make it a match where either side has an equal chance of winning, depending on how well the coaches manage to adapt to the changing tides of the game. In part, this is why the limits on each inducement would be removed... the limit is more or less silly, and if you had a reroll every turn, 16 wizards, every special play card, and every star player, your snotling team would be pretty nasty.

No injuries or deaths? Then I think people would play oneturner teams in this case, I think playing vs teams with one or more oneturners would become boring after a while, the same goes for playing vs stat-freaks team (WD ST 4, AG 4, ST 5 Bull centaurs with AG 3 Blodge etc.);

Again, the teams those monster teams would be playing against would be just as nutty as them, either though high development or through masses of inducements. Maybe you face a team of insane one-turning elves.... and maybe you drop a fireball on the closest elf every single turn. I suspect those matches wouldn't be boring at all, they'd be slapstick comedy. To avoid that type of thing you'd just have to engage in TV management, as needless TV bloat would go right into the other guy's pocket before each match began.

about no SE, well, since there is no long term attrition and so no need to replace often players, why remove them?

It serves no purpose, and at high TV levels it would discourage coaches from engaging in controlled TV management by way of firing bloaty players to give themselves some breathing room. With a 1M treasury cap you'd have what you needed to rebuild your team at any time.

[dode74]

Fair enough. Which leads me to ask if any criteria for matching is necessary at all: could you just match randomly and let TVPlus do its job? Presumably this would be less good with newer teams, so perhaps a "games played" filter - once you've played 30 then you match randomly, for example?

Also, wouldn't this lead to (eventually) teams will all level-7 players? Personally I don't think that'd be a lot of fun (ymmv), especially the lack of development, but others might. Firing the occasional player in order to level up a rookie could be an exercise in farming stats, too.

[VoodooMike]

Fair enough. Which leads me to ask if any criteria for matching is necessary at all: could you just match randomly and let TVPlus do its job? Presumably this would be less good with newer teams, so perhaps a "games played" filter - once you've played 30 then you match randomly, for example?

You can match using anything or nothing, though certainly a performance based measurement will not be very good until the thing being measured has had time to perform... which is why TVPlus also includes TV (which is a crappy measure of team strength, but it's better than nothing, especially before performance has had a chance to kick in). Matching based on games played gives time for performance to be measured. After a few games it won't matter much who someone is matched against, or how.

Also, wouldn't this lead to (eventually) teams will all level-7 players? Personally I don't think that'd be a lot of fun (ymmv), especially the lack of development, but others might. Firing the occasional player in order to level up a rookie could be an exercise in farming stats, too.

I imagine there would be some people who would play until they had a team of players that absolutely could not develop any more, sure. It wouldn't make those teams any more likely to win each match than if it were a team of all level 3 players. Past a certain level of development the ratio of TV-to-effectiveness is going to drop off, but the amount of effectiveness that the bloat provides your opponents will not, making it the best strategy to streamline and optimize rather than develop ad infinitum.