Thinking ahead: NTBB 2013

[MattDakka]

They didn't design for those TV matched leagues at all.

It's up to those leagues to be brave and design their own TV if they dare.

I would guess that they kept it simple for easy calculation on TT. If the computer is doing it it can be as hairy as you like.

Let's imagine a short private league (10-12 matches long) with 9 Dwarf teams and 1 Amazon team: do you think Amazons will win it roughly 50% of their matches, assuming coaches of equal skill?
If the answer is yes, then they designed well the game, if the answer is no, then their game design is heavily flawed.

[dode74]

Let's imagine a short private league (10-12 matches long) with 9 Dwarf teams and 1 Amazon team: do you think Amazons will win it roughly 50% of their matches, assuming coaches of equal skill?
If the answer is yes, then they designed well the game, if the answer is no, then their game design is heavily flawed.

Do stop making strawman arguments. It's not the specific win% in small leagues which tiers were based upon, or even the win% in individual matchups against specific races, it's the average overall win%. What you want it to be is something different entirely, as well you know because you keep making this irrelevant point and you keep being told that this was never the intention.

My point is exactly the same as Shteve0's: the data does not support the changes, so claiming that the changes are made based on data is misleading at best.

[MattDakka]

Do stop making strawman arguments. It's not the specific win% in small leagues which tiers were based upon, or even the win% in individual matchups against specific races, it's the average overall win%. What you want it to be is something different entirely, as well you know because you keep making this irrelevant point and you keep being told that this was never the intention.

What ensured that a league had a fair distribution of teams to collect data from for the overall win % when BBRC made the CRP?
No rule in BB tells that in a private league you must have a fair distribution of bash, hybryd and agile teams, so the overall win % is not reliable, because it doesn't take in account leagues with a bad distribution of teams (like in my example), yet they are regular short private leagues and not MM.
It's not a strawman, it's just an example to show how much flawed TV is.
TV should be either more accurate or relative (i.e. calculated by match up), not like it is now (TV 1000 Amazons are weaker than TV 1000 Dwarfs: I don't give a damn about the winning rate overall when I have to face only Dwarfs in a private league or even in a single game, wasting 1 hour just to lose due to a built-in match unbalance).
You keep on saying that BBRC meant to do the game this way (i.e. overall win rate % against all the other teams), but it's not the way to go in order to make a really balanced game (i.e 50 % win rate against every possible team).

[dode74]

What ensured that a league had a fair distribution of teams to collect data from for the overall win % when BBRC made the CRP?

Large amounts of data. No data collected in any amount (Plasmoid's, FUMBBL, FOL etc) has any such skewed distribution as the example you used.

You keep on saying that BBRC meant to do the game this way (i.e. overall win rate % against all the other teams), but it's not the way to go in order to make a really balanced game (i.e 50 % win rate against every possible team).

Who said the game was meant to be really balanced? If that's what you're after then maybe this isn't the game for you.

If you want balance then the TV system is flawed and would require a total rewrite, but unless you balance matchups on a per-game basis to determine inducements (which would require some serious statistical analysis to determine the relative values of each skill and in what numbers as well as the inducements themselves within that particular matchup, and would only really calculable by computer each game) then it will still not be accurate.

[Darkson]

Thankfully a 50% balanced game isn't what was intended. If you want one, I suggest you find a perfectly weighted coin, and play heads & tails.

[voyagers_uk]

c'mon Heads, it is almost always heads!!!

[VoodooMike]

And he just happened to come up with the kind of changes that you think he should have? \o/ for coincidence. Or was it gut feel.

NTBB has a LOT of varied changes. I have no doubt some of them will match up with what the data suggests should be done, and many more will not. You know what they say about broken clocks...

If you play enough games you'll have enough data. If you don't you won't. If I understand you and Dode correctly the more games the smaller your margin of error. The larger your margin of error the less confidence and the better case for the manual override.

The underlined part is where your understanding apparently failed. Not having enough data does not give you carte blanche to just make crap up and say it's based on more than making crap up.

But what i'm asking is where are your rules? What data do you have to back them up?

Part 2 of a creationist style argument? I don't have any alternate rules to offer you, so I'm not sure what I'm supposed to back the non-existent rules up with... not having the answer doesn't mean its impossible or even unreasonable to point out the error in someone else's answer. Having blanks in our knowledge is not an invitation to fill them with whatever speculation and wishful thinking you happen to have in your left pocket.

It seems to me that Amazons and wood elves are pretty under used considering that they are winningest teams around. Maybe definitions of 'good' vary.

9 out of 10 people who taste-tested what later became "New Coke" preferred it over what later became "Coke Classic", yet when they pulled coke off the shelves and replaced it with new coke people refused to drink it. Human behavior is based on perception of fact, not actual fact, so it bears little relevance when examining those facts. Rational people can get past their erroneous perceptions while less rational people are slaves to them, instead.

Given that NTBB states its goal as "narrowing the tiers", we're not really worried about what people guess the tiers of teams are, we're more worried about the specific definition of those tiers, which is straight win%s. That's data discussion, not guesstimation.

What do you guys have that is better than CRP? Where is your evidence?
What do you have that is better than CRP+? Where is your evidence?
What do you have that is better than NTBB? etc etc

Again, irrelevant to the conversation - I don't need to tell you how and why the universe was created to argue against the theory it was willed into creation by God. I don't need to solve "5 + 5" to tell you the answer is not "-4", I can just say "you never get a negative number when you add two positive numbers together".

I would guess that they kept it simple for easy calculation on TT. If the computer is doing it it can be as hairy as you like.

My guess would be they have no grasp of the long-term effect of skills or stats on win%, so it's rather arbitrary. The only way to have an accurate TV calculation that really tells you what two team's relative chance of victory is, barring coach skill, is to calculate out the actual effects of those skills and stats on win%, which thus far nobody has managed to do... if they did, then you could very easily refactor the rosters using that information to make them line up correctly (or appropriately incorrectly if you're a fan of the tiers).

Let's imagine a short private league (10-12 matches long) with 9 Dwarf teams and 1 Amazon team: do you think Amazons will win it roughly 50% of their matches, assuming coaches of equal skill?
If the answer is yes, then they designed well the game, if the answer is no, then their game design is heavily flawed.

This is not necessarily true, Matt. Consider one of the major game design paradigms that we've seen get very popular in the last decade... the "rock, paper, scissors" design. The idea is that nothing beats everything - everything is strong against something, and weak against another, forcing a form of diversification. Now, apply this to your question: "what if the league has 9 rocks, and one scissors... will the scissors win 50% of the time?" no, it'll lose every time... but that isn't a design flaw, just a compositional flaw in the league. If you had an infinite open league, you'd find their w/l/d rate would be approximately even, however... as composition imbalances would result in an influx of whatever element is missing, because for a short time that element would give people an edge, and so on.

You can make your same argument, though, by pointing out that in an environment where all teams are represented... at the same TV and within the same tier, the win% differences don't lack statistical significance... though each tier is ridiculously wide, and typically about even with that of a coin toss, unless they have a handicap, as anything but tier 1 does.

Thankfully a 50% balanced game isn't what was intended. If you want one, I suggest you find a perfectly weighted coin, and play heads & tails.

Or Chess, or Go, or some equally trivial and strategy-lacking game!

[MattDakka]

What ensured that a league had a fair distribution of teams to collect data from for the overall win % when BBRC made the CRP?

Large amounts of data. No data collected in any amount (Plasmoid's, FUMBBL, FOL etc) has any such skewed distribution as the example you used.

A skewed league can still happen. The fact that no league had such skewed distribution (a theoric example I used just to show my point) doesn't mean that such league can't exist and rules must be written considering every scenario. The fact is that, at the end of the day, playing Amazons against Dwarf is an unbalanced match, especially at low TV. "Don't worry, you will lose this but you will win against other teams" is what the BBRC says.

Who said the game was meant to be really balanced? If that's what you're after then maybe this isn't the game for you.

Nobody said that, but claimin that the game is balanced according to a silly overall win % is misleading and creates frustrating match ups.

If you want balance then the TV system is flawed and would require a total rewrite, but unless you balance matchups on a per-game basis to determine inducements (which would require some serious statistical analysis to determine the relative values of each skill and in what numbers as well as the inducements themselves within that particular matchup, and would only really calculable by computer each game) then it will still not be accurate.

I agree, but greater accuracy is better than bad accuracy. The goal should be better accuracy, even if perfect accuracy is impossible to achieve.

This is not necessarily true, Matt. Consider one of the major game design paradigms that we've seen get very popular in the last decade... the "rock, paper, scissors" design. The idea is that nothing beats everything - everything is strong against something, and weak against another, forcing a form of diversification. Now, apply this to your question: "what if the league has 9 rocks, and one scissors... will the scissors win 50% of the time?" no, it'll lose every time... but that isn't a design flaw, just a compositional flaw in the league. If you had an infinite open league, you'd find their w/l/d rate would be approximately even, however... as composition imbalances would result in an influx of whatever element is missing, because for a short time that element would give people an edge, and so on.

In rock scissors paper the win rate is perfectly even and balanced amongst them, unlike BB teams.
In a private league you are forced to use the same kind of team, you can't change it in every match (thus evening the win rate over time), so, in your example, you have scissors and your opponents rocks and you keep on playing at a disadvantage every match (the win rate doesn't even over time in this case). Every team should have roughly 50% against another team, this way every match would be balanced and interesting, regardless the race involved (and no, this doesn't mean "let's play with 6 3 3 8 linemen teams" ). From the coach's point of view I can say that it's very boring playing a bad match up between two tier 1 teams.

though each tier is ridiculously wide, and typically about even with that of a coin toss, unless they have a handicap, as anything but tier 1 does.

I agree, win brackets are too wide.

[dode74]

A skewed league can still happen. The fact that no league had such skewed distribution (a theoric example I used just to show my point) doesn't mean that such league can't exist and rules must be written considering every scenario. The fact is that, at the end of the day, playing Amazons against Dwarf is an unbalanced match, especially at low TV. "Don't worry, you will lose this but you will win against other teams" is what the BBRC says.

As Mike said, Rock Paper Scissors.

In rock scissor paper the win rate is perfectly even and balanced amongst them, unlike BB teams.

Only when there is a perfect distribution of rocks, paper and scissors.

Nobody said that, but claimin that the game is balanced according to a silly overall win % is misleading and creates frustrating match ups.

If nobody said it then why are we arguing about it? The game is balanced within the tiers as laid out. Given that you understand those tiers then what is misleading?

I agree, but greater accuracy is better than bad accuracy. The goal should be better accuracy, even if perfect accuracy is impossible to achieve.

Then a rewrite of TV is what is needed. That's not within the scope of NTBB though, as I understand it.

[MattDakka]

As Mike said, Rock Paper Scissors.

(I believe) I explained the difference between Rock Paper Scissors and playing a BB team. You can't change your team when the league is in progress, unlike Rock Paper Scissors (you can change every round).

Only when there is a perfect distribution of rocks, paper and scissors.

Yes, and over time the odds will even up.

The game is balanced within the tiers as laid out. Given that you understand those tiers then what is misleading?

Using the word "balanced" might make it sound balanced, but it's not really balanced (by single match up), as I pointed out. Private leagues are generally short, while overall win rate works well with a great number of matches and a fair distribution of teams. You can have the latter, but in a short private league you can't have the former, thus I pointed out the relevance of the balance by match up and not by overall win rate.

Then a rewrite of TV is what is needed. That's not within the scope of NTBB though, as I understand it.

Of course. I never claimed that NTBB had such goal.

[plasmoid]

Hi all,
heck of a time to have a birthday (the missus) and sick kids on top of that - what with an angry mob at the gates and all
I do intend to reply to all of this, but getting around to it is proving tricky.

I shoulf probably say that I never claimed to be creating Perfect Bowl.
I'm merely trying to improve on CRP - or rather to make a move in a direction that some will find to be an improvement.

But I'm certainly not in a position to make said 'Perfect Bowl'.
*I'm not speaking from any position of authority, so I can't make sweeping changes (without being ridiculed/ignored). If a made a massive rewrite of the TV-system, for example, I doubt I could get anyone to test it.
*I have no means of getting the ammounts of data that seems to be required to make Perfect Bowl , (and even if I did, as VoodooMike said I would still be forced to make aducated guesses, because in a complex game like BB you can't isolate any one factor anyway).

I've been foolish enough to mention percentages in this thread, but in others I have emphasized that I would never be able to generate the kind of data to check that anyway.

My goal is to make more teams and strategies viable.
One way is to weaken the impact of CPOMB.
The second is to take the most powerful teams and weaken them a bit, without accidentally pushing them through the bottom of tier 1.
The third is to take the tier 2 teams, and buff them enough that they become as powerful as a low end tier 1 team.
The 4th is to take the tier 3 teams and buff them enough to make them better, without accidentally pushing them into tier 1 - i.g. they will still feel weaker than the other teams.

I think such nerfs/buffs could be done in a number of ways without breaking the game.
Case in point: Ogre team. 10K off the ogres. Better statlines for snots.
Is there anyone who will claim that such a move could make the team worse by mistake?
Could these changes have no effect? Literally no effect in any games? Hardly.
Pure buffs will improve the team. How much? Good question. But my gut (ooooh, he said it) along with feedback from the playtesters tells me, that they still feel like a less than tier 1 team. Ace! And unfortunately I will never be able to get enough games playd to get a statistically significant ammount of data.

Now - before you get your pitchforks out, rest assured I'll post more. Then you can get your pitchforks
Cheers
Martin

PS (Shteve0, I'll get to you first. You seem quite upset, and some of the things you write I just don't recognize).

[dode74]

Using the word "balanced" might make it sound balanced, but it's not really balanced (by single match up), as I pointed out.

As I've said already, it's not intended to be balanced by single matchup. If you don't like it...

[koadah]

But what i'm asking is where are your rules? What data do you have to back them up?

Part 2 of a creationist style argument? I don't have any alternate rules to offer you, so I'm not sure what I'm supposed to back the non-existent rules up with... not having the answer doesn't mean its impossible or even unreasonable to point out the error in someone else's answer. Having blanks in our knowledge is not an invitation to fill them with whatever speculation and wishful thinking you happen to have in your left pocket.

Is anyone doing anything else?

No one is forced to use it. If anyone likes it good for them

[Darkson]

As Mike said, Rock Paper Scissors.

(I believe) I explained the difference between Rock Paper Scissors and playing a BB team. You can't change your team when the league is in progress, unlike Rock Paper Scissors (you can change every round).

Who said every round? Oh wait, you did, to try and prove your disproved statement.

Then a rewrite of TV is what is needed. That's not within the scope of NTBB though, as I understand it.

Of course. I never claimed that NTBB had such goal.

Then why are you arguing about it in the NTBB thread? If you want to argue/discuss it, start a thread about it and do so with those interested in it.

[MattDakka]

Thankfully a 50% balanced game isn't what was intended. If you want one, I suggest you find a perfectly weighted coin, and play heads & tails.

Err... thanks for your suggestion but I want a luck management, tactical turn based game with a fantasy theme.
Coin tossing is just sheer luck, it lacks thinking, luck management, tactics and fantasy theme.

Who said every round? Oh wait, you did, to try and prove your disproved statement.

Ok, so, let's consider 1 match only: playing a league match with Amazons against Dwarfs has the same odds of winning a Rock Paper Scissors round for you? I said "every round" because you play more than one match in a private league. A match is like a Rock Paper Scissors round.
The overall win rate doesn't take in account league composition, that's a design flaw.
You can break the overall balance win rate simply with a bad league composition, I know that my example is extreme, but I used it in order to make more evident my point.
Balance teams by match up and each league, no matter the composition, will be balanced.

Then why are you arguing about it in the NTBB thread? If you want to argue/discuss it, start a thread about it and do so with those interested in it.

Apòlogies for the off-topic.