51 coaches online • Server time: 16:23
Forum Chat
Log in
Recent Forum Topics goto Post Creating a custom to...goto Post Secret League Americ...goto Post DOTP Season 4
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
pac



Joined: Oct 03, 2005

Post   Posted: Dec 04, 2008 - 10:21 Reply with quote Back to top

The good point worth discussing that did emerge out of the other thread: basing (B)BR by team and not by coach. This would be an invisible and non-restrictive change that would address quite a few concerns. I'm still not convinced it's necessary as such, but it warrants its own separate discussion.

Some relevant posts (think I got most of them):

Symmetrical wrote:
westerner wrote:
However, said coach also has other teams that are trying to win. And I object to the scheduler assigning lower-TS matchups to his other teams as a result of losses with his "killer" team.

I'm not sure if this particular point has been addressed, forgive me if so.

I think westerner raises a fair point. I don't have a problem with anyone in [b] playing any particular style with whatever race they see fit, but I can see a potential problem in the situation that westerner describes. If someone has multiple teams in [b], and is tanking with one team to gain an advantage with another, I don't think this is positive for the division.


pac wrote:
... a solution would be to attach BR (and BBR) to the team, not the coach, so that one of a coach's teams could not get easier games thanks to the poor performance of another.

The new problem introduced by this solution is that a coach could reset his ratings by starting a new team. This in turn could be solved by giving any coach's new team starting ratings based on the average from his existing teams.


PhrollikK wrote:
... a collective BR for the coach could [also] be calculated from these for cosmetic reasons (bragging rights).

Right now if you have 4 Bashers teams and one elf team then the elf teams are forced to (probably) play basher teams if the other 4 teams have caused lot of net carnage :/


westerner wrote:
BR and BBR need to have a per-team component so that coaches cannot metagame one team's losses to offset another team's wins.
JanMattys



Joined: Feb 29, 2004

Post   Posted: Dec 04, 2008 - 10:27 Reply with quote Back to top

The point of Christer (if I remember well from a conversation we had in channel blackbox) is that indicators such as BR and BBR only work with big numbers due to their mathematical nature (he used the specific word but I don't know and I don't recall it).

Using them for teams would make them useless. At least that's what Christer said to me when I questioned him about the same problem.

_________________
Image
PhrollikK



Joined: Nov 04, 2006

Post   Posted: Dec 04, 2008 - 10:34 Reply with quote Back to top

JanMattys wrote:
The point of Christer (if I remember well from a conversation we had in channel blackbox) is that indicators such as BR and BBR only work with big numbers due to their mathematical nature (he used the specific word but I don't know and I don't recall it).

Using them for teams would make them useless. At least that's what Christer said to me when I questioned him about the same problem.


For statistical reasons you need a large sample of data to be able to draw trends. Yet a given team has to be matched by it's innate values, otherwise the matching is not that even.

Personally I would love to see the BBR value calculated as Difference in CAS caused/received divided by number of games played. In that way a team that causes alot of net CAS per game will be slightly nudged towards playing other CAS-causing teams.

_________________
Visit my MiniCity
Mega Punk!
Malicious Rex, Daemon Adoptable
pac



Joined: Oct 03, 2005

Post   Posted: Dec 04, 2008 - 10:37 Reply with quote Back to top

Jan, that is why I would suggest that a new team should have (B)BR values based off all that coach's past games, thus providing a sufficiently large number.

Thus, if a coach has just two teams, each of which plays lots of games in a quite different style (and perhaps with a different degree of success) from the other, they could end up with quite different ratings. But if a coach is making and retiring several teams per week, all their ratings will be roughly the same.
DukeTyrion



Joined: Feb 18, 2004

Post   Posted: Dec 04, 2008 - 10:51 Reply with quote Back to top

For the moment I would just be happy to see BR published.
Shane70



Joined: Oct 03, 2008

Post   Posted: Dec 04, 2008 - 12:15 Reply with quote Back to top

i'd love to know my BR for Tr@itor

Smile
Wreckage



Joined: Aug 15, 2004

Post   Posted: Dec 04, 2008 - 13:27 Reply with quote Back to top

JanMattys wrote:
indicators such as BR and BBR only work with big numbers ).


i highly favor the idea to make it team dependend.....

about the points being made:

1. using more teams will provide more statistical data but much less accurate
2. i dont see even really an exploit in creating a new team to reset the own rating.... we shouldnt forget that br and bbr should have a rather small impact on the game anyways....

i talk too much mathematical stuff so i put this in a quote wrote:

i dunno how br is exactly calculated but assuming its only basing on win/loss percantage a good way to deal with this would be to apply a significance value... for example a team that has played one game may have won or lost 100% of its games but the data is 0% accurate... if it has played two games the data is 50% accurate... three games 66% accurate... four games 75% accurate and so on... ie ... 1- 1/n(games)...

now lets say a team has (i dunno) 900 suitability... and the BR would add 100 BR points to the suitability..... but there are played only two games....a 50% accuracy is multiplied and it adds only 50 BR.... or removes only 50BR .... or nothing if its 1/0/1

the sifnificance wouldnt need to be 1 - 1/n tough


even tough we say these aspects to balance the league into fair matchups its reasonable to carefully increase the difficulty rather slow then too hasty...


the basic problem is that a coach will make very different expierences if he chooses to change the team he is playing with... every race is played in an individual style... if you play on a competetive level, you will try to advance the team basing on whats best with its racial abilities and extend your playstile to chosen skills...this playstyle may favor scoring, cas'ing, fouling or other stuff but will alyways be more dependend on the race then coach.....


looking at my expierences in ranked i gotta say... people arent really looking on the coachrank but almost everyone respects a team with a 20/0/1 winning stat...
Grod



Joined: Sep 30, 2003

Post   Posted: Dec 04, 2008 - 14:35 Reply with quote Back to top

Question:

Does your bashing rating accumulate, or is it only based on currently active teams? For example if I play 10 games as Khemri, retire the team, then start with a skaven team, do I still have a high bashing rating?

_________________
I am so clever that sometimes I don't understand a single word of what I am saying.

Oscar Wilde
Wotfudboy



Joined: Feb 17, 2004

Post   Posted: Dec 04, 2008 - 14:56 Reply with quote Back to top

Grod wrote:
Question:

Does your bashing rating accumulate, or is it only based on currently active teams? For example if I play 10 games as Khemri, retire the team, then start with a skaven team, do I still have a high bashing rating?


I believe the answer is "Yes you do" - it is all coach based and not team based as far as I am aware.
westerner



Joined: Jul 02, 2008

Post   Posted: Dec 04, 2008 - 16:19 Reply with quote Back to top

JanMattys wrote:
The point of Christer (if I remember well from a conversation we had in channel blackbox) is that indicators such as BR and BBR only work with big numbers due to their mathematical nature (he used the specific word but I don't know and I don't recall it).

Using them for teams would make them useless. At least that's what Christer said to me when I questioned him about the same problem.

Thanks for that input, Jan.

I think it's difficult to propose specific per-team weighting changes since we don't know yet how BR/BBR are being calculated. But it is useful to discuss what results/outcomes would be the most fun.

Wreckage wrote:
the basic problem is that a coach will make very different expierences if he chooses to change the team he is playing with... every race is played in an individual style... this playstyle may favor scoring, cas'ing, fouling or other stuff but will alyways be more dependend on the race then coach.....

Wreckage raises a good point. Different teams from the same coach will have different styles. For example, coach X is famed as a fouler, but his Woodie team will likely cause much less CAS than his Khemri (or even another coach's Khemri). I think it's good fluff to match successful teams (whether by Blood or by Bowl) against each other based on their individual team records.

I picture [B] as a "Gladiator"-type league where your team shows up and gets "booked" by the promoter into whatever match will generate the closest contest, since that's how they attract their fanbase.

_________________
\x/es
Gatts



Joined: Jun 18, 2004

Post   Posted: Dec 04, 2008 - 17:42 Reply with quote Back to top

wasnt the point of BBox that we shouldnt encourage cascausing teams to play other cascausing teams but to encourage any and all kinds of matchups. Personally I just dont get what we are trying to achieve with BBR. Some teams use force to win others use mobility, should we add a Mobilityboxrating as well?

I can see the point of the regular BR, since its intent is to create more even games between sperienced and newer coaches, but what exactly is the intent of the BBR?
To make sure that bashy races have to play up against elfs? To make sure bashy races only play other bashers? Isnt this how it is in R isnt this what we wanted to get away from in B?

Ill try to be openminded and understanding since the box is still in its aplha stages, but i think that before we try to determine how and in what way the BBR should be applied to teams I think we should think about what it is we want BBR to acccomplish. In my view I can't even understand what it is we want this BBR to do and why just bashing out of all the different statistics that could be used merits its own subranking. Here are some other BR's that could have been used instead: Blackpassrating, BlackDodgerating, um, my point is that bashing is a means to an end. And that end is already meassured by the BR.

If we want to target people who bash wihout aiming to win, then what do we think we'll accomplish by this meassure, do you think a 4 mummy 12 dp khemri team that isnt interested in winning will care the least bit if your elfs are so skilled and so many that he cant possibly win the game. He can still Make a foul and a blitz / turn all game long. All the BR does is makes life more difficult for the bashier races, and they seriously dont need that. Stop racial discrimination in the box!
(written sort of in a hurry so my appologies on the inevitable typos and oddities)

_________________
Players die, touchdowns are forever!
Hogshine



Joined: Apr 04, 2007

Post   Posted: Dec 04, 2008 - 17:52 Reply with quote Back to top

@Gatts, BBR was intended so if a team plays solely to destroy the other team (eg Khemri with blanket DP that foul every turn and ignore the ball unless the pitch is clear) then it won't play softer teams. I personally think I agree with you, in that I don't see the point in having BBR in the formula. I know (at least I think I do) what it's there to achieve, but I feel it won't do it very well and the collateral effects of it outweigh the possible benefits.

When I play my Khemri against elves, I will probably lose. However, the challenge of playing Khemri is to win those games. Is it hard? Yes. Is it impossible? Not even slightly. (I know, my Khemri don't have a great record, but in two games...)
JanMattys



Joined: Feb 29, 2004

Post   Posted: Dec 04, 2008 - 17:53 Reply with quote Back to top

Gatts wrote:
wasnt the point of BBox that we shouldnt encourage cascausing teams to play other cascausing teams but to encourage any and all kinds of matchups. Personally I just dont get what we are trying to achieve with BBR. Some teams use force to win others use mobility, should we add a Mobilityboxrating as well?

I can see the point of the regular BR, since its intent is to create more even games between sperienced and newer coaches, but what exactly is the intent of the BBR?
To make sure that bashy races have to play up against elfs? To make sure bashy races only play other bashers? Isnt this how it is in R isnt this what we wanted to get away from in B?

Ill try to be openminded and understanding since the box is still in its aplha stages, but i think that before we try to determine how and in what way the BBR should be applied to teams I think we should think about what it is we want BBR to acccomplish. In my view I can't even understand what it is we want this BBR to do and why just bashing out of all the different statistics that could be used merits its own subranking. Here are some other BR's that could have been used instead: Blackpassrating, BlackDodgerating, um, my point is that bashing is a means to an end. And that end is already meassured by the BR.

If we want to target people who bash wihout aiming to win, then what do we think we'll accomplish by this meassure, do you think a 4 mummy 12 dp khemri team that isnt interested in winning will care the least bit if your elfs are so skilled and so many that he cant possibly win the game. He can still Make a foul and a blitz / turn all game long. All the BR does is makes life more difficult for the bashier races, and they seriously dont need that. Stop racial discrimination in the box!
(written sort of in a hurry so my appologies on the inevitable typos and oddities)


The original reason for BBR is that without it to "balance" the BR, we will have a situation where teams that don't want to win get easier and easier matchups.

Chaos killer team first match: it loses causing mayhem.
Second game: it loses again causing more mayhem
Third game: it loses again causing more mayhem.

If a system only counts for BR, the team will get easier and easier matchups because, you know, it's a crappy team with a crappy winning record...
Coach it enough times, and you will get a team that will grow stronger and stronger by *losing*, as long as it is able to get 2 casualties per game to balance the loss in terms of fan factor.

In the end, you have a team with a record of (say) 5/10/30, but very healthy (because it gets paired against inferior teams who try to win... and often those teams won't be able to fight back in terms of damage) and much much higher in TR and TS than it should have been able to reach while playing "serious" games.

When you finally see the 250/230 Chaos monstrosity with 5/10/30 wtl record playing consistently +25 TS, you realize that the "only BR and no BBR" system contributed to the creation of that monstrosity.

...which becomes a team able to take on all honest blackboxers in terms of dealing damage. Especially if the coach persists in his goal of ignoring the ball and only playing for destruction and mayhem.

BBR is there to avoid the spiral... counterbalancing BR, BBR makes it so that the killer team will get harder matchups the more it deals damage, somehow limiting the spiralling effect of easier and easier matchups based on WTL record.

_________________
Image
arw



Joined: Jan 07, 2007

Post   Posted: Dec 04, 2008 - 18:17 Reply with quote Back to top

Hear hear!
Team based is fine.
Races should play a bigger role, too-
for instance for new teams (# of games=data=0).
Just TS? Never ever- that'd be an exploit competition. Glad the discussion isn't as bold anymore.
Team/Race (B)BR would be brilliant!
I am tired of people whining about me playing Orcs as I am tired of overpowered Orc matchups.
I never asked for easy matches with Orcs. That's a waste! Orcs are powerful. All races are different in strength anyway and balancing is futile. Point is: Let 'em prove their strength in appropriate matchUps!

Actually I'd wish for a bright future without a dominating TS influence...
TS is a nice indicator but not a solid criteria. Or maybe just too solid.

Flexible concepts like BR that are influenced by gamestats are way better. Auto-adjusted to exploits if the formula works well.

Btw: Although I am sure that there are quite some problems with the scheduler I haven't faced them too often. Just two matches (of 32) were somewhat ridiculous. One in my favor one against me.
Thanks Christer- you created a brilliant BBox. Even as it is now it is easily the best devision on fumbbl.
nin



Joined: May 27, 2005

Post   Posted: Dec 04, 2008 - 20:36 Reply with quote Back to top

pac wrote:
The good point worth discussing that did emerge out of the other thread: basing (B)BR by team and not by coach. This would be an invisible and non-restrictive change that would address quite a few concerns. I'm still not convinced it's necessary as such, but it warrants its own separate discussion.

1) Why invisible?
Is there any important reason for that?
(you can allready see the cas ratio of the team)

2) There could still be Coach BR/BBR, no need to have only one thing or the other.
(we allready have team records, win%, cas. ratios...)
(teams could start with Coach BR/BBR attached and then grow their onw... I supose that this is close to what you meant with the average thing)
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic