56 coaches online • Server time: 23:41
Forum Chat
Log in
Recent Forum Topics goto Post Creating a custom to...goto Post ramchop takes on the...goto Post NBFL Season 32: The ...
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
DukeTyrion



Joined: Feb 18, 2004

Post   Posted: Oct 30, 2014 - 13:55 Reply with quote Back to top

Just had another weird matchup, from the scheduler;

https://fumbbl.com/p/blackbox?op=scheduler&r=2014-10-30+13%3A30

The current scheduler is built around getting the 'best for most' even if it means shafting 1 coach;

In this example, am I the only one that sees this;

Total Suitability: 3160

TV 1000k Chaos (0)
TV 910k Halfling (0) @866

TV 1040k Norse (1)
TV 1140k Skaven (4) @854

TV 2000k Nurgle (929)
TV 1740k Orc (148) @762

TV 1460k Orc (56)
TV 1750k Dark Elf (35) @678

As better than this;

Total Suitability: 3254

TV 1740k Orc (148)
TV 1750k Dark Elf (35) @995

TV 1000k Chaos (0)
TV 910k Halfling (0) @866

TV 1040k Norse (1)
TV 1140k Skaven (4) @854

TV 2000k Nurgle (929)
TV 1460k Orc (56) @539
Christer



Joined: Aug 02, 2003

Post   Posted: Oct 30, 2014 - 13:56
FUMBBL Staff
Reply with quote Back to top

JimmyFantastic wrote:
It depends on the % applied.
And the intention is to give soft teams easier teambuilding than cpomb teams, in terms of opposition faced at least.


No, the intent is to increase diversity in the environment. The suitability padding would be placed on elves should they end up being dominant.
Christer



Joined: Aug 02, 2003

Post   Posted: Oct 30, 2014 - 15:09
FUMBBL Staff
Reply with quote Back to top

DukeTyrion wrote:
Just had another weird matchup, from the scheduler;


Let's look at this from the scheduler's perspective. What it comes down to is which of these two sets is better:

A: 1740 vs 1750 (@995) -- 2000 vs 1460 (@539)
B: 1460 vs 1750 (@678) -- 2000 vs 1740 (@762)

or, in TV difference:

A: 10 and 540
B: 290 and 260

The suitability formula is based on win probability using an ELO based formula, which inherently compares TV difference.

So, let's look at the win probabilities of these TV differences (for the lower value team):

A: dTV 10 => 49.2% dTV 540 => 0.7%
B: dTV 290 => 7.4% dTV 260 => 9.6%

The scheduler chose A here, and you could roughly say that's because 49.2 + 0.7 > 7.4 + 9.6

You could also say that the point where the scheduler chooses to "shaft" a coach is when TV difference causes the win percentage to drop below 25%, which happens roughly at dTV 150.
Calcium



Joined: Apr 08, 2007

Post   Posted: Oct 30, 2014 - 15:22 Reply with quote Back to top

koadah wrote:
Do you basher boys have no stones?

A person might think that you like to dish it out but can't take it.


Think what you like, but the same ppl that are now crying about clawPOMB etc. are the same ppl that cried about LRB4 DP.
Simply put, their definition of blood bowl is one without the blood, because they can't stomach their pixels being RIP'd. If coaches like you get their way that will damage this site far more than clawPOMB ever will.....

_________________
Image
DukeTyrion



Joined: Feb 18, 2004

Post   Posted: Oct 30, 2014 - 15:46 Reply with quote Back to top

Christer wrote:
DukeTyrion wrote:
Just had another weird matchup, from the scheduler;


Let's look at this from the scheduler's perspective. What it comes down to is which of these two sets is better:

A: 1740 vs 1750 (@995) -- 2000 vs 1460 (@539)
B: 1460 vs 1750 (@678) -- 2000 vs 1740 (@762)

or, in TV difference:

A: 10 and 540
B: 290 and 260

The suitability formula is based on win probability using an ELO based formula, which inherently compares TV difference.

So, let's look at the win probabilities of these TV differences (for the lower value team):

A: dTV 10 => 49.2% dTV 540 => 0.7%
B: dTV 290 => 7.4% dTV 260 => 9.6%

The scheduler chose A here, and you could roughly say that's because 49.2 + 0.7 > 7.4 + 9.6

You could also say that the point where the scheduler chooses to "shaft" a coach is when TV difference causes the win percentage to drop below 25%, which happens roughly at dTV 150.


Christer, I understand the logic, from an overall point of view it's a better 'group' set of matchup and I appreciate (and am amazed by) all the work you do on the formula.

But, is it not possible to find some level of protection somewhere, where perhaps even the other 3 games took place, but the game with < 1% didn't?

I realise we do not want to go down the path of multiple coaches not getting scheduled, but this is another area of frustration for many.

Of course, it would mean that coaches like myself would be more affected, as I normally only activate 1 team, so it would be another tweak to the benefit of the multiple team activaters, which surely could only be a good thing?
Endzone



Joined: Apr 01, 2008

Post   Posted: Oct 30, 2014 - 16:18 Reply with quote Back to top

I think the work Christer puts into this is amazing. Personally, I would prefer there to be a minimum threshold because I don't think a 0.7% win probability for the underdog based on TV is a match up worth having.
Mr_Foulscumm



Joined: Mar 05, 2005

Post   Posted: Oct 30, 2014 - 16:27 Reply with quote Back to top

Calcium wrote:
koadah wrote:
Do you basher boys have no stones?

A person might think that you like to dish it out but can't take it.


Think what you like, but the same ppl that are now crying about clawPOMB etc. are the same ppl that cried about LRB4 DP.
Simply put, their definition of blood bowl is one without the blood, because they can't stomach their pixels being RIP'd. If coaches like you get their way that will damage this site far more than clawPOMB ever will.....


Honestly, this thread is actually one of the few semi-productive and less whiny threads on the subject that we have in the forums. I'm not sure if all the tough guy posturing will be helpful in that regard. Smile

_________________
Everybody's favorite coach on FUMBBL
JackassRampant



Joined: Feb 26, 2011

Post   Posted: Oct 30, 2014 - 16:31 Reply with quote Back to top

Would it have been better if those two teams just didn't get a match?

_________________
Veni, Vidi, Risi
PainState



Joined: Apr 04, 2007

Post   Posted: Oct 30, 2014 - 16:54 Reply with quote Back to top

Mr_Foulscumm wrote:
Honestly, this thread is actually one of the few semi-productive and less whiny threads on the subject that we have in the forums. I'm not sure if all the tough guy posturing will be helpful in that regard. Smile


I think it is refreshing because it is a break from the drone of sensible ideas being floated around. I like good ole fashion chest pounding and taunts, reminds me we are playing BLOOD BOWL after all.

_________________
Comish of the: Image
PaddyMick



Joined: Jan 03, 2012

Post   Posted: Oct 30, 2014 - 17:32 Reply with quote Back to top

If I was playing Orcs 1460tv Vs Nurgle 2000tv i'd take:
2x babes, 2x wandering apo, 3 x merc linemen*
-if i wanted survive with most of the team intact, then it's off to the next game
It's only if this happens to you all the time that you wouldn't want to play box

*I wouldn't really. 3 bribes, a saw and a DP merc would be my actual choice
Balle2000



Joined: Sep 25, 2008

Post   Posted: Oct 30, 2014 - 17:45 Reply with quote Back to top

PaddyMick wrote:
3 bribes, a saw and a DP merc would be my actual choice

I'd spec that

_________________
Join the SWL
Image
Get your team bios here!
Putting the romantic in necromantic since 2010
Endzone



Joined: Apr 01, 2008

Post   Posted: Oct 30, 2014 - 17:57 Reply with quote Back to top

Take what inducements you want, then roll your D100 and see if you can get over 99.3!
Dunenzed



Joined: Oct 28, 2011

Post   Posted: Oct 30, 2014 - 20:42 Reply with quote Back to top

Endzone wrote:
I think the work Christer puts into this is amazing. Personally, I would prefer there to be a minimum threshold because I don't think a 0.7% win probability for the underdog based on TV is a match up worth having.


I think some more data on actual results at TV differences over 15% is really needed. I'm not 100% on the scheduler math but I think that 0.7% win probability for an underdog at extreme levels might be understating their chances.

I had a quick look at the Wild Dragons looking for games with more than a 15% TV difference. I saw 4 games in this category, but one of those was an FC game last year so before the scheduler change and nothing to do with the current issues that prevent this team from being played.

Of the 3 others, one was a loss vs High elves, where the Dragons were the overdog by 280TV.

The next two were wins, one vs chaos at 850TV gap, and one vs nurgle at a 340TV gap.

So thats 3 match pairings outside 15% in about 20 games played with this team since the scheduler change, but the only one they lost was when they were the overdog.

Granted this is anacdotal evidence from a coach that is much better at this game than me, and is playing one of the top tier races, so those factors (along with luck) obviously influenced the outcome more than TV.

And maybe the coach hated beating a chaos team playing up 850TV, and would have had more enjoyment beating a TV matched chaos team.

But my point is solid data on TV differences and match results would be of interest. To what extent is this perception a gap of more 15% TV creates an unfair game actually reflected in the match results. Is it more than the advantage certain racial pairings might deliver?
Dominik



Joined: Oct 29, 2004

Post   Posted: Oct 30, 2014 - 21:30 Reply with quote Back to top

Calcium wrote:
koadah wrote:
Do you basher boys have no stones?

A person might think that you like to dish it out but can't take it.


Think what you like, but the same ppl that are now crying about clawPOMB etc. are the same ppl that cried about LRB4 DP.
Simply put, their definition of blood bowl is one without the blood, because they can't stomach their pixels being RIP'd. If coaches like you get their way that will damage this site far more than clawPOMB ever will.....


No, I did not cry about Dirty Player. There was IGMEOY and there was the disadvantage to give up your position if you want to remove a player by fouling properly which is by assistance of team players, these things are not necessary now with CPOMB plus you can only foul once per turn but you can use your CPOMB eleven times per turn theoretically.
harvestmouse



Joined: May 13, 2007

Post   Posted: Oct 30, 2014 - 21:46 Reply with quote Back to top

Yes, but you were most famed for playing Khemri in lrb 4. DP to Khemri lrb 4 is what eucalyptus is to a Koala bear.

Fouling in lrb 4 was too strong. It needed tweaking, but without gtr it wasn't too bad. Claw/RSC was worse but there weren't that many of these in number.

BBRC are probably right in that you need something to balance the cas chart and lrb 4 and CRP had/have it around the right number. The problem is if it's all coming from one avenue, that's the problem.

What I'd like to see looking into Serious Injuries.

1. Players are forced to miss more than 1 game.
2. Players take perm injuries that do not result in retirement. Right now that's a problem. What's the point in a perm injury when most educated coaches count it as a kill anyway?

Saying that, I'm surprise why most coaches retire new niggle players outright.
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