94 coaches online • Server time: 22:45
Forum Chat
Log in
Recent Forum Topics goto Post Designer's Comm...goto Post Why did GW nerf guar...goto Post Request to Maintain ...
SearchSearch 
Post new topic   This topic is locked: you cannot edit posts or make replies.
View previous topic Log in to check your private messages View next topic
Purplegoo



Joined: Mar 23, 2006

Post   Posted: Apr 19, 2011 - 11:18 Reply with quote Back to top

There is also the point (looking at Garion and anisdrin's posts regarding high TV, bashing and 'balance') that the bashers have all their toys already. Remains to be seen when the agility teams get their main weapon (i.e. the Wizzie) if that will encourage more Tackle on heavy bashers, changing that playing field a little, more coaches to run Elves (although, looking at the ’winningest’ teams on BB central, it seems winning races don’t attract lots of coaches all the time) or more high TV losses for the heavy bashers. It is a bit unfair to lean on the Claw/MB/PO thing when a good 2+ Zap! Would be a big difference maker in many of the games we see at high TV. Too often I seem to see high TV Elves / more fragile teams just load up on Apos to live to fight another day, perhaps the magical option would skew the coaches of those teams more toward winning.
Carnis



Joined: Feb 03, 2009

Post   Posted: Apr 19, 2011 - 11:26 Reply with quote Back to top

PorkSol wrote:
Yet I have only played Orc vs Orc 5 times in 164 games.

Its because you are lower TV than most orcs, orcs tend to get higher TV quickly.

My unlimited Nurgle team has played 210 games (oldest nurg in the box)!

The top10 matchups:
Dwarf 28 20/4/4
Chaos 22 13/5/4
Orc 19 13/5/1
Undead 16 11/3/2 | Necrom 16 6/4/6 (shared 5th)
CDwarf 14 9/3/2
Nurgle 13 9/2/2
D Elf 12 7/5/0
Norse 11 9/2/0
Wood Elf 10 8/2/0 | Khemri 10 5/4/1 (shared 10th)

There seems to be bias towards chaos and not nurgle, but I find the # of undead more surprising to be honest Wink.
Smeesh



Joined: Oct 29, 2005

Post   Posted: Apr 19, 2011 - 11:37 Reply with quote Back to top

Quote:
Or, sack of the scheduler and just do random pairings. Everyone likes inducements!


Fine for me. Just don´t overstretch enducements. Morg ist enough Wink .


Gretings

Smeesh
Kelkka



Joined: Aug 14, 2009

Post   Posted: Apr 19, 2011 - 11:43 Reply with quote Back to top

The Murker said:
Quote:
I'm using observations.

Quote:
The solution is simple, and I'll type extra slow. Activate all the teams you want to play. The BOT then looks ONLY at the TV numbers presented to it, and gives every coach a match-up based on the closest TV number match to one of their teams.


Sorry to tell you, but that's not any more "fair" than the current system.
If you like observations over statistics then there is little I can tell you, but in case you are interested there is this
statistics page that has win% of races in
different TR-ranges. The data was collected in old LRB4 time, and AFAIK they are collecting new FFB data now that is being used
(just not shown to us yet).

That particular page tells you for example, that TR150-170 skaven wins 65,72% of games against equal TR ogre team.
Paring teams only based on equal TV is not a silver bullet to make it all fair suddenly.
As you are a new coach, I'd advice you to stay away from Box for a while and play ranked. Box is a nasty place and
it does actually say "This division is not for newcomers to the game." in division description when you create a new team.
The_Murker



Joined: Jan 30, 2011

Post   Posted: Apr 19, 2011 - 11:49 Reply with quote Back to top

Ehlers: I appreciate your at least taking the time to look into the data available a bit before posting. I submit the page(s) you are looking at include a great deal of pre-LRB6 teams. I also submit the scheduler is running on data collected pre-LRB6. LRB4 is no more, and this is a new game, requiring new data. At the very least we could ditch all of the invalid LRB4 data collected and gather new data of racial match-ups before we alter match-ups with it. I would submit that investigating active Chaos teams would revel that they face other chaos based teams rarely. I looked at RandomOracle's team last week and he had only played another Chaos team once in 50 games. Hmmmm.

PGoo: I am not surprised an elf player (I'm guessing you are usually) does not feel the same CRUSH of Chaos. I am pleasantly surprised when I play elves, and get many interesting matches. (this is at sub 1700 TV thus far) I notice the imbalance when I play humans. More importantly, Chaos notices and is benefiting in that they don't face other Chaos. (often enough) This is the part that is hard to swallow. If everyone and their dog is going to jump on this busted-ass clawpomb bandwagon, everyone should get a taste of what box has become, Chaos included. Nice to see you think random should be random though. Mirror being boring is an opinion of course.

I have no problem facing my FAIR share of claw. I have no problem playing a team that gets smacked around in box. I have a problem with an UNFAIR share of claw coming my way, and a Chaos player getting a pass instead of a taste of his own medicine.

I'm still waiting for a valid argument against "TV alone" pairings.

_________________
Image
Join the wait-list. Watch the action. Leave the Empire. Come to Bretonnia!
Hitonagashi



Joined: Apr 09, 2006

Post   Posted: Apr 19, 2011 - 11:49 Reply with quote Back to top

For my oldest Box lizzie team entirely under new client(65 games), my opponents are:

Dwarf 11 4/1/6
Nurgle 8 6/0/2
Orc 7 3/1/3
Chaos Dwarf 6 5/0/1 | Necromantic 6 4/0/2
Chaos 5 4/1/0 | Norse 5 1/2/2
Chaos Pact 4 4/0/0 | Undead 4 2/1/1

It matches up fairly well to what Carnis has with his Nurgle in terms of racial balance.

[Edit] PS. Hates the stunties I does! Hates them! [/Edit]
RandomOracle



Joined: Jan 11, 2004

Post   Posted: Apr 19, 2011 - 11:58 Reply with quote Back to top

The_Murker wrote:
I also submit the scheduler is running on data collected pre-LRB6.

I would submit that investigating active Chaos teams would revel that they face other chaos based teams rarely. I looked at RandomOracle's team last week and he had only played another Chaos team once in 50 games. Hmmmm.



Is "submit" a synonym for "make stuff up"?
The_Murker



Joined: Jan 30, 2011

Post   Posted: Apr 19, 2011 - 12:04 Reply with quote Back to top

Kelkka, maybe you don't get it. Your stats page is being used by the scheduler for what? To keep Ogres from being paired against Skaven too often if there is a match-up they might have a better chance at winning? Boo-hoo, lets help the Ogres have more fun. NO! You said it, box is a nasty place. Let's not help the Ogres. Don't help anyone. And let's NOT help Box be a warmer, fuzzier place for the nastiest Chaos teams out there. Those efforts to make things "more fair" are largely responsible for box being so nasty, IMO. Chaos will pair itself down if given the chance. I don't want the pairing systems help. It's not helping anyone but bashers avoid bashers.

Yes, yes pyth, I said warm fuzzy box.

_________________
Image
Join the wait-list. Watch the action. Leave the Empire. Come to Bretonnia!
BillBrasky



Joined: Feb 15, 2005

Post   Posted: Apr 19, 2011 - 12:05 Reply with quote Back to top

I think the "scheduler is broken" argument may be based mostly on feeling. I think a true analysis would be exceptionally time consuming, as it seems like the small number of teams activating at any time are the major factor determining who your team will play against. Certain time frames lend toward certain coaches playing, and of course each coach enjoys using certain teams... There are just so many variables.

I'm all for a [H]ard Core Blackbox that is completely random with maybe a 600 TV difference limit or something, and nothing based on race, coach, etc. This would obviously not be for the faint of heart...
The_Murker



Joined: Jan 30, 2011

Post   Posted: Apr 19, 2011 - 12:07 Reply with quote Back to top

4 in the last 50 games. I stand corrected. Well done RandomOracle. Must be nice.

_________________
Image
Join the wait-list. Watch the action. Leave the Empire. Come to Bretonnia!
Carnis



Joined: Feb 03, 2009

Post   Posted: Apr 19, 2011 - 12:09 Reply with quote Back to top

With the amount of 4 coach draws, a change to 20min scheduler may also be a good idea for variety.. Just a thought. The murker is wrong about what is being used for matchups though, its all LRB6 data.
RandomOracle



Joined: Jan 11, 2004

Post   Posted: Apr 19, 2011 - 12:15 Reply with quote Back to top

The_Murker wrote:
4 in the last 50 games. I stand corrected. Well done RandomOracle. Must be nice.


4 chaos, 7 nurgle, 5 chaos pact. Yes, it's been a nice and cuddly ride for me, as my team used to be TV 2580 and now they're at TV 1740. Clearly, high-TV chaos teams never face other teams that can bash them.

Btw, my team has faced much more varied opponents lately now that they've dropped in TV. At high TV, it's pretty much just CDs, nurgle, chaos, orcs, and dwarves.
Christer



Joined: Aug 02, 2003

Post   Posted: Apr 19, 2011 - 12:20
FUMBBL Staff
Reply with quote Back to top

Right then. Some facts for you:

- There is a 0.97 factor for same race matchups in the suitability calculator (3% penalty)
- This penalty was added because mirror matches are generally seen as less interesting
- There is a racial factor in the suitability score. This works using a BWR type calculation and looks at race vs race in different TV brackets (huge amounts of data points). This factor is strictly based on LRB6 races/matches and does NOT use old LRB4 data.

To get an idea of how many same race matches have been played, I gathered complete (LRB4 as well) data and generated a graph on it. Basically, it's two graphs showing the average number of matches against each opponent race (Effectively # of matches played total) and the number of matches against the same race. These two values are on the Y-axis. X has the different races, sorted after average number of games, descending.

Image

#1 are Chaos, #2 are Orcs, #10 (the spike) are Ogres.

The same distribution only counting LRB6 is:

Image

Here, the top2 are Chaos and Nurgle

As you can see, the # of same-race matches seems to be roughly exponentially correlated. Since the number of matches played for a given race is directly related to the number of teams available for that race, this correlation is absolutely expected.

The 3% dampening isn't overpowering the fact that the most played races play more against themselves than anything else. Bottom line is that I don't see a major issue here.

(Edited to show Total vs LRB6-only stats)


Last edited by Christer on %b %19, %2011 - %12:%Apr; edited 4 times in total
thedevil



Joined: Jul 31, 2009

Post   Posted: Apr 19, 2011 - 12:21 Reply with quote Back to top

Ive played a few games with the scheduler in favour of humans, maybe thats because I play a range of races. I think that in general the scheduler needs looking at in regard to inducments and thier wieght in TV.
The_Murker



Joined: Jan 30, 2011

Post   Posted: Apr 19, 2011 - 12:37 Reply with quote Back to top

Christer: Thanks for taking the time. Quick question. Do you give any consideration to the idea of removing the 3% penalty? I'm guessing near 70% would like it removed based on the recent poll. Perhaps a poll on mirror being "less interesting."


Last edited by The_Murker on %b %19, %2011 - %12:%Apr; edited 1 time in total
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   This topic is locked: you cannot edit posts or make replies.
View previous topic Log in to check your private messages View next topic