harvestmouse
Joined: May 13, 2007
|
 Posted:
Feb 04, 2013 - 22:12 |
 
|
Hi, this thread is to iron out an exact formual for play testing (and to see how feasible an idea) of my theory on match making in a perpetual environment.
This isn't official in anyway, but it's how I plan to run teams in private leagues in the future.
The idea is to match teams by their success more than their TV. All new teams of 1000TV would have a score of 0, and success/failures new players/injuries affect this score.
The formula in my mind is the following.
Fan Factor
1 point of FF = 1 point
Win/Losses
Each win in your last 5 matches = 1 point
Each draw in your last 5 matches = 0 points
Each Loss in your last 5 matches = -1 point
Each win in your previous 5 matches (6-10) matches ago = 0.5 points
Draws 0 points
Losses 0 points
Changes in TV since your last match
+15-50 TV 0.5 points
+50-100 TV 1 point
etc
-15-50 TV -0.5 points
-50-100TV -1 point
etc
Handicapping/inducements obviously not possible on FUMBBL at the moment, however possibly in the future. As a ball park figure I'll say:
1 point = 100 GP
I thought about taking some teams as an example of how they would be affected by this formula, however they wouldn't look like that in this environment, so it's a little pointless.
Anyway, your thoughts. |
|
|
Overhamsteren
Joined: May 27, 2006
|
| Posted:
Feb 04, 2013 - 22:30 |
|
Fan factor will often make up something like 50-80% of the points?
Not saying that's good or bad.  |
_________________
Like a Tiger Defying the Laws of Gravity
Thanks to the BBRC for all the great work you did. |
|
harvestmouse
Joined: May 13, 2007
|
| Posted:
Feb 04, 2013 - 22:47 |
|
Yes, if your TV and win/loss ratio is balanced and your TV is constant, it then comes down to mainly fan factor.
However the more you win the harder your match ups, the more you lose the easier your match ups.
If you have had a large change in fan factor, this all changes things also. |
|
|
happygrue
Joined: Oct 15, 2010
|
| Posted:
Feb 04, 2013 - 23:15 |
|
It's interesting to think about such systems, but it is also very hard to come up with something that isn't abused. Setting aside stuff like starting teams with 7 FF, here is an example that comes to mind.
Rookie ogres (5 ogres, 2RR TV: 960) win two games but get no skills from them. They buy an apo after the second game, so TV increase after the second game is 60K (apo + 1FF)
FF -> 2
TV change -> 1
wins -> 2
so these guys have a rating of 5, with 1030 TV of ogres snots RR and no skills.
Now suppose some rookie wood elves (1000 TV) play their first game, skill up a WD with tackle and win it. Then they play another game and skill up the other WD with tackle and lose the game. They then buy an apo.
These WE have rating:
FF -> 1
TV increase (50K+20K)-> 1
wins -> 0
So we have a WE team with tackle on both wardancers and a rating of 2, with TV 1100
Under the current system, this would be a horrible matchup even with the 70K of inducements that ogres are getting. But under the new system, the WE get 300 gp of inducements... so wiz and babe and apo on top of a stacked matchup. If it's 50 gp per point of rating then that's still 150 in this case. And if it gets smaller then we can dream up other situations heading the other direction where teams of VERY different situations get no inducements at all.
This isn't to say there isn't merit to improving the matchmaking system, but I'm just trying to point out that as such things get more complicated the ease of abusing such a system is something to think about. |
|
|
harvestmouse
Joined: May 13, 2007
|
| Posted:
Feb 04, 2013 - 23:45 |
|
Hmmmm Ok, plenty there for me to mull over there.
2 points though.
1st one that teams will be required to start with 0 FF, starting with anything else causes a system imbalance of course.
2nd one is that this system does and is meant to give successful teams harder and harder match ups. Losing, brings them back to parity, if they've had a few lucky results.
I'd like to see it as sort of 'divisional' each area is like a division with promotions and relegations. And like such a system the matches should be harder the higher you go. |
|
|
Nelphine
Joined: Apr 01, 2011
|
| Posted:
Feb 05, 2013 - 00:06 |
|
I think the biggest issue that jumps out at me from this is that you are penalizing people for choosing stats when they skill up. You are also penalizing people for the game after their recovery game.
Example: You are playing skaven, and you get smashed to shreds, with your TV dropping by 1000, all from non retiring MNGs. This means the next game you are playing at -10 points. The game after you are playing at +10.
Obviously it won't usually be that big, but if the score is calculated each time for each game, then the skaven, in recovery, will be playing either really bad teams, or really new teams. Which is good. But the game after their recovery, they'll be up around 23-32 points, which seems like it might be so high that they don't get matched with anyone, or if they do, they'll be giving away massive handicaps, when in reality they should be back where they were previous to the recovery game. |
|
|
harvestmouse
Joined: May 13, 2007
|
| Posted:
Feb 05, 2013 - 00:15 |
|
I don't think your first point is that much of a point (stats and doubles). Changes in TV generally don't make much difference.
The second point is valid, I was thinking of adding changes in TV from the match up previous (50%), but didn't want to for making the system too complex. Possibly running it the same as the last match up.
I think that's probably better. |
|
|
Synn
Joined: Dec 13, 2004
|
| Posted:
Feb 05, 2013 - 01:03 |
|
I like this as the premise to a supplemental Rating system (or CR even).
The idea is, games 'closer' by scores should be higher rated.
But matchmakers need to be first and foremost a matchmaker.
__Synn |
|
|
koadah
Joined: Mar 30, 2005
|
| Posted:
Feb 05, 2013 - 01:23 |
|
|
Muktar
Joined: Dec 02, 2012
|
| Posted:
Feb 05, 2013 - 03:09 |
|
To deal with the mng's and other things, here is my idea (numbers maybe are semi-random).
FF=1 pt
Each earned skill or stat= 1pt
Each win/loss in the last 6 games = +1/-1
Each player that earned a mng or dead from last game = -1
This would put the focus on pitting built teams against built teams who have won. Would even help recovering teams to still only play teams in which they still match up in skills and/or player depth. Like I said, the numbers I stated are semi-random and not meant to be anything but a talking point. |
|
|
|
|
Gnome Box ranking pa...
Search
