New rating system - suggested values

[Bjorn]

....this seems to put a lot more work on the GM's shoulders.
-Raw

Any attempt to implement a rating system will require some additional work on the part of the GM. For that reason we should try to keep it simple. It should not be too bad if all they keep track of is when a player is RIPed and who did it to them.

[Raw]

Any attempt to implement a rating system will require some additional work on the part of the GM. For that reason we should try to keep it simple. It should not be too bad if all they keep track of is when a player is RIPed and who did it to them.

I agree with this Bjorn! That would be simple.

I just think that if the GM had to do a +.0001, -.0003 for each player each turn, it would get pretty sickening for the GM, who is just out to have a little fun .

-Raw

[Saladin]

I tend to agree with Bjorn and Raw. It should be kept simple for the GM. Though i would still like to see a form of negative bonus for players that go M-3. So maybe not have bonusses for sending in turns, but just a little 'punishment' for abandoning a game.

We should also try and think of something that would counter the fact that somebody who surrenders in the end game gets treated differently than somebody who gets ripped, to me it's the same.

[gm_al]

VPs are taken into account for lots of reasons, but why do they have have to matter for this rating system? Why treat WOK4 and WOK5 differently? The plan is to maintain separate WOK4 and WOK 5 ratings and average them. My feeling is that a win in WOK4 should have the same impact on your rating as a win in WOK 5.

WOK5 takes way more effort then WOK4, and this should reflect somehow, just like VPs. Maybe increase all the numbers a bit for WOK5 games.

.... your rating will fall below 0.

Its clearly stated in the rules that rating can go from 0.000 - 9.999. Of course if a Player would fall under 0.000 you would get some inbalace in the rating sum, but I doubt anyone is around long enough to 'achieve' this.

Players RIPed on the same turn should pool and split points to avoid having the OOP influence this.

100% d'accord.

Now I can live with that ratings as well, IF a few things happen:
- WOK5 gets some higher rating changes then WOK4
- RIPping is increased to +/-0.02
- everything is put clearly into some tables and on a webpage

M-3 could be included in the system. Just give the M-3-Player a -0.02 rating and, to keep the sum equal on both sides, split the 0.02 as bonus for the Players still in game. (or something similar)

GMs will still need to track rating changes every Turn, in a column on the Group page, to keep the workload low.

Of course I can already hear people and 'good' Players screaming once they get ganged up early on in game by 2 or 3 aggressive neighbors and end up as 10th or 9th....

And this will not stop people to miss a Turn or two just because they are lazy. The final conclusion is, as previous threads have shown, that you can NEVER have a perfect system that pleases EVERYONE I guess. Some will always shout "Unfair !"

So what shall it be, Players ?

[Saladin]

We're talking about the system proposed by Bjorn right?

So:

+.10 for 1st
+.06 for 2nd
+.04 for 3rd and 4th
-.04 for 4th through 8th
-.05 for 9th and 10th

What i have a bit of a problem with is that the gap between 1 and 3 (0.6) is smaller than the gap between 4th and 5th (0..

Why not have a simple sliding scale:

1: +0.10 (for a single win)
2: +0.07 (so +0.85 for a shared win)
3: +0.05
4: +0.03
5: +0.01
6: -0.01
7: -0.03
8: -0.05
9: -0.07
10: -0.10

Now something that i'm also a bit concerned about is that all games will drag on longer because if one or two players are far stronger they will continue till they have ripped all remaining players to get as much bonus as possible (i know i will). So if a player surrenders after a vote that should be considered a rip as well, so there won't be too much unnecessary rounds played.

What do you think?

[Underdog]

I would guess the winners split credit for the Rip at the end?

[Saladin]

Good point Underdog! Yes, they would split the credits for the extra rips at the end.

[Aussie Gaz]

Is it possible somehow to weight the points for a RIP based on the two players ratings?

This would require that everyones rating is either recorded at the start of each game or the system is updated continuously (after each turn of every game).

Player A has rating of 2.000
Player B has rating of 1.000

If A RIPs B - A gets 0.005 B loses 0.005
If B RIPs A - A loses 0.02 b gains 0.02

My old rating system was based on 50 points for each vp, -20 for going M3 and +-20 for RIPs.

[Bjorn]

Is it possible somehow to weight the points for a RIP based on the two players ratings?

My old GMT rating system would determine the average KVAL for the entire group and the number of rating points won or lost for each player would vary based on their rating relative to the average. Actually, the way it worked was that I created a pool of 200 points for WOK4 and 300 points for WOK5 from the players in the game. Players with a higher rating would contribute more points to the pool. So, if you were below the average you would win more points if you won and lose fewer points if you lost. It doesn't work well for this system because the numbers we use are too small. It was still a zero sum system.

Two comments on your chart, Sal. Keep in mind that we are now considering a +/- .02 rating points for RIPs. If the winners are going to get RIP points for the survivors, then the survivors would each lose an additional .02 rating points. On the assumption that the winners may have RIPed one or two players along the way, you are going to see the same problem we have now, where the winners get all the rating points and the 3rd through 5th place finishers barely break even. It may turn out that the rating points reward RIPing people much better than a third or 4th place finish. Is that what we want?

In addition, I don't want to see the first player out punished too much. I would prefer to see the first three or four players out share the pain.

1: +0.10 (for a single win)
2: +0.07 (so +0.85 for a shared win)
3: +0.05
4: +0.03
5: +0.01
6: -0.00
7: -0.03
8: -0.07
9: -0.08
10: -0.08

No one loses more than .1 rating points in a game. Even the 6th place finisher can break even if he RIPs someone. The first 3 players out of the game take about the same damage, which is OK by me.

[Bjorn]

As a note to everyone, we can initiate this process without agreeing on the exact points awarded or lost. If we agree to keep it simple, then GMs can add one or two columns to their group pages and all the information we need can be collected from them and recorded. We can play with the numbers later to see what makes sense.

If the only things that affect rating are RIPing players and final position in the standings, then all the GM needs to do on the group page is to show what turn a player is RIPed and the player who RIPed them. As gatekeeper I always go to the group page when a game ends and update my records with the information there. I can just record the information and we can decide what to do with it later.

Keeping track of the number of turns each player submits orders or misses the turn puts an added burden on the GM which I don't believe will be consistantly recorded. That is the primary reason I don't want to see it made part of any rating system.

[gm_al]

I am against considering Players that havent been RIPped as such and give the winner(s) additional credit they dont deserve.

Id like to set a deadline for putting things into motion, as several games are about to kick off, and more will come to the list shortly. Sunday ?

I am positive to have Bjorns system to be put into place, as long as RIPs are worth +/-0.02 - yes, RIPping should always be a real reward, as it will make games much more lively. We can use his latest suggestion of values too, seems good.

Open points I still see are:
-1- tweaked values for WOK5 ? yes/no ?
-2- going M-3 has no effect on ratings.... yes/no ?

Im not too convinced on weighing the individual pre-game ratings into the whole thing - it makes calculations more complicated and games are usually set up in a way that you play with players of the same skill then yourself (hopefully). Keep it simple.

Please put the rules and values onto some linkable website....

[Saladin]

It may turn out that the rating points reward RIPing people much better than a third or 4th place finish. Is that what we want?

Ehm...yes. What does it matter that you were the #3 player with 600 points left and therefore beat the #4 player with only 200 points left. If this would mean that we'll see more ripping of players that's only good for all the action. Personally i'd rather see a player who rips 2 players and then finishes 8th get more points than a player who sleeps the whole time and finishes 4th.

In addition, I don't want to see the first player out punished too much. I would prefer to see the first three or four players out share the pain.

1: +0.10 (for a single win)
2: +0.07 (so +0.85 for a shared win)
3: +0.05
4: +0.03
5: +0.01
6: -0.00
7: -0.03
8: -0.07
9: -0.08
10: -0.08

Ok, this looks good as well although this wouldn't solve the 'problem' of the numbers 3 and 4 getting a lot less than the number 1 and 2 if they get a negative bonus for getting ripped/surrendering. Actually by having to make sure that the end total sums up to 0, you have the problem that if you want to reward good play you have to extra punish the lesser players. For instance i'm fully in favour of giving a bonus for ripping a player, but i don't think there should be a negative bonus for getting ripped. In the end everybody except the 2 winners gets ripped.

If i understand correctly the reason for having the end total sum up to 0 is so that the ratings don't go through the roof? But it doesn't actually stop that because a good player will see his rating rise and rise, creating a big gap with new players who have played less games. So i don't think that we actually need to have a balanced end total. Or are there other reasons to have it like that Bjorn?

[korexus]

If i understand correctly the reason for having the end total sum up to 0 is so that the ratings don't go through the roof? But it doesn't actually stop that because a good player will see his rating rise and rise, creating a big gap with new players who have played less games. So i don't think that we actually need to have a balanced end total. Or are there other reasons to have it like that Bjorn?

Not quite. The reason for havng the total being a constant 0 is so that people's ratings don't go through the roof just by playing. A good player will always have his score rise, but with out the 0-sum element an average player who plays a lot of games will get the same thing. In theory, a player who is exactly average in the game should keep his score at 1 so length of time playing is only an advantage if you learn from it.

On a related note. This system would result in newby's scores dropping initially as they would almost certainly be below the "average" level of WoK play. So it may not be as encouraging as could be hoped...


korexus.

[Underdog]

There is a way to fix that and that is to not rate beginner games. It would mean we would have to provide beginner games on a regular basis but it would keep the beginners from getting discouraged too much to start. it would also allow them to maybe figure out how this system works if they put some time into it.

[Dameon]

Wow, y'all have sure been busy during my trip out here. The rating system is hardly a new idea, but it seems to me the reason that it hasn't ever been effected is because it would mean a fair amount of extra work for GMs, who are doing this as volunteers after all. I really don't understand what is wrong with keeping it with a simple VP system. In any case, when I play, I already know my opponents and how good they are, and a rating of some sort really has no practical value. The only value in it is for a person's ego really, and we already have Kaohalla for that. It feels like this would be a GREAT deal of work for some ego-stroking, if you ask me.

I don't think that a ratings system would harm Kaomaris by any means but I just do not feel that any benfits from it would be worth the work that goes into it, unless Al or Matt could come up with a program that does this automatically. Putting effort into developing new games like WOK 6 or MW would have more of a positive effect on the community as a whole, it seems to me.

[Mullog]

There is one other possible way to avoid the ratings rising too much: We could let the ratings be for a limited period, say one year or two. Or a limited number of games.
If we were to keep track of the ratings manually this would be a horrible system, but as soon as the kaobase2 is ready (very soon! hehe) this can be handled automatically. This also mean that we need not worry about the complexity of the system. There will be some extra work for the GM even with the new kaobase, but this will be limited to registering when a player is RIP'ed and who did it.

While I am at it. Have we disbanded Hannibals suggestion about a new VP system? I agree most of his suggestions, and completely with his reasoning. WOK needs a better way to encourage new players to stay, and changing the VP system could help.


Mullog

[trewqh]

The only value in it is for a person's ego really, and we already have Kaohalla for that. It feels like this would be a GREAT deal of work for some ego-stroking, if you ask me.

The idea behind this is that ego-stroking will make more newbies and other players stay in Kaomaris. You can't argue with that.

I say we should also put a naked woman on every webpage within and associated with kaomaris.com! That would help even more

[Sorry, couldn't resist ]

trewqh

[trewqh]

Or at least change that ElBarachmi warrior sketch into something prettier

t.

[gm_al]

Yes, ratings are meant to be some kind of extra-reward that should encourage people to stay with us and keep playing. Like an extra reason to keep trying to get better. Thats why Im also against not rating beginner games, in fact ratings are just right for them (even if some will start with losing some rating points)

What I suggest is to put some rules and values in effect and re-evaluate them when the 2K5 champs start next January. I am against totally dropping something that Players have fought to build up over months.

And stop joking about the "effort" it will take a GM to keep track of everything. Its just adding up one or two numbers every Turn, putting it into a new rating column and handing over the final results after the game to the Gatekeeper (who will then enter it into the new KaoBase V2). Not really much of an effort, its more work to post in here if you ask me.

And while we can discuss this endlessly (and will come to no result that will please everyone) I have set the deadline for the finalisation of the rating rules for this Sunday. We will take what we have established by then and try it for this season. If it REALLY pi**es off too many people we can always learn from the experience and drop or modify it.

PS: To stir up the discussion a bit, how about the GM sets a rating level requirement for people that want to sign up for his game....?

[trewqh]

PS: To stir up the discussion a bit, how about the GM sets a rating level requirement for people that want to sign up for his game....?

With the amount of players and games we have I think this is pointless.

Games would take waaay too long to fill.

trewqh