CircularLogic
Joined: Aug 22, 2003
|
 Posted:
Aug 23, 2004 - 13:32 |
 
|
OK.. I guess I file it as bugreport now, because there is something wrong with the formula, when I play up 40 TS points (=40% - my team had TS 100 vs his 140), manage a draw against the odds, and my coachrating goes down.
But before I start the discussion, I want to set a point.
If you don´t care about coachranking, please save your and my time and do not post. The fact, that a coachranking exists indicates that there are people who care about ranking. Also statements like "if you don´t like the coachrating-system, don´t play ranked" are not constructive. It´s like (WARNING: exaggregation) I would advise to stop living, if you have ANY problem.
The problem with the actual formula forces you only to play game you will win for sure, because in even games, your coachrating will statistically drop.
A major part of the problem lies within the fact, that the coachrating of the playing coaches seems to play a major part. While this is OK, the influence of the coachrating-difference is WAY to much.
Equal, avarage luck assumed, there are in fact matchups, you can hardly win.
Lets take a look at the actual formula:
A player, having 10 points more coachrating is expect to play a matchup, where his opponent has an 17,5%-advantage in terms of teamstrenght.
That´s the difference of a rookie team and a team with 3-4 games, as for every handicap the higher coach gets, he has to play an additional 6 TS-points up. Just browse the the gamefinder and compare rookie-teams with teams at a TS of 120. You will see, that without being exceptionally lucky, you will have a very hard time to win. The same goes for teams at TS 150, who have to challenge teams around TS 175.
As it is VERY unlikely to win those games even on a semi-regular basis, your coachrating will drop, if you continue to play even games. The only choice is to play down, and get small gains from constant winning.
Some people might say, that the more experienced coach should play with an inferior team in order to have an even game. That´s right, but ithe relation between coachrating-dif and TS-dif is wrong in my opinion.
First of all, there is a huge difference between a 140-coach vs a 150-coach and a 160-coach vs a 170-coach. While coaches below 145 make mistakes quite frequently, on which the opponent can capitalize on, a 160-coach plays a solid game. Maybe he is not playing so much games then the 170coach, but he won´t make huge mistakes. At this level, the 170-coach may excell with some clever plays, but he won´t a a similar advantage then the 150-coach over the 140-coach, especially as clever skillchoices are partly already calculated into the teamstrenght.
Thus the difference between a 140-coach vs a 150-coach should weight heavier than in a 160 coach vs 170-coach matchup.
In addition, the Coachrating-dif should NOT weight more than the TS-dif.
Proposals:
Make the coachrating-dif count based on the percentual difference.
Weight TS-dif and CR-dif equal.
Any other suggestions? |
|
|
Tommi
Joined: May 17, 2004
|
| Posted:
Aug 23, 2004 - 14:19 |
|
I think the current rating system works quite well, and TS should be only very small part of the formula.
Why:
Poor players play badly and part of playing badly is making poor teams. Ive played several times teams up to 20 ST stronger, and Ive known as the match starts that I have a superior team.
Of course you can play a match against a team whose ST is up to 40 better and you know it will be near impossible to win, but how the formula can know it?
Only good evaluation of the match difficulty can be made by players themselfs, and team ST should affect rating income of the match only when both players are as good evaluating the match outcome = they have about same coachrating. |
|
|
Mezir
Joined: Aug 02, 2003
|
| Posted:
Aug 23, 2004 - 15:32 |
|
The point of the rating of the coach having such a large influence is that people will now no longer be able to reach, and maintain, very high coach ranking, and it encourages people to play against coaches who have a ranking as high as theirs - not picking on low-ranked coaches as was previously the case. |
_________________
Build a man a fire and he's warm for a day; set a man on fire and he's warm for the rest of his life. |
|
Mully
Joined: Aug 02, 2003
|
| Posted:
Aug 23, 2004 - 15:46 |
|
Circlularlogic - I argued the same exact points a few months ago when the rankings were reset and Christer adjusted the rankings formula. DO a thread search for some detailed analysis I threw in.
The simple truth is the rankings care about what COACH you play and not what TEAM you play. If you want your ranking to go up then you have to find coaches at or above your ranking. Team TR is rather irrelevant in theis process. So go ahead and cherry pick weaker teams if you can.
The catch is the higher ranked coaches will also be wanting to cherry pick easier games so it's harder to "trap" them into a contest where you have the edge. I don't like the current concept but I understand the thought behind it. |
_________________
Owner of the REAL Larson
Come join the CCC League |
|
Britnoth
Joined: Aug 02, 2003
|
| Posted:
Aug 23, 2004 - 15:49 |
|
Mezir that is totally wrong. All that was changes is the relative values to what is considered a 'high' coach rating. What was 220 under the old value is now 190... we had people reaching 220 before the ranking reset, and now we are reaching 190 under the new value. All that has changed is people get upto the high group of players through cherry picking much much faster.
PS. Tried pointing this out to Christer when he made the change, he ignored it. |
|
|
PurpleChest
Joined: Oct 25, 2003
|
yeah, as i understand it didn't the change just squash us all together a bit more? |
|
|
Glomp
Joined: Jan 04, 2004
|
| Posted:
Aug 23, 2004 - 16:19 |
|
Possibly the fairest way to do things is to have a yearly reset. When I joined I hadn't read the rules of Blood Bowl for a long time and wasnt used to playing with the lrb or the inability to take silly mistakes back.
Consequently my rating slipped into oblivion at 114. I gave up on ranked and ignored the figure. After the ranking reset it gave me the chance to compete again and far from losing i've actually done quite well considering my previous standard. At the time of writing im at 161.71 (within the top 100 coaches - previously I'm pretty sure I was at the bottom) and I have been within the top fifty coaches at some point before my experiment with darkelves lost me alot of games. I really suck with them 
Essentially I suggest re-setting the rankings on a given date each year to level the playing field and give the people who previously had a losing record to get back into ranked in a big way.
(note- PurpleChest and Circularlogic arent allowed to respond to this - your ratings are just too high  ) |
|
|
Mully
Joined: Aug 02, 2003
|
| Posted:
Aug 23, 2004 - 17:18 |
|
I dont like a rankings reset but some type of ageing would be nice. Simple too. Maybe the rankings roll and only count your last 300 games played or maybe only games played within the last 500 days or some combination of the 2.
I think that would be much fairer to NOOB coaches who need 20+ games to catch on and be competitive. Why penalize them forever for a 2-20 start? Also, some type of time bleed would force coaches who achieve the 200 ranking to keep playing games. |
_________________
Owner of the REAL Larson
Come join the CCC League |
|
Glomp
Joined: Jan 04, 2004
|
| Posted:
Aug 23, 2004 - 17:35 |
|
Whats wrong with a simple reset? Have coaches on a level playing field.
Its more exciting that way you have the yearly challenge of not sucking |
_________________
Forum terrorist. |
|
phil78
Joined: Jul 29, 2004
|
| Posted:
Aug 23, 2004 - 17:50 |
|
The race would be on each year .The first coach to reach 170 (or something) after the reset.
Or for us who suck the first to reach 130  |
|
|
PurpleChest
Joined: Oct 25, 2003
|
Resets dont worry me at all personally as the ratings are, at best, a mere rough indicator of ability.
But the causialties are usually the newer and lower ranked players. Under any kind of weighting system it is actually fairly easy to pull out of the nose dive, and just as easy to fall from the top. My own development graph saw me plummet over the first 30 or so games, then swing sharply up. Resets just allow people to hunt lower ability players and newer players without the risk of an acurate rating and hence the small gain/large loss equation.
Degrading points over real time or judging on a fixed sample (last hundred games weighting =1, next hundred back weighting=.75 etc) would be fine.
but what exactly is wrong with the current system?
And isn't it possible there will always be problems with any system? so as long as we have one that is stable, usuable and not screamingly unfair isn't that good enough? |
|
|
Mully
Joined: Aug 02, 2003
|
| Posted:
Aug 23, 2004 - 18:28 |
|
| Quote: |
Resets just allow people to hunt lower ability players and newer players without the risk of an acurate rating and hence the small gain/large loss equation
|
Well put - that is the major problem with resets (like the one we had). For the rankings hunters, its a race to beat up on all the sucky coaches while their rankings are still aorund 150 ish.
And like PC said, under the current system it's easy for scuky coaches to get their ratings back up. Just TIE vs average (150 ish) coaches and you will rise in rankings. |
_________________
Owner of the REAL Larson
Come join the CCC League |
|
CircularLogic
Joined: Aug 22, 2003
|
| Posted:
Aug 23, 2004 - 23:49 |
|
Well.. I have 2 problems with playing higher ranked coaches:
1) There are not many.
2) Most of then are.. hm... careful, when it comes to the selection of their opponents.
It would be great, if I could play EVEN games, against lower ranked coaches, by challenging them with a weaker team. BUT if I do so, I am really punished by the formula. |
|
|
Glomp
Joined: Jan 04, 2004
|
| Posted:
Aug 24, 2004 - 00:02 |
|
How do you hunt out poor players when everyones around 150?
I can go into chat now and see loads of coaches lfg with ratings close to or below the 140 mark.
The prefect hunting ground for the unscrupulous. |
_________________
Forum terrorist. |
|
Renno
Joined: Jun 17, 2004
|
| Posted:
Aug 24, 2004 - 00:03 |
|
Wouldn't it be better if the coach rating was based of the percentage of wins you obtain?
My biggest problem with the coach rating is the number of high ranked coaches I see playing stunty or unranked now (which I assume) is to maintain their high ranking.
My second biggest problem is the lack of play weaker teams recieve because people are afraid to lose their ranking.
I've won several games and my rating goes up minimal compared to the drop after the one game I lost.  How fair is that?
<shrug> |
|
|
|
|
| |
Gnomes are trash
Search
