Posted by garyt1 on 2019-10-09 22:04:31
I think when you have such a high rating any loss to someone else without a high rating gives a big hit. The racial factor effect is small right?
Does unfortunately probably contribute to picking.
Posted by MonkeyMan576 on 2019-10-09 22:39:24
I wasn't aware race came into play on CR. I thought it was just CR and won vs. loss game.
Posted by Malmir on 2019-10-09 22:49:22
Maybe it doesn't which explains it. Cheers to the person who turned up to cheer my every move the game after. Must feel like 50 min well spent.
Posted by awambawamb on 2019-10-09 23:25:14
I would love to opt out the CR system when playing outside L. until that, I'll keep playing L.
Posted by SzieberthAdam on 2019-10-09 23:54:05
Posted by MonkeyMan576 on 2019-10-09 23:57:09
I know you have individual race CR rankings, I think what he meant is that you should get more CR credit if you are playing a stunty race vs. a basher race.
Posted by Malmir on 2019-10-10 00:04:09
If you aren't going to enter into it anyway Awambawamb, why do you care about coach rating?
Posted by Wozzaa on 2019-10-10 01:58:56
@awambawamb you could just not look at the top right corner of your page, then you are all good.
Posted by ArrestedDevelopment on 2019-10-10 02:38:44
Yep, been this way for as long as I can remember. Which is why 90% of the top rated coaches had never used ogres, halflings, goblins or even vampires in box until the trophy appeared.
Posted by The_Murker on 2019-10-10 07:57:30
The trophy sure isn't a way to climb the rankings or to stay there, due to the zero and one point teams.
Christer helped me with an 'MR' rating system once, where you only move up or down significantly by playing those who are rated close to you. A Legend winning or losing to a noob was not counted for much, up OR down. Everyone with a high MR rating was a total FUMBBL badass.
I wish I had that link. The system was suppose to reset each year.
Posted by det on 2019-10-10 08:04:53
ah well it happens...but I think CR is not a good predictor of coach ability anyway due to many factors, so in the end who cares.
Posted by Joost on 2019-10-10 09:17:10
CR predicts how well a coach can manage all three of:
The actual game.
Racial CR at least eliminates the first bullet, the most interesting statistic to me would be something that also eliminates TV management from the equation. We probably have the data to make steps in that direction but I can imagine it would be hard.
Posted by Civ on 2019-10-10 10:56:12
I thought getting good at bb would automatically make me stop caring about cr. But if even Malmir gets this not upset about it.. :o
Guess I have another thing to work on. Thanks for the heads-up Malmir!
Posted by MattDakka on 2019-10-10 14:00:44
I agree about the CR loss being too big.
Playing a tier 3 team should have a bigger CR increase when winning than a tier 1 team, and a lesser CR loss when losing (a good way could be comparing the expected win rate of a tier 1 race and the relative CR gain and the expected win rate of a tier 3 race and the relative CR gain, then adjusting the CR gain proportionally, so if you win with a tier 3 you win lot of CR, more than if you had won with a tier 1, and if you lose you lose little CR).
I can accept to lose when playing a tier 3, but not if on top of that I suffer a massive CR drop just because I wanted to play a tier 3 and bring variety to the Box.
And that leads to another issue, the lack of balance in BB amongst the different races, which narrows down the variety, unless there is some incentive to play weak teams (like the BB trophy).
Another CR issue is the lack of decay in the competitive division, which allows a coach to be rank 1 forever even if he stops playing in that division.
He has just to be active on FUMBBL, even in a non-competitive division such as League, and his competitive division's ranking will be frozen.
Posted by Malmir on 2019-10-10 18:09:20
I (genuinely) wasn't upset about it Civ. I just found it interesting and thought it might start an interesting chat (which it did).
Posted by tussock on 2019-10-11 06:23:27
There used to be a roster vs roster adjustment, but it had TV-steps, because obviously some teams are much better at High TV and others thus relatively worse, and of course there were never enough games to estimate the correct adjustment for Goblins vs Orcs at 2000 TV or whatever, despite my best attempts.
I suspect it was salvageable by estimating trend lines, but really, if you wanted high CR, it would still be that you should play using better teams. Roll more +ST, eh, makes you a much better coach.
Posted by Lorebass on 2019-10-12 01:28:52
Why do we care about CR?
Posted by Garion on 2019-10-12 10:49:40
I agree. I have been playing goblin lately and a couple of losses on the bounce, and you drop from 10th on the site to 200th. I don't give a monkeys, but it shows CR is pointless nonsense.
I would personally rather it wasn't there at all really. I think it is a heavy contributor to the power gaming types.
But equally I don't care enough about getting rid of it. If people really like it then fair enough. To each their own. there are more important things to take care of, like making the client screen bigger so we can see the icons.
Posted by MattDakka on 2019-10-12 14:10:41
I don't think CR is pointless nonsense, and Christer invested time on the CR formula, so getting rid of it would be wasting his time and effort.
CR provides a sort of abstract "goal" when playing in a perpetual league without any prize, I find rewarding and useful to have an objective number showing if I'm improving or getting worse and it helps as well showing quickly how good my opponent is.
I think that dropping a lot if you play Goblins just shows that the CR calculation could be improved further, but no need to remove CR altogether.
As an aside, making the client screen bigger would reduce misclicks too.
The mouse pointer is roughly as big as a pitch square in the current client, and that increases the chance of misclicking. I don't mind too much about the client graphics, but on the practical side, having a bigger pitch would be a great improvement in terms of playability.