|
Post by Moo on Aug 6, 2013 13:39:54 GMT
Moo - 90.4 (3...108.4) Bairn - 85.2 (2...97.2) Stu - 85.9 (1...91.9) Hornet - 85.2 (2...97.2) DC - 102.0 (2...114.0) Elth - 103.3 (1...109.3) First figure is total points from last year. In brackets are the number of 6 pointers we all had and how that changed your score. Surprisingly, Dave would have won the league with his early cavalier approach and Stu went from 4th to 6th. Stu - this may change your view of the bonus system.
|
|
|
Post by Mr Bismarck's Electric Donkey on Aug 6, 2013 20:04:44 GMT
What happens if you start off with a five or so points each week and lose one for every game you get wrong? Does that make any difference to how the results look?
You can put up the excel sheet, (I know you have one!) and I'll work this stuff out myself if you want.
|
|
|
Post by Moo on Aug 7, 2013 9:04:47 GMT
THe results will look a lot lower and there may be a situation where you end up in negative numbers. I'll have a look at this now.
|
|
|
Post by Moo on Aug 7, 2013 9:11:40 GMT
To clarify... you would get negative weekly numbers. Interesting.
Moo - 90.4 (38...52.4) Bairn - 85.2 (42... 43.2) Stu - 85.9 (39...46.9) Hornet - 85.2 (42...43.2) DC - 102.0 (42...60.0) Elth - 103.3 (38...65.3)
It's very similar to without any bonus, looking at that. I understand why you would want to look at that as an option though as your low scoring wins are offset by your losses, pretty much.
I think I would prefer to be more positive about it though and reward being good.
|
|
|
Post by DC on Aug 7, 2013 17:20:53 GMT
The bonus should be paid based on the proportion of correct answers.
If I get one match right for 6 pts then I get 1/6th (or 1) bonus point. If I get 2 matches right for 6 points then I then I get 2/6ths. So in any week I you got 6 right out of 6 for say 12 pts you would get 12 bonus pts.
In contrast if you got 2 right out of 6 for 12 pts then you would only get 4pts. Someone getting 5 right for 12pts in the same week would get 10pts etc.
There would still be some benefit of getting the big scoring games (i.e. The big 6pts to start with) but knowing your bonus will only ever likely be 1 or 2pts when others may consistently be scoring 3 or 4.
If that doesn't boost it enough, then you simply make it a first past the post bonus with an F1 scoring scale for runners up. Say 10pts for those guessing the most, then 6pts for those guessing next highest, and 3pts for 3rd and nothing for any lower.
In theory this certainly wont punish anyone in bad weeks, and will help promote positive play.
My preference is the first system because it's more nuanced and gives a progressive bonus based on the value of your picks (so in an easy week, the one guy who gets the outlier will still see a benefit).
|
|
|
Post by Moo on Aug 8, 2013 8:19:41 GMT
Eh? I thought we were on about a bonus for getting all results right, not just one or two six pointers? I'm still favouring a 6 point bonus (representing 6 players) for someone getting all predictions correct in any week, mainly because it's easy to do from my point of view.
|
|
|
Post by DC on Aug 8, 2013 15:04:13 GMT
How does a bonus of 6 for being 100% right really matter when I was just as right just as often as anyone else and therefore just as likely to receive it? Or are you suggesting everyone else was also trying to get the wrong right answers?
What disincentive is there for me to not score 12pts a week by picking outliers?
Trust me, I'm a mentalist.
|
|
|
Post by Moo on Aug 8, 2013 15:27:58 GMT
Your example above says that you should get a bonus for just getting predictions correct. Isn't that what the points are for? THe bonus was supposed to be for people who get all the games correct, not just whoever wins the week.
If you can get 12 pts a week for picking two outliers, then fair enough. I don't understand your first paragraph, you'll have to dumb it down a little.
|
|
|
Post by Mr Bismarck's Electric Donkey on Aug 8, 2013 17:02:39 GMT
It's very similar to without any bonus, looking at that. I understand why you would want to look at that as an option though as your low scoring wins are offset by your losses, pretty much. I think I would prefer to be more positive about it though and reward being good. It was just a thought that came to me while on the beach, (fighting off Greenpeace, etc), because right now five wrong games doesn't impact you at all. You still end up with the zero you started off with. Ultimately it looked similar to vanilla scores though, so I'm going to go back to my gin.
|
|
|
Post by DC on Aug 8, 2013 19:37:24 GMT
Your example above says that you should get a bonus for just getting predictions correct. Isn't that what the points are for? THe bonus was supposed to be for people who get all the games correct, not just whoever wins the week. If you can get 12 pts a week for picking two outliers, then fair enough. I don't understand your first paragraph, you'll have to dumb it down a little. Because the point scoring system is biased by other people also effecting your scores (i.e. a sure fire win for a team will only ever get you 1pt) your odds of getting 6pts or more in any given week are proportionately lower if you guess all the results correct (because it's also more likely that everybody else will also choose the same correct answers). In fact choosing the most obvious results will statistically result in the worst return. It's like betting on the 1/4 favourite vs putting an each way bet on a horse at 13/1. I may have no higher chance of winning, but I do have a better chance of winning enough money to buy Boony's Mac with if just one of them comes in. As it turns out, having ran the numbers through an excel spreadsheet it would have made differences in occasional weeks - but Elth still wins and the margins are proportionately the same. I thought the whole idea was we didn't want me winning anyway?
|
|
|
Post by Moo on Aug 9, 2013 10:53:54 GMT
Because the point scoring system is biased by other people also effecting your scores (i.e. a sure fire win for a team will only ever get you 1pt) your odds of getting 6pts or more in any given week are proportionately lower if you guess all the results correct (because it's also more likely that everybody else will also choose the same correct answers). In fact choosing the most obvious results will statistically result in the worst return. It's like betting on the 1/4 favourite vs putting an each way bet on a horse at 13/1. I may have no higher chance of winning, but I do have a better chance of winning enough money to buy Boony's Mac with if just one of them comes in. As it turns out, having ran the numbers through an excel spreadsheet it would have made differences in occasional weeks - but Elth still wins and the margins are proportionately the same. I thought the whole idea was we didn't want me winning anyway? Read between the lines, DC, it's Elth we don't want to win. He always wins everything and if he doesn't it's because he chose not to. Also, he's the best. And the most handsome. Seriously though, I don't quite know if you're winding me up or not. The point of a bonus is to reward those who predict all games correctly. It's irrelevant if someone manages to get a 6-pointer for one of the games, if that's all they get then nobody will receive a bonus for that week. If someone is daft enough to predict Cleveland beating Green Bay, for instance and luckboxes the 6 points then I'm all for that because the odds are that I'm getting 1.2 points and a better chance of a 6 point bonus for getting all games correct. [aside] I've just spent about 30 seconds thinking about what you said earlier... receiving 1 pt per win PLUS the point share based on correct predictions. SO instead of getting 1.2pts for being one of the 5 of 6 "winners" you get 2.2 pts. If you're one of 2/6, you would get 1+3pts. I'll see if I can work that out today to see how that works. In my head, I don't think that would make any difference to the overall results from last season though, as I think that me and Elth were tied on games won, so if anything, I would be nearer 2nd place, but that's about it. If that's the case, I don't see the point in making things more complicated. [/aside] Your horse analogy doesn't work. There are very good reasons why the 1/4 favourite is that short compared to the 13/1 horse. It's unlikely the 13/1 shot will win, so you saying you have "no higher chance of winning" is not correct as the odds indicate what the chances are. If you're will to put your name aginst the long-odds horse (like Cleveland) then you deserve to return more points, as we have currently.
|
|
|
Post by DC on Aug 9, 2013 17:07:18 GMT
In that case, I shall pick the losingest ftw.
My only thought was to encourage maximising the number of right picks vs trying to chance a 6pt'er by being a contrary cunt.
Unfortunately, in practice, I largely picked the same number of winners as everyone else doing it seriously so the impact on my decisions was minimised.
|
|
|
Post by Moo on Aug 9, 2013 17:43:03 GMT
Welcome to NFL parity, population: 31.
Fuck you, Jacksonville.
|
|
|
Post by elth on Aug 10, 2013 2:41:15 GMT
I'm all for bonuses for being lucky smart enough to pick upsets as well, but I don't think we should get away from a scoring system where the best practice is simply to pick the correct winner.
The point for a win plus the bonus for picking against the crowd (ie. being right when everyone else wasn't) seems like a good compromise if it's actually effective.
|
|
|
Post by Moo on Aug 10, 2013 11:45:30 GMT
That's settled then. Predict all games correctly and get a six point bonus.
|
|