JellyBelly
Joined: Jul 08, 2009
|
  Posted:
Nov 06, 2018 - 20:14 |
|
Hey, something I've been thinking about is that the Black Box Trophy scoring page is obviously skewed towards those that have already played a lot of games, so it is quite difficult for someone like me (who has played 21 games) to see how I compare, in terms of win rate (which is what really matters), with those that have played ahead.
So, my suggestion: how about having a button on that page that toggles the scores (and rankings) to be based on win rate, rather than total score? One way this might be done is to rank by:
([win rate] x 150) + [bonus points from unused team budget]
I would suggest that unused budget points be included, as those are fixed at the start, whereas the extra point per 10 games will cancel out if everyone plays them all and the racial firsts will likely be very fluid (and again lean towards those that have played more games at a given point in time).
If this was done, perhaps only people that have played at least 10 games should be included, to avoid it being swamped by people that have played 2-3 games, with a 100% record.
Just a suggestion for an alternative way of visualizing the scores that would 'normalize' out the '# of games' effect. |
_________________ "Opinions are like arseholes, everybody's got them and they all stink." - The protagonist, Fallout 2
"Go for the eyes, Boo! Go for the eyes!!" |
|
awambawamb
Joined: Feb 17, 2008
|
  Posted:
Nov 06, 2018 - 21:25 |
|
however, you could make that game weighing towards the point rate 1/n, where n is the number of games played by the team that played most games.
so it would be [WR]*[1/n]*150 + [bonus points] |
_________________ "la virtù sta nel cielo e nella terra, ma anche nelle nuvole e nelle stelle"
|
|
JellyBelly
Joined: Jul 08, 2009
|
  Posted:
Nov 07, 2018 - 00:26 |
|
awambawamb wrote: | however, you could make that game weighing towards the point rate 1/n, where n is the number of games played by the team that played most games.
so it would be [WR]*[1/n]*150 + [bonus points] |
Well, n would be 150, because Chivite has already played 150 games, so the 1/n and 150 would cancel in that formula. In which case, the effect of the bonus points would be disproportionately large (because WR would be between 0 and 1). You could divide the bonus points by 150, but that would then just be the same as my formula divided by 150. |
_________________ "Opinions are like arseholes, everybody's got them and they all stink." - The protagonist, Fallout 2
"Go for the eyes, Boo! Go for the eyes!!" |
|
Chivite
Joined: Sep 04, 2017
|
  Posted:
Nov 07, 2018 - 08:19 |
|
How about this?
G= total games
E= extra points( E = race leadsx1.5 + unused points x2.5 + G/10
Total = (Score-E)/G |
|
|
Rawlf
Joined: Jul 15, 2007
|
  Posted:
Nov 07, 2018 - 08:40 |
|
To display the raw points per game was one of the first things suggested by several people when Christer created the trophy page and it was denied.
This does follow a logic, as the challenge is to play a high number of games. End result, not relative succesfulness, is in the focus here.
Also Christer wanted to keep it as simple as possible. |
|
|
Strider84
Joined: Jun 03, 2009
|
  Posted:
Nov 07, 2018 - 09:07 |
|
Well it doesn't matter too much anyway, does it? In order to be in the top 10 you probably will need a win rate around 75%. So if you're around there you're competing, if not you better get your game on
Whether 75 or 72 or 80 is enough to win you only know at the end anyway. |
|
|
JellyBelly
Joined: Jul 08, 2009
|
  Posted:
Nov 07, 2018 - 09:14 |
|
Rawlf wrote: | To display the raw points per game was one of the first things suggested by several people when Christer created the trophy page and it was denied.
This does follow a logic, as the challenge is to play a high number of games. End result, not relative succesfulness, is in the focus here.
Also Christer wanted to keep it as simple as possible. |
Ok, I wasn't aware of that. Although, I would say the challenge isn't just to play a high number of games, it is to do that whilst achieving the best win rate you can. So, a way to visualize how your win rate compares would be useful.
Strider84 wrote: | Whether 75 or 72 or 80 is enough to win you only know at the end anyway. |
Well, I think it would be helpful to have some idea of that before 'the end', hence my suggestion |
_________________ "Opinions are like arseholes, everybody's got them and they all stink." - The protagonist, Fallout 2
"Go for the eyes, Boo! Go for the eyes!!" |
|
JellyBelly
Joined: Jul 08, 2009
|
  Posted:
Nov 07, 2018 - 09:16 |
|
Chivite wrote: | How about this?
G= total games
E= extra points( E = race leadsx1.5 + unused points x2.5 + G/10
Total = (Score-E)/G |
This is basically the same as my suggestion, minus the unused points bonus. |
_________________ "Opinions are like arseholes, everybody's got them and they all stink." - The protagonist, Fallout 2
"Go for the eyes, Boo! Go for the eyes!!" |
|
bghandras
Joined: Feb 06, 2011
|
  Posted:
Nov 07, 2018 - 09:19 |
|
I think a win% without any bonus points would do the trick. |
_________________
|
|
JellyBelly
Joined: Jul 08, 2009
|
  Posted:
Nov 07, 2018 - 09:27 |
|
bghandras wrote: | I think a win% without any bonus points would do the trick. |
Yes, if the toggle button is too complicated then just another column showing straight up win % would be nice. |
_________________ "Opinions are like arseholes, everybody's got them and they all stink." - The protagonist, Fallout 2
"Go for the eyes, Boo! Go for the eyes!!" |
|
bghandras
Joined: Feb 06, 2011
|
  Posted:
Nov 07, 2018 - 09:30 |
|
Another option would be a prediction about the full score. That would need:
1. Taking win % with each race, and see how that projects. How many points...
2. Adding the extra points on top of it. Also the racial lead based on the prediction.
For solving small sample problems i would then suggest to exclude races with less than 5 games played.
As you can see, this is also not perfect. Could not include all races with 1 games played. Not reasonable to assume someone having 100% or 0% win rate after 1 game to do equally. And once those are excluded this might be no good to predict the winner. |
_________________
|
|
|
| |