For Clan League 3, I will only be scoring the 8 teams who qualified for the playoffs. All games will count but I am considering the playoff teams to be the Premier League for that season.
Kind of a funky season as the first part was played against the general competition, thus certain statistics will not be eligible for All-Time Rankings as they are quite inflated. Really good season from the top players either way.
CL 3 Top 25 Adjusted Scores Spoiler
1 Chariot of Fire - TOFU
2 Redfelt - IA
3 cachejob - TOFU
4 Dako - TOFU
5 Forza AZ - IA
6 thebest712 - KORT
7 dwilhelmi - EMP
8 jetsetwilly - AOC
9 Ace Rimmer - TOFU
10 jpeter15 - TOFU
11 Eyestone - AOC
12 Nicky15 - AOC
13 Chuuuuck - EMP
14 ballenus - KORT
15 josko.ri - KORT
16 VampireM - IA
17 FarangDemon - IA
18 TheSpaceCowboy - EMP
19 steve066 - KORT
20 jj3044 - EMP
21 Namor - AOC
22 loes - KORT
23 Qyu - AOC
24 laughingcavalier - TSM
25 jabajabba - IA
I believe that I'm going to call this part of the project complete until Clan League 13.
Clan League 1 was all doubles and is not easy to find via Game Finder.
Clan League 2 had no limit to the amount of games a player could play nor does it seem to have any real set of rules I can find, thus there is no real way to compare or have relevant statistics.
I will cherish that single 22nd place for the rest of my CC life. A little bit more than I will cherish that 148th place in CL6, with a 100% win rate. Did I play one game or something?
Any chance you're going to do something similar for CCup? I guess it's harder to do a properly equal analysis when the game loads are dependent on how far you make it.
I’ve always looked at cup as something that should be weighted by round with more weight going on the deeper rounds. There is no discernible difference between 11-0 in the first round and 9-2, but in the semi finals or finals 9-4 vs 7-6 can be a huge difference.
Another factor I look at is when your clan loses how do you perform? For instance, Caymanmew’s performance against us in the latest round is outstanding even though it will be 8-8. Why? Because the overall score will be 27-22, which means he took on a large number of games and gave his clan the opportunity to win the match. 8-8 would be not nearly as impressive if your clan won by a large margin unless you took on a huge number of away games.
This is just how I see it. Others may see it differently.
niMic wrote:I will cherish that single 22nd place for the rest of my CC life. A little bit more than I will cherish that 148th place in CL6, with a 100% win rate. Did I play one game or something?
Any chance you're going to do something similar for CCup? I guess it's harder to do a properly equal analysis when the game loads are dependent on how far you make it.
You were a solid 3-0 in CL6 ha
The larger the sample size, the less that matters. I might consider looking at teams who have participated in more than 4 Cups maybe, that gives you a decent amount of games to look at.
I have 334 games in Cup with my team recording 219 wins in those games. Just over 57% were away games. Josko may have more wins and maybe more games played (although I doubt that part), but I think that is close to the most if it isn't.
iAmCaffeine wrote:i know you worked hard making the formula rule out people who repeatedly get carried in games rather than actually leading teams to victory, but that top 25 shows it clearly hasn't worked fully. seems good for the most part though.
CL4 shows that to be true if I was No. 2, but I'll take it!
Swimmerdude99 wrote:This is actually kinda neat. Second the interest in actually seeing the formula if you do indeed have one.
Dubs + trips + quads x each respective luck factor (25/16.7/12.5%) + home wins + away wins - home losses x overall win %
Each game is awarded a total of 6 points, each home win a point, each away win 2 points, each home loss -1 point
So since there are only two players in a dubs game, each win is worth three points, negatively impacted by a 25% luck factor. Trips is 2, quads is 1.5. Reasoning there was simply that in a dubs game you in general are going to have a larger impact on the outcome of the game than you would in a quads game. There are obviously going to be flaws because I cannot discern who takes instruction and who truly contributes or leads but in general, if you win your games and do it in a well rounded fashion you're going to rank high. If you only play home quads games because you only play 4 maps, you probably won't rank high. If you play a lot of games, you can end up with a slightly inflated number just because you would in theory pick up wins. Of course opinions will exist but I think the results of the games generally tell a true story.
I figure there is always one player on every team that does less than the others; maybe it is only a little bit less, or maybe it is complete dead weight. Whatever the case, this person is the outlier and should be removed. If we adjust for this one player that would mean removal of 50% for dubs, 33% for trips and 25% for quads. This could be a better representation of a player impact constant for your formula if you ever decide to tinker.
Yeah, I hear you, definitely looked at actually counting number of different teammates in wins and rewarding the players who had a larger variety of teammates but the data collection on that was waaaay too labor intensive. And it becomes difficult to assign a blanket number to that given there is no way to tell who is dead weight, if ever. I would love to be able to award or take away based upon leading vs following but I just dont think there is a fair and even way to do that. Also, if you win your games, you win your games. There will of course be instances where players who simply tag along with a better player get wins and rank high but they still have to be cooperative and communicative teammates.
I think its sometimes difficult to determine if someone is a "tag along" or being carried. They may not contribute a lot on most turns but they may make 1 or 2 very important points in a game, in which case I think they are invaluable to the war cause and not dead weight in any way.
super interesting info to comb through and lots of trips down memory lane
