|
How to succeed at FPL: Calculating a Player's Cost Efficiency
If you read this through, I think the FPL understanding you will gain from this makes a lot of sense.
Back when Proleague had 7 weeks in 1 round, a top 100 Team in FPL must gain 3 points for every 1 cost it took. So if  soO cost 7 points, on average, he had to acquire 21 points. This benchmark is critical in creating an effective FPL team.
Now that Proleague has 6 weeks in 1 round, each player must get 2.57 points for every 1 cost it took.+ Show Spoiler + So if SoO costs 7 points, he has to acquire 18 points. But wait, there’s more! Every time a team wins, the player gains 1 point.
Based on approximations, a team will win:
SK Telecom T1: 5 times out of 6
Jin Air: 4 times out of 6
CJ Entus: 3.5 times out of 6
KT Rolster: 3.5 times out of 6
Samsung Galaxy: 2.5 times out of 6
Afreeca Freecs: 2 times out of 6
MVP: 1 time out of 6
And since every time a team wins, the player gains 1 point, then we can subtract that number from a player’s expected performance so the player is purely based on their performance. So if SoO needs 18 points, and SKT1 is expected to win 5 times, then SoO by himself needs 13 points. 13 points is 4.33 wins, so for SoO to be cost effective, he must win 4.33 times.
Now we can do this for every single player in FPL, which is what I did.
Savor this.
![[image loading]](http://i.imgur.com/bpGu3z6.png)
![[image loading]](http://i.imgur.com/pkDZP2n.png)
Trends we can see, etc
- Players that cost too much are not worth it. If you look at players that cost 8 and up, they must win more than 6 times, which is impossible, except for ace matches.
- Teams are very efficient at getting points. I didn't bother putting the math onto here, but teams in general are safe to bet on. It's okay to splurge on SKT1!
- It's important to note how often a player is going to be fielded. If a player is on KT Rolster, they are likely to be fielded almost every game. CJ Entus and Jin Air this time around only have 6 players, so a player on their team is very likely to play as well.
If you want to immerse yourself in more FPL theory or analyze previous years' statistics, here you go: http://www.teamliquid.net/blogs/490521-a-fpl-guide-and-statsistics
I think this formula deserves more time than I'm putting into this, but meh. Go win!
Tl;dr: look at bold numbers in the spreadsheet. don't buy players that cost 8+ (7 is also risky)
|
Yay more numbers!
While I mostly agree with you about the 8+ players being not worthy, I think it is not because of the cost per se but because of the risk of not being fielded as ace. I think herO and Zest are the only ones that have that place locked in their teams, herO usually does a good job but Zest not so. Jin Air and SKT have more than one dude at ace level so that makes them even a worse choice to spend bucks on them, in SKT's case the team is so stacked that you are also in risk of not being fielded at all!
|
|
|
I'd say TY and Losira are pretty strong. If KT does what KT does best, TY and Losira are a steal.
I'd also say that Ragnarok is a pretty good pick. He has to win 2 games and since CJ only has 6 players, he should be getting some decent playing time.
I'd try to get Ryung instead of Choya since it looks like Ryung is getting some playing time, but it seems you only have one protoss. I'm also not 100% sold on Bbyong because:
1. Bbyong costs 6, and needs to win 4 games in order to be cost effective. Can he do that?
Looking at last year's proleague and aligulac score, he got fielded 5 out of 7 times, winning 3 of them. (one of them didn't get to be played) His first round opponent is aLive, who is doing not bad in LotV
If you think that Bbyong can be fielded 5 times this round, I think it'd be worth it. I considered bbyong for a while too. He's an okay pick.
Reality is also an interesting choice. Interesting thing about Reality is that he got fielded a lot of times last year, so he does have a decent chance of winning 2.6 times. If you think he'll be fielded 5+ times, he'd be a good pick.
Your team is not bad. I don't see any significant flaws, but I might try to fix a few things. I really feel like Choya is a waste of a point, so I might try to rearrange my team so I don't choose a Choya and instead choose ryung and get another protoss e.g. hush, creator are not bad picks.
I have life on my anti too. I'd be careful of having Blaze on anti since he might play a decent amount on MVP since MVP doesn't have that many good players.
|
On January 31 2016 14:45 Silvana wrote:
Yay more numbers!
While I mostly agree with you about the 8+ players being not worthy, I think it is not because of the cost per se but because of the risk of not being fielded as ace. I think herO and Zest are the only ones that have that place locked in their teams, herO usually does a good job but Zest not so. Jin Air and SKT have more than one dude at ace level so that makes them even a worse choice to spend bucks on them, in SKT's case the team is so stacked that you are also in risk of not being fielded at all!
Yes I agree that players that cost 8+ are likely to be the ace players. Winning an ace match earns you approximately 3-4 points (memory fails me) which counts as another win. Even so, an 8+ player must still be fielded every round and win almost every game, (and the ace match) which is still very very difficult.
|
Thank you for the wise advice...!
I have the least confidence in MVP winning matches, and little confidence in Blaze as well...
I'm a little wary of putting other players from other teams on the anti team..
Also Blaze plays vs Maru first game lol
|
FPL sucks ! i can't even take a full KT rolster team  i had to pick some random pllayer to complete it
|
On January 31 2016 20:01 FFW_Rude wrote:FPL sucks ! i can't even take a full KT rolster team i had to pick some random pllayer to complete it
Yes you can! Example:
|
|
|
|
|
|
EPT
