General FPL Chat 2018/19

Rikand

I'm exactly where I should be lol

Dearg81

fplreview wrote:

So effectively your left with 5 scores, FPL, xG and Odds pts should be obvious from the above. Fair pts is just the average of xG and Odds, ie. considering it from both perspectives for a more overall picture. Luck is the difference between your fair score and the actual result.

Thanks for the explanations. I'd question the use of the word luck though as your making a big leap saying that anyone that beats the bookies odds is lucky, there is skill there too. I know you can't measure that and you are also using xG but calling it luck is misleading.

Brian201888

Anyone beating the bookies odds on a Premier League game is lucky, they out a phenomenal amount of resources into pricing it accurately. It's one thing to beat them on a less traded market but for a Premier League game there's little to no value there

fplreview

Dearg81 wrote: »

Thanks for the explanations. I'd question the use of the word luck though as your making a big leap saying that anyone that beats the bookies odds is lucky, there is skill there too. I know you can't measure that and you are also using xG but calling it luck is misleading.

It's can labelled underperformance/overperformance rather than luck (though try fitting that in a table column header on a phone!).

Having analysed massive amounts of bookmaker data prior to making this site (this was just a natural follow on), I would heavily question the idea of a significant amount bettors being better than the bookies.

Check out Squares & Sharps, Suckers & Sharks by Joseph Buchdahl who does very nice analysis on bookies data and knows it all inside out. He shows very clearly from thousands upon thousands of bettors data that success in sports betting comes appears like a random distribution rather than one indicating skill.

I can also tell you that according to the data the best indicator of future FPL pts is not past FPL pts but actually how good a managers decisions where based on bookies odds (displayed as Odds pts), with xG data being better than FPL pts but worse than Odds.

This to me just confirms what is known, the markets are incredible forecasters and it really takes inside knowledge or incredible skill way beyond 99% of people to actually better them over a large sample.

All in all I definitely would value bookies data more than xG, especially after doing analysis, but from running twitter polls I know that most managers tend towards short term xG.

fplreview

Deleted User wrote: »

How do use odds to get accurate scoring for assists?

Do you calculate bonus points into your calculations? (as it would be fairer for xg but not bookies odds once you are using xg stats against the bonus system FPL use) ((I've no idea but it might be fairer just to leave them out altogether))

How you calculate the clean sheets for Defenders also interests me (in xg terms it kind of makes sense) but when you have a 4,0,-1,-2 ratio. I cant really get my head around it.

Also fill us in on some interesting trends

Great and interesting site,

From my point of view you need to fix the table to be easier read although I'm not 100% sure how

Nice questions.

As there are no odds I use historical xA data to determine a players assist rate. When you know the bookies think Arsenal will score 2 or 3 goals it is then easy enough to determine a prediction. It also is adjusted to account for FPL assists rather than actual assists.

Bonus pts are included and done in a probabilistic manner using relationships from past season data

xCS is not too difficult - if a team concedes chances with 10%, 20% and 30% chances of going in when he's on the pitch his xCS is (1-0.1)*(1-0.2)*(1-0.3)= 50.4%. The - 1,-2 etc. can be determined from a poisson distribution once you know the teams expected goals conceded.

Some trends: 1. Past Team Odds pts is a better indicator of future FPL Performance than past FPL pts and past xG pts. 2. There are a very small amount of teams in the top 1k who the data suggests are displaying the decisions of a 200k team.

Dearg81

fplreview wrote:

I can also tell you that according to the data the best indicator of future FPL pts is not past FPL pts but actually how good a managers decisions where based on bookies odds (displayed as Odds pts), with xG data being better than FPL pts but worse than Odds.

That's interesting, I would have thought that xG was a better indicator. It's a good tool anyway, keep up the good work.

steve_r

Thanks for taking the time to answer questions - think the tool is very interesting.

How do you factor in the use of BB, TC and FH?

fplreview

steve_r wrote: »

Thanks for taking the time to answer questions - think the tool is very interesting.

How do you factor in the use of BB, TC and FH?

They are all accounted for, I can detect which chips and how many transfer hits a manager has taken

nullObjects

fplreview wrote: »

They are all accounted for, I can detect which chips and how many transfer hits a manager has taken

Good work so far.

How about the opportunity to torture yourself by adding a column showing the pre wildcard team returns compared to post wildcard?

Chesty08

Not sure how to post Screenshots or that from Twitter.....

A team in the Top 10k was deleted. Apparently had 2 teams with same email address was the reason but appears to have been an FPL regular

km79

Chesty08 wrote: »

Not sure how to post Screenshots or that from Twitter.....

A team in the Top 10k was deleted. Apparently had 2 teams with same email address was the reason but appears to have been an FPL regular

Which is against the rules ?

eh i dunno

Chesty08 wrote: »

Not sure how to post Screenshots or that from Twitter.....

A team in the Top 10k was deleted. Apparently had 2 teams with same email address was the reason but appears to have been an FPL regular

Hopefully they do the same with a few hundred thousand more accounts and I might have a decent season

Chesty08

km79 wrote: »

Which is against the rules ?

Yes against the rules.

Scorpion Sting

Norwich and Sheffield United promoted. West Ham vs Sheffield United should be interesting next season!

Mr.Prodston

Scorpion Sting wrote: »

Norwich and Sheffield United promoted. West Ham vs Sheffield United should be interesting next season!

Norwich vs everyone will be interesting

Step forward Teemu Pukki!

Dr. Bre

Enda Stevens on my watchlist..

sheroman01

I was looking at possible last day punts, and noticed Fulham vs Newcastle. Was thinking Babel could be nice addition, and Newcaslte could totally be on the beach by then. Looked up their last match day results over the last few seasons and to my surprise, they have dominated them!! Granted they've been at home for the last few:

2015 - Newcastle United 2–0 West Ham United
2016 - Newcastle United 3–0 Barnsley (Championship)
2017 - Newcastle United 3–0 Chelsea
2018 - Newcastle United 5–1 Tottenham Hotspur

Win % - 100%
GF - 13
GA - 1

I'm not necessarily saying go for Newcastle, more so I might personally stay away from Babel

dsmythy

sheroman01 wrote: »

I was looking at possible last day punts, and noticed Fulham vs Newcastle. Was thinking Babel could be nice addition, and Newcaslte could totally be on the beach by then. Looked up their last match day results over the last few seasons and to my surprise, they have dominated them!! Granted they've been at home for the last few:

2015 - Newcastle United 2–0 West Ham United
2016 - Newcastle United 3–0 Barnsley (Championship)
2017 - Newcastle United 3–0 Chelsea
2018 - Newcastle United 5–1 Tottenham Hotspur

Win % - 100%
GF - 13
GA - 1

I'm not necessarily saying go for Newcastle, more so I might personally stay away from Babel

Rafa already said he's not going to make it easy for anyone regardless of game. I say they play to the end. There's no real evidence of any team beaching right now even though they could.

Green&Red

Dr. Bre wrote: »

Enda Stevens on my watchlist..

Used to play GAA with Enda, nice lad

Bleating Lamb

How can you check again which chips you have used?

F1ngers

On my team page and gameweek history.

AdpRo

https://www.anewpla.net/fpl/report/index.php

Handy website for anyone interested in stats!

Some interesting ones for me:

Total captain pts: 624
I've had 7 captain blanks for the season
4 biggest captain hauls are: 40/42/38/38 - 3 from Kun and 1 from Sterling, maybe I should captain city assets more often!
I've captained Salah 16 times getting an average of 17 pts from him.
Salah has played 34 weeks for me, next best is Jimenez on 23!

I really thing sticking with Salah when loads were offloading him has been the key to my excellent season, instead of swapping him for Mane I got both and dumped Kun to fund it, keeping Sterling.

Special mention to Jonny and PEA, both have played 1 GW for me getting 17 and 12 points in the process, if only every transfer worked out as well!

sheroman01

Play Boruc at home to Spurs or Ryan away to Arsenal?

The real mccoy 91

sheroman01 wrote: »

Play Boruc at home to Spurs or Ryan away to Arsenal?

Boruc. Arsenal are a different animal when at home

Scorpion Sting

AdpRo wrote: »

https://www.anewpla.net/fpl/report/index.php

Handy website for anyone interested in stats!

Some interesting ones for me:

Total captain pts: 624
I've had 7 captain blanks for the season
4 biggest captain hauls are: 40/42/38/38 - 3 from Kun and 1 from Sterling, maybe I should captain city assets more often!
I've captained Salah 16 times getting an average of 17 pts from him.
Salah has played 34 weeks for me, next best is Jimenez on 23!

I really thing sticking with Salah when loads were offloading him has been the key to my excellent season, instead of swapping him for Mane I got both and dumped Kun to fund it, keeping Sterling.

Special mention to Jonny and PEA, both have played 1 GW for me getting 17 and 12 points in the process, if only every transfer worked out as well!

Total Captain Points: 512
4 biggest captain hauls: 40, 32, 30, 30
Total Captain Points as % of Overall Score: 23.14%

But the most frightening thing I found out was I played Schindler 7 times this season and he returned a grand total of 1 point. :eek: Now I know why I've been doing so badly for the last six months.

nullObjects

"Possible Captain Points if you Always Captained Highest Gameweek Scorer : 946"
"Total Captain Points : 358"

Oops!

Thanks for posting the link, some helpful stats in there

normanoffside

I should have just picked one keper from the start:

Total Played Keeper Points : 79
Total Benched Keeper Points : 85
Times Played 'Wrong'(fewer points) Keeper : 15

CSWS101

Interesting

https://twitter.com/lego_fp/status/1123934701943570432

km79

Just after realising the Champions league final is THREE WEEKS after the end of the league????? Europa is only the midweek before that
So no fear of rotation for any of the teams that may be involved
The opposite in fact

dsmythy

km79 wrote: »

Just after realising the Champions league final is THREE WEEKS after the end of the league????? Europa is only the midweek before that
So no fear of rotation for any of the teams that may be involved
The opposite in fact

Way things are going neither Champions League team need worry about finals.