| Player | Club | Factors | |
|---|---|---|---|
| No Change |  Raya Martin | ARS | C, Vs |
| No Change |  Alisson | LIV | C, Vs |
| No Change |  Donnarumma | MCI | C, Vs, Sus |
| Up |  Ramsdale | NEW | C |
| No Change |  Sels | NFO | C, Vs |
| No Change |  Dubravka | BUR | C, £ |
| Player | Club | Factors | |
|---|---|---|---|
| No Change |  van Dijk | LIV | B, Vs |
| No Change |  Muñoz | CRY | A, Sus |
| No Change |  Jurriën Timber | ARS | C, Vs |
| No Change |  O’Reilly | MCI | A, Vs |
| No Change |  de Ligt | MUN | C, Vs |
| Up |  Thiaw | NEW | A |
| No Change |  Andersen | FUL | A |
| No Change |  Milenkovic | NFO | C, Vs |
| No Change |  van Hecke | BHA | B, £ |
| No Change |  Mukiele | SUN | C, £ |
| Player | Club | Factors | |
|---|---|---|---|
| No Change |  Saka | ARS | A, Vs |
| No Change |  Bruno Fernandes | MUN | A, Vs |
| No Change |  Foden | MCI | A, Vs |
| No Change |  Semenyo | BOU | C |
| Up |  Gakpo | LIV | A, Vs |
| No Change |  Minteh | BHA | A |
| Up |  Anderson | NFO | A, Vs, £ |
| No Change |  Fernández | CHE | B |
| Up |  Bruno Guimarães | NEW | B |
| Player | Club | Factors | |
|---|---|---|---|
| No Change |  Haaland | MCI | B, Vs |
| No Change |  Mateta | CRY | B |
| No Change |  Thiago | BRE | A |
| No Change |  Welbeck | BHA | B |
| No Change |  Woltemade | NEW | B |
Paul Our usual post-match delve into the Gameweek data arrives this evening as we check in with some of the standout players from the previous round of fixtures.
Gameweek 31 remains a factor in our selection, with Everton and Bournemouth midfielders catching the eye.
Watford’s defence under Javi Gracia is also analysed, while Brighton’s playmaker and Leicester’s attack-minded centre-half more than merit their place under the spotlight.
Kabasele to Schidler gives me this:
-Foster
-Alonso(c)-VVD-Davies-Schidler
-Son-Salah-Mahrez-Walcott
-Kane-Firm
---Elliot--Ogbonna--Wilson--CM
Son to D.Silva for a hit if plays 90min vs Juve.Decent?
Jonty's point on the Scoutcast about double digit hauls was particularly bad and really used statistics in the wrong way:
- It would be better to look at variance and expected value rather than double digit hauls since getting 20 is very different to getting 10 and getting 9 is very similar to getting 10.
- As Mark pointed out he looked at absolute number of double digit hauls and not as a percentage. Also the sample sizes were too small, he needed to look back over multiple seasons to make a conclusion from this.
- It's still advantageous to take a hit for a player if they get 5 points in a DGW, you don't need them to get a double digit haul.
- By taking a hit, the max you can lose is 2 points, but the max you can gain is 10+ points.
- As he acknowledged at the end, he seemingly ignored the fixture which affects the % chance of getting a double digit haul.
Why are we talking about DGW? This is a BGW. Yes, you gain a point if you take a hit and they score 5 points, but look at the risk you're taking for just a 1 point profit (which won't really help you). You're relying on a player from one of these weak sides to get 5 points in a single match.
By taking a hit, the max you can lose is NOT 2 points. Do yellow cards not exist in the game anymore? Do players not get injured/subbed off?
Obviously that's a typo and it should read BGW not DGW.
Depending on the player, the probability of a YC is very low (maybe 10%) and the probability of getting subbed or injured is also very low (maybe 5%), so it really is negligible.
You have to think in terms of expected values, let's take Walcott for example who was heavily debated on the Scoutcast and let's use data based on his last 6 games for Everton
- He has played >60 minutes every game and averages 0.33 YCs per game so his expected value from playing is 1.67.
- He is expected to score 0.38 goals per game based on his last 6 games, goals are worth 5 points so he is expected to get 1.9 points (5 x 0.38) from goals (https://understat.com/player/503).
- He is expected to get 0.19 assists per games based on his last 6 games, assists are worth 3 points so he is expected to get 0.57 points (0.19 x 3) from assists.
- He has 3 bonus points in 6 games so we can expect he will get 0.5 points from bonus points.
- Everton have a 29% chance of a clean sheet in GW31 according to the betting odds so he is expected to get 0.29 points from keeping a clean sheet.
Therefore his expected value of points in GW31 is: 1.67 + 1.9 + 0.57 + 0.5 + 0.29 = 4.93 meaning that the hit is worth it and you're expected to gain 0.93 points from the hit.
A better prediction would be a quick regression but don't have the time at the moment to fully develop the model (would take a couple hours) and the data is a little too small.
Exactly. I think some people just don't quite get the maths to be honest. They see '2 pointer' as the standard and therefore can't see beyond that.
Okay come on now, I like xG and xA data too, but this calculation is ridiculous. This isn't science. If it was, we'd all be in the top 5k, just by following some numbers. A player will not necessarily score based on these numbers, they are just useful to give you an idea of how they're performing.
Are you really predicting whether or not he'll get a yellow card based on whether or not he got one in the last 6 games?
0.38 xG is low. If a player has 0.38 xG in a game, he is more likely to not score, which is 0 points. You don't just multiply that with 5 and conclude he'll get 1.9 points out of it. 0.19 xA is nothing. If a player has that in a game, he will almost certainly not assist.
It is the complete opposite of a prediction. It is statistics. That's how they work. Go and talk to a professional poker player. Of course he's not predicting yellow cards; he's simply laying out the scenario based on past recent form. These are Walcott's average returns per game, and therefore a reasonable guide to what he would score, on average, if he played this fixture another 10 times. That's all we can do, but the maths is correct. You're blinded by the idea that blanking would be a huge fail, and you're not looking at how big a difference it would make if he scored.
"and you're not looking at how big a difference it would make if he scored."
- If he scored 9 points at least, yes, it's a big enough difference, imo (5 net points at least). And I don't think he will, so I won't take the risk. If you think he will, you should go ahead.
Why 5 net points at least? I'd happily take 1 net point.
Agreed with OP here, the double digits hauls argument is fundamentally flawed and the wrong barometer to use.
Mate it’s just how expected value works! It’s not a prediction but rathervhow many points the player will score on average in the long run, obviously in each individual week it may be different.
It’s how I’ve made all my decisions for the last 4 years where I’ve ranked 548, 788, 1558 and now 5555 this year and rising.
So yes I genuinely believe if everyone used this sort of method they’d be ranked in the top 5k.
If u are going to calculate expected value from only 6 games, surely the variance and the risk also should be presented. I am sure the variance will be very high.
I calculated it and it is 5.32. This just goes to show that he easily can score negative points as well as huge hauls.
Then it eventually comes down to the fact that whether you have the balls to go for it or not. Simple as that.
Yes you'll notice I mentioned variance in my opening post.
For me personally my utility is only a function of expected value so I don't consider variance. You shouldn't actually be considering variance if you all you care about it overall rank.
If you care about winning a mini-league (and even more so if you mini-league is based on win-loss and not overall points), then variance will be hugely important.
Ideally I would've liked to use more than 6 games but Walcott only has 6 games for Everton so nothing I could do.
Yeah it really was one of those rants that didn't actually make any sense, and made less sense the longer it went for.
The key point that constantly gets glossed over is the fact there is essentially a clear limit (or 'floor') on what a hit can cost you in this scenario. You have a -4 replacing a guaranteed 0, so you accept the risk. That turns into a -2 if your midfielder simply turns up and doesn't get a yellow card. Any attacking returns or bonus immediately puts you ahead. And if you take 4 hits and all 4 players blank, you still only really lose 8 points.
This is completely difference to a normal hit, where the player you take out might even outscore your new player. Regular hits come with massive risk. Hits in a BGW come with very little. The potential upside is then anything up to 20 points (if they score a hat-trick).
Nobody would ever argue that Walcott is likely to get a double digit haul. Frankly, it's irrelevant. But even playing 60 minutes and not conceding gets Wally to 3 points (-1 for the hit taker). I'd call that pretty low risk. And if Wally scores a brace and gets max bonus? Only the dullard can look silly in this situation. The hit taker accepts minimal risk for the potential of high reward.
"And if you take 4 hits and all 4 players blank, you still only really lose 8 points"
- What do you mean ONLY 8 points? Lol. 8 points is a lot.
"But even playing 60 minutes and not conceding gets Wally to 3 points (-1 for the hit taker)"
- That's just bending the argument to convince yourself. I think we can all agree that the most probable outcome is a 2 point blank. If not, he'd be in all our teams. So the most probable outcome is you getting a -2 which is worse than getting 0 points. And for you to really make up any sort of ground, he has to give you at least a 5 point profit imo, which means after the hit, he has to score 9 points. There's your risk - because the odds are low for that.
See above. We have to base our decisions on expected value, hopefully I showed above that the probability is that the hit for Walcott will pay off. Each player needs to be evaluated individually to determine their expected value to see whether it is greater than 4.
It makes total sense. Some can't analyse that way though, and refuse to try.
That is based on the unlikely assumption that ALL FOUR attacking players with good fixtures completely blank. It is the absolute worst case scenario.
Have you ever taken a hit in any GW this season? If you have, that hit came with more risk than any hits anyone takes in a BGW. It can't burn you any more than the burn you take (-4).
Let's say I bring in Stanislas, Shaqiri & Mane for a -12 (replacing 3 x 0). You really think I'm going to be devastated if they all score 2 points and I lose 6 points? What if Mane gets a hat-trick, Shaq goes double digit, etc? The risk averse will be kicking themselves.
I've taken only 2 hits this season and they both paid off. This is because a hit is often paid off in several weeks together, not one single gameweek. Also, the hits were for much better players than this bunch.
In this BGW, you're relying on them to pay you back for the hit in a single match, because you'll wildcard and get them off after. And for that single match, Salah, Firmino or even Mane is fine, but the rest are just risky, imo.
"Let's say I bring in Stanislas, Shaqiri & Mane for a -12 (replacing 3 x 0). You really think I'm going to be devastated if they all score 2 points and I lose 6 points? What if Mane gets a hat-trick, Shaq goes double digit, etc? The risk averse will be kicking themselves"
- That's the point. I don't think Mane is going to score a hat-trick. If you really think he will, you should be captaining him. Are you? I don't think Shaqiri getting double digits is likely either. And yes, I care about losing 6 points.
Okay and what if I told you that in the Mane, Shaq, Stanislas example, ONLY ONE of them needs to score ONE GOAL to break even.
2 + 2 + 8 (goal and 1 bonus pt or a CS) = 12
3 x -4 = 12
3 x 0 = 0
So, forget the magical Mane hat-trick. Will Mane OR Shaq OR Stanislas score 1 goal between them?
Will they score more than that? Because every goal, assist, bonus pt after that is making you ground.
It's like your saying none of these players can score goals or assists. Shaq has 3 goals in 4, Mane has 2 in 3 (not including a cup hat-trick).
Very minor attacking returns will pay off the hit. It's not all about double digits and hat-tricks.
The most probable outcome is 2 points? Again, horrible use of stats.
Shaqiri, for example, has a points per match of 4.5, so his 'average' score is more than double your 'most probably outcome'. Shaqiri performing to his average gets your hit points back. If he scores a goal the hit pays off handsomely. If he doesn't, sure, you might lose 2 points.
It comes down to whether you are willing to take even the slightest risks in this game, like paying to replace a ZERO pointer and hope for a goal or assist.
Yes that was the point I made in the 4th bullet point above - thanks for elaborating on it!
Every time any player plays you take the negative risk of yellow cards injuries. Even if the player is a saint, other peoples judgements, refs etc, can have an affect on the outcome. Therefore that risk can be nullified from the theory. So -2 is maximum loss. Even if the player is a saint playing for a locked out defence, there are other judgments that can affect a players points. All of which are possible to any player who plays. So limited risk if you think that player will score. Last season i took a lot of these risks in a double game week and smashed it. Just depends if you need, want and/or desire the possible returns.
Otamendi has scored 1 point or less 8 times this season (rested twice), so a -4 hit for him would have resulted in a loss in the range of -5 to -3 in those 8 weeks. Where he had 2 points the loss is as you say, a -2
I am not sure you can count any FPL chickens before they are hatched 😉
Then don't take a hit for defenders who lose points for conceding goals. Take hits on attacking players with a decent fixture (Stanislas, King, Firmino, Mane, Pritchard, Benteke, Shaqiri etc).
Something that doesn't get brought up a lot either is that you don't need all or even most of them to haul or mini-haul for it to work. You actually spread the risk by bringing in multiple attacking players to replace guaranteed zeroes.
The ones who don't take any hits will be the ones hiding behind the sofa, not the other way around.
Attackers have different issues and do get subbed before 60 minutes, miss penalties etc though
I agree, don’t take a hit for a defender, spread risk etc .... but you still have to pay and start at -4 for it ... it is all about what you get back thereafter 🙂
Let's play a scenario out:
- We play GW31 100 times over
- I take 3 hits every single time for a -12 + 6 (if they play 60 mins) = -6 worst case scenario for all 100 times. The variation will be negligible on my 'worst case scenario'.
Now imagine how often, out of 100 times, a Shaq / Walcott / Stanislas / Mane type gets a haul (1 goal plus 3 bonus is already 10 points).
Now imagine a Josh King style hat-trick you took a hit for.
If you don't take the risks you will never catch the hauls. The game is full of hauls. You could rise up the rankings if you pick wisely and get a bit of luck. Or you could do absolutely nothing and be guaranteed to make absolutely no progress.
It’s a one game shot though - not an average - it is all or nothing
I am using my FH and effectively buying more lottery tickets for free 🙂
Yes I think it is less justified for defenders since they can concede to lose points, have a lower ceiling and most importantly have a lower expected value compared to attackers.
I see your Otamendi and raise you Benteke. He has outscored the hit (4+ points) in 7 of his last 8 games.
And he has done so by while scoring precisely ONE goal.
So all Benteke has to do is continue his form. If he actually knocks in a goal or two against Hudd you're laughing. Unless you have fielded a team of 6 and just accepted a rubbish GW no matter what.
Yip, he is a good gamble and in my FH team 🙂
The correct mathematical way of getting the likely-hood for scoring is simply:
the probability of the them scoring (bookies) multiplied by for example 4 points for a goal.
then taking the 4 points off the "hit" player and then comparing the 2 players. If the total points is close, it is worth taking the hit as the hit is usually for the short to mid term. (and not just the gameweek's points)
With the blank player, unless its a defender, you can pretty much take off 2 points rather than 4
Yes I just used the understat statistics rather than the bookies odds just because I couldn’t find the odds for each individual player.
That is a very valid way of doing it (indeed I used bookie odds to determine the clean sheet expected value). If you can find bookie odds on goal scorers I’d be happy to update the calculations, but you have to also take into accounts assists and bonus points.
Anyone else fancy Alonso (c) this week?
Sure! Between him and Kane for me. I think Chelsea is going to bring a good performance after that Man City game.
Morning all
Just to say new pod is out: http://whogottheassist.com/pod-30-gameweek-29-out-the-window/
We talk CBs v WBs, 3rd striker options and teams to think about targeting for the run in!
Cheers!
What to do with this lot for next GW? 2FTs, 0.7 itb
DDG
VVD Otamendi Mee Vertonghen
Sterling Mahrez Shaqiri Salah
Kane Firmino
subs: Speroni** Ogbonna Wilson RLC**
will use FH GW31
I would Probably go Otamendi to Schindler, keeping in mind GW31. And upgrade RLC
Not this GW though. Ota has Stoke next.
Plan nowp for the doubles and bench boost ...
Wilson to Ayew
RLC to any budget nailed mid with a double game week coming up
Speroni to a double game week keeper
Etc ...
Repost -
pickford
lowton alonso Schindler
Mane Sane Salah Shaqiri Ramsey
Kane(C) Firmino
Bench - JAyew, Pope, Mbemba, Martina
Is this team Good to go ? Or do I bench someone for JAyew?
Gtg to me
u not concerned about Rambo? (im a rambo owner)
i dont agree on bps removed when a player is fouled..its not his fault if he gets tackled..lol..atleast he is trying to do a positive move by dribbling..i agree on dispossessed but not fouls suffered..anyone else?
I didn't know bps will be reduced or removed if a player is fouled. Sounds not right.
Mkhi > 8.0 ?
Already have Sterling, Salah, RLC, Walcott
Siggy
Double Everton, no way!
Lil Shaq
what is a better move. pogba to Mahrez and then bring in 1 more player in gw31. currently have 7 gw31 players (including dcl)
Anyone for a WC now? Nice fixtures this week for players to bring in, a FH next week could pay off
no chance
Ludicrous
Hmm interesting...
Hi lads,
Play one:
A) Schindler (SWA)
B) Chilwell (wba)
C) Stanislas (TOT)
Thanks
A
Anyone else with less than the full compliment of players for gw31 suffering a bit of FOMO?
What is FOMO?
fear of missing out
Nope
Fantasy Football Scout daily hot topic
Which players is it worth using a Free Hit Chip on or taking a hit for in Gameweek 31?
This follows on from Ragabolly's research into GWK 31 player ownership among the top 10K - well worth reading. https://www.fantasyfootballscout.co.uk/2018/03/06/how-the-top-10000-are-geared-up-for-gameweek-31/
Ideally we are looking for high ceiling players that are capable of a double digit haul.
As discussed on last night's Scoutcast it’s not many across the eight GWK 31 teams.
Salah on 13 double digit hauls – essential
Firmino on 7 – near essential
Rest are v unlikely to haul.
Just five players have 3 double digit hauls.
Niasse, Wilson, Mounie, C Moting, Doucoure.
Just 12 have two double digit hauls
Fraser, Stanislas, Zanka, Mooy, Pickford, Rooney, Shaqiri, Butland, Allen, Richarlison, Mane, Can.
Rest have either none or only one.
It looks as though there are only two players worth getting spending points or a free hit chip - Salah (13) and Firmino (7).
The rest are very unlikely to bring in a haul that will make a points hit of the deployment of this chip worthwhile.
If you are willing to take such a risk then the best of the rest are worth considering – These are the players with 3 double digit hauls this season
Wilson
Mounie
Niasse
C Moting
Doucoure
But even they come with a health warning around rotation, lack of form or poor fixture.
There is also an argument that most players people are thinking of spending points to get in for GWK 31 won’t even make back the four point investment.
Ian Wilson , who can be found on Twitter as @FantasianPL has carried out some excellent research into players who average 4pt a match.
https://twitter.com/FantasianPL/status/971121533023264768
It’s grim reading….with barely any reaching this milestone.
Jonty, going to copy and paste my post from above on this page to show why you're reasoning of looking at double digit hauls over expected value is flawed:
- It would be better to look at variance and expected value rather than double digit hauls since getting 20 is very different to getting 10 and getting 9 is very similar to getting 10.
- As Mark pointed out he looked at absolute number of double digit hauls and not as a percentage. Also the sample sizes were too small, he needed to look back over multiple seasons to make a conclusion from this.
- It's still advantageous to take a hit for a player if they get 5 points in a DGW, you don't need them to get a double digit haul.
- By taking a hit, the max you can lose is 2 points, but the max you can gain is 10+ points.
- As he acknowledged at the end, he seemingly ignored the fixture which affects the % chance of getting a double digit haul.
You have to think in terms of expected values, let's take Walcott for example who was heavily debated on the Scoutcast and let's use data based on his last 6 games for Everton:
- He has played >60 minutes every game and averages 0.33 YCs per game so his expected value from playing is 1.67.
- He is expected to score 0.38 goals per game based on his last 6 games, goals are worth 5 points so he is expected to get 1.9 points (5 x 0.38) from goals (https://understat.com/player/503).
- He is expected to get 0.19 assists per games based on his last 6 games, assists are worth 3 points so he is expected to get 0.57 points (0.19 x 3) from assists (https://understat.com/player/503).
- He has 3 bonus points in 6 games so we can expect he will get 0.5 points from bonus points.
- Everton have a 29% chance of a clean sheet in GW31 according to the betting odds so he is expected to get 0.29 points from keeping a clean sheet.
Therefore his expected value of points in GW31 is: 1.67 + 1.9 + 0.57 + 0.5 + 0.29 = 4.93 meaning that the hit is worth it and you're expected to gain 0.93 points from the hit.
Each player needs to be evaluated individually to determine their expected value to see whether it is greater than 4.
Also sorry for the wall of text but was collating a couple posts together.
Cheers.
So at best Walcott in for a hit could bring in 1 pt if you believe that Everton at home is the same as Everton away. Worth maybe redoing the figures for last four away for example... as the GWK 31 game is away. On current form Everton are far stronger at Goodison than they are on the road.
Cheers for the response!
It's not "at best 1 point", it is an expected value of 1 point, meaning you will gain 1 point on average in the long-run. At best it could be 10+ points.
Changing it to the last 4 fixtures is an entirely valid point and may mean that we shouldn't be getting Walcott. The only point I'm making is that your methodology of just looking at double digit returns is flawed and we should focus on expected value instead.
Whether it is an expected value over the last 4 gameweeks, 6 gameweeks or the whole season is where the subjectivity comes in.
Good point about home/away, that may tip it in favour of not taking the hit.
Yep - that's what the discussion was about - is he worth taking a hit for and looking at away evidence he isn't.
Also as said the double digit haul aspect is just one to look at and shouldn't be the only one - lots like Hall of Fame number one Jay have come to the same conclusion looking at ppm.
If I recall correctly in the same video he did also say that every point will count in a gameweek like that. Would have to re-watch to see if he said he would take a hit or 2 though.
its his method and therefore his opinion don't try and prove jonty wrong when hes just voicing his opinion
Also looking at double digit hauls is just one way of looking at it...there's lots of other ways.
Jay Egersdorff and Ian Wilson (see link below) come to the same conclusion as I did - but using average points per match. In fact their research is even more damning as there are few players that will even earn that 4 pt hit back.
pritchard only player.. is he an option jonty?
He has a slightly higher chance of rewarding that hit.
Check on his fitness though ahead of GWK 31.
Yep just think you’ve got the potentially the right conclusion (it’s borderline) using the wrong method.
Thanks for the discussion!
sorry Jonty where's the link below you mention?
Can't see it (maybe dirty specs)
Haha and I'm just voicing mine.
There are different ways of predicting points and I just thought his was a little skewed - nothing wrong with challenging other's opinions! Surely that is the point of this site...
Indeed. 🙂
What a ridiculous reply to a hot topic trying to promote discussion
I think that's a little unfair. My post was designed to contribute to the discussion in a meaningful, unique way.
If you look above where it's repeated it has promoted some interesting discussion itself.
Still, if you want to delete it because it's too overbearing I would not be offended, especially since I've made the same comments above (just wanted to make sure Jonty saw it).
I think Andy was referring to Adam28193's comment.
Haha I think we all got confused 🙂
Oh so he was - my mistake!
Andy, the reply does actually promote discussion and raises an important point. Jonty is absolutely right in using research to identify the best GW31. He is however not right when focusing on double digit hauls only (or mainly). The argument is rather straightforward and backed by statistics: >4 points expected do justify a hit. In fact, Andy, it justifies your whole strategy of taking hits before GW31 (if of course they are rewarding in expectation)
Expected values are useful over the long term but not for a once off as values like 0.93 can't be achieved. Probable values are morr useful and for most gw31ers it's 2.
I agree with you on the second point, but this is compatible with the use of expected value. In simple words, if you believe a player has expected value of 2 then you definitely should not bring them in for a hit (unless you have a longer term strategy)
This is what i called half cooked statistics to make your argument look good. With his calculation and sample size, the deviation from his expected value is around 5.32 which means he can easily haul a negative points as well.
Also he has not taken into consideration many important points, like what is the correlation between assist and goal. How likely will he assist if he scores?
Also, he need to take into consideration the spread of point depending on the opposition and venue of the match. His expected value is far fetched.
Last point to add, his entire stats of last six match is skewed because of the 15 point haul which in the long run happens occasionally leave alone over 6 match. It is by luck that the 6 match he considered happens to have a 15 point haul.
Argue is invalid.
Half-cooked stats can be a better fit for a purpose than other methods. This is one of those situations 😆
My post wasn't so much about Walcott but it was more just suggesting that we should be aiming to use an expected value rather than simply total number of double digit hauls as our barometer. I could've made it more accurate for Walcott by taking more games (but he's only played 6 for Everton) or by using bookie odds for him scoring in GW31 rather than just using xG and xA (since then it would take into account the exact fixture), but I could not locate those. If they are anywhere online, I'd be happy to update the calculations. I suspect it would still be > 4 points.
You can't just discount the 15 points, he did score that so it must be included. He is also capable of those big hauls as we've seen in the past few seasons, so it's not really an outlier.
Also a couple points you raise are irrelevant: (1) the correlation between goals and assists doesn't matter because I'm taking an expected value, not running a regression; (2) the spread of points is irrelevant since we should only care about expected values if all we care about is overall ranking (variance is only important if we also care about mini-leagues).
The correlation does matter because he cannot be scoring and assisting at the same time. The correlation in this case is negative and increases the possible deviation from expected points scored and thereby again increasing the possibility that he wont even score the expected return.
Actually, the correlation between scoring and assisting would be positive because if it is a goal heavy game he is more likely to both score and assist. Why can't he score and assist in the same game? Obviously he can. Even still, correlation has no relationship with expected value so it is irrelevant. Remember, we should only be caring about expected value and not variance provided that all we care about is overall rank.
Overall, your comment here is nonsensical.
Very good point!
Excellent analysis, cheers.
Haven't watched the scoutcast yet but Jonty keeps bringing up on here how bad Everton are away, and he is correct to do so. But I feel he is overlooking Everton's excellent home form. Based on the last 8 home matches (incidentally since Allardyce has been at goodison), Everton are 5th in the form table, ahead of Manchester United and Arsenal, teams known for their home form.
Therefore the time to bring Walcott in for a punt, (for those who haven't already) is not in GW31, its now, ahead of a fixture against a Brighton team who haven't had a single win in their last 8 away games.
Specificaly on whether Walcott is worth a hit for GWK 31 away to Stoke.
Over his last four away matches Walcott has only had two shots on target, which is pretty poor. Yes he’s had 7 shots in the box over that time – but if they are not on target then that shows he is lacking accuracy. Failing to hit the target from close range so often is a negative against him.
Here’s a FFS Members table I’ve made with more info on Everton.
https://members.fantasyfootballscout.co.uk/my-stats-tables/view/5423/
Of course, anyone can score at any time in theory – but the stats suggest that he may not be the best to bring in for a hit in GWK 31.
Thanks for this, I’m also leaning towards not bringing Walcott in. True, he can return. But we need to give numbers! How much? And, most importantly, how likely?
For me, I’ve approached GW31 with the mindset of playing the Free Hit.
While we can say that one team is more likely to score than another, other than Liverpool, I think the reliability of the remaining seven teams is questionable.
Huddersfield have good form and should come out on top vs Palace.
Bournemouth should beat a poor West Brom.
Stoke should beat and keep out Everton who are desperately poor on the road.
But a Free Hit manager is unlikely to take liberties with this. They are probably likely to hedge their bets and try to cover as many eventualities as possible.
With a FH I’d probably go with 3 Liverpool and then try to cover the other teams most likely to get goals or clean sheets – Huddersfield, Bourenmouth and Stoke.
Having said that, in attack – I’d still limit myself to probably one Stoke (Shaqiri), one Huddersfied (up for grabs) and one Bournemouth (Stanislas/King).
Then I’d probably look at the individuals most likely to produce attacking returns in the other teams.
In that category I’d say Free Hit managers will be looking at Walcott, Benteke and maybe, Deeney – though his task at Anfield is surely the toughest.
Given that, I think that Walcott is likely to be in 80% of Free Hit squads, maybe more.
So, say 20% of the managers above me are Free Hitting and Walcott is likely to be in 80% of those teams, there is extra incentive to consider him if you don’t already have him.
Would I take a hit for him? It's debatable.
But I look at it like this.
The cost of the mistake should he blank is 2 points.
The cost of the mistake should he score could be six points or more, plus the fact that the vast majority of Free Hit teams have him.
The fact that he’s a player that is heavily involved in attack compounds the risk.
Since he made his first start for Everton in GW24, his xGI (Expected Goal Involvement) is 2.53. The only midfielders in action in GW31 who beat that are Mane and Salah.
I’m not sure we can rely on him to keep missing chances. Last season he hit the target with 46.8% of his shots on his way to 10 goals in 23 starts. No midfielder with 10+ goals bettered that accuracy. Okay he played for Arsenal, but he's still getting big chances for Everton and, so far, he's not taking enough of them but by writing him off, we're assuming that this continues. His data last season does not indicate that he will go on missing the target.
Overall, I think you have to be very certain on Stoke vs Everton will go to form.
It probably will, it might not and when you're Free Hitting, you're unlikely to take that risk. I think that means Walcott will be a hugely popular FH player and therefore, that has to be considered when debating the GW31 hit for the rest of us.
Mark, one question: are you considering the fact that 40% of top 10k managers have already used their FH chip? This number has gone unmentioned, but it is truly striking and can have a huge impact in the calculations above.
Maybe you could loan him Diane Abbott to help look at the figures?
Don’t spoil my strategy
I'm surprised by that. But I'm not that concerned about the top 10k - I'm nowhere near it! 😉
That number means one thing: you (as well as I) should not give up hope.
Everton are one of those teams where you have to look at away form carefully. They are far poorer away - which is key for GWK 31 as they are away then.
If you were Free Hitting, would you have Walcott?
As I mention above, it's surely about covering eventualities. The form and data points to a Stoke with with a clean sheet but I don't think Stoke vs Everton is the kind of fixture we can call with any degree of certainty.
Personally, if I was Free Hitting I wouldn't trust Stoke to perform or Everton to fail. I'd cover both with Butland/Bauer and Walcott.
You bought Niasse. Surely you did that with some thought that Everton must upset the form book in GW31?
For example Walcott's xGI last four away matches is 1.53.
Around 20 players have a better xGI.
To be honest I'm already regretting getting Niasse in and wish I'd gone for King 🙂
Yes but how many of the 20 actually play in GW31? I make it, again, that Mane and Salah are the only two midfielders above him for xG based on the last four GWs.
And 1.53 is actually pretty good. Look at the midfielders who haven't managed that.
It's no guarantee of points but it does underline the risk of ignoring him. It shows he gets chances and last seasons data shows he can take them. Everton might not create anything at Stoke. Walcott might keep on missing if they do. But it's a risk.
He - like Shaqiri - is one of the tier two players I'd want going into GW31 just in case Everton vs Stoke doesn't stick to script.
Would I pay a hit? Dunno, it's still marginal.
I don't understand the rationale behind covering players that FH temas will have. You've chosen not to FH and now you're looking to get a FH team with hits. It's an argument from ownership instead of points, which I always struggle to get my head around.
"So, say 20% of the managers above me are Free Hitting and Walcott is likely to be in 80% of those teams, there is extra incentive to consider him if you don’t already have him. "
Why?
The points you make above to justify the hit - do just that. They justify spending 4 points on Walcott. But that has nothing to do with the FH teams owning Walcott, and everything to do with the debate about how many FPL points he might score (the only useful question in my view).
It's like you narrowed down the options by looking at ownership, and then justified it post-hoc by estimating points gained for your team independent of other FPL teams (and their players andd strategies).
Perhaps you could just look at the players you think will score most points in GW31 (and GW30), and pick the best, regardless of ownership. You might end up choosing Walcott still, but at the moment I think you are trying to find reasons to buy him, rather than choosing him as the best option of many.
Good post and interesting - and should be used in conjunction with Ragabolly's excellent article posted yesterday.
One downside of using averages to estimate future returns is that we often get convinced that these averages will be maintained, and of course, that rarely happens over one gameweek. To put into context, the fixtures for Bournemouth, Stoke and Huddersfield in GW31 are among the easiest they will have all season - so would we expect players from these teams to score more than average because of this? I'd say there's a reasonable chance of this - and this should be taken into account.
Actually did this as a follow up to Ragbolly's -I'll put a link in now to the HT. Thanks for the reminder.
This is the way I'm looking at it too, although I do have rose tinted glasses on following last years blank.
Stanislas and King have not hauled for a variety of reasons but I wouldn't bet against them doing well versus the worst team in the league.
There's an element of risk to it but I don't think the risk is that high.
Definitely agree with this.
It also depends on when you're taking the hits, doesn't it? For example if you take a hit this week for a 31 player, they have two game weeks to make up that deficit, therefore aren't relying just on one week.
I've still not got Shaqiri but I think he will be worth a hit to bring in next week regardless of if he has only had 2 double digit hauls. I think he is more than likely to get more than 4 points. What do you think?
On current form I'd say he should reward a hit...maybe not double digit....but certainly would expect at least 4 from him.
My thoughts exactly, thank you!
Think Shaqiri is probably one of the better ones to get as it looks right now. Although part of me is tempted to bet against him and triple on Everton defence 😆
Whilst not worth a hit, a FT for Zaha could be a great shout for GW31 and isn't being mentioned anywhere if no other pressing moves ... Could be back for Chelsea if reports are true!
Yip, and you have auto subs to cover it in case he doesn’t play
I think you need to ignore a double digit haul or average points and simply target players that have rhe higher ceiling because they are ...
- players on set pieces
- players on penalties
- nailed players that are heavily relied upon
- nailed players OOP or attacking
Pick players which have the best chance of scoring the most points - after all it is a free hit 🙂
Lossl
VVD, Bauer Ake/Van Aan)
Salah Mili Shaqiri Siggy/Walcott
Firmino Benteke King
This only really works with a FH because you are buying more lottery tickets to catch the haul 😉
Stanislas for me. Good underlying stats and has showed in the previous two seasons he's capable of big hauls.
I don't think we necessarily need to be worrying about double digit hauls.
5 points minimum is a win in this case, although I'd still say if you aren't wildcarding in GW32 the hits should probably be kept to a minimum unless you're happy to keep that player.
It's whether you're prepared to gamble points to potentially get only a small return
This is the whole point of the FH in 31, you are getting 7 lottery tickets for free (every one will have Shaq) hoping that two or three are winners 🙂
Playing a team of 8 or so means you only have 4 tickets and you are hoping two of are winners 🙁
*(Everyone will have Shak & 3 Pool)
Agree!
Not taking a free hit but currently have Hennessey, VVD, Schindler, Salah, Walcott, Richarlison and Firmino for week 31.
Thinking of doing the following transfers for -4 to give me 9 players. Don't see anyone else I want.
Sterling > Shaqiri
Ogbonna > Daniels
Will wildcard in week 32.
Putting Mane is that group on players is really misleading. Mane has looked just as dangerous the last few. Capable of being explosive unlike the options you’ve baskets him with.
I'm taking BGW31 as a personal challenge. Re my post just below, I plan X1 starters. I'm 100 ahead in my main ML, and I know most will only have 5 or 6 players as they've already burnt most of their chips except for one guy who still has all of his. I have only used TC so far. So if I outscore the rest by circa 20 points I think the league is done given the chip advantage as well.
I'm basically a relentless points accumulator screwing the rest into the ground this season in my ML. My target was top 50k after three poor seasons, maybe a top 10k is still possible and will be helped by getting X1 out for BGW31!
The proper way to approach this is to do some modelling. You should get a player in for a hit if you expect him to score more than 4 points (if you WC in 32). The Rate My Team model, as an example, expects more than 4 points from only five players: Salah (7.92), Firmino (6.48), Mane (6.15), van Dijk (4.31) and Shaqiri (4.17).
All this fuzzy thinking about hauls and ceilings is making me nauseous. 😯
Currently listening on Scoutcast and one fixture related question just hit me.
To be able to go all in with gw31 players and Wildcard in 32, you really need the gw35 extra fixtures to added by then.
Will they?
Sorry, the extra gw34 fixtures I mean.. 😉
We know the doubles, just don’t know which week they will be placed in
But very likely, since several teams will end up with two doubles, the games will be re-scheduled before gw32, just 14 days before gw34?
The issue is that the 31 blanks could have been moved to 34 already, but they have not, which makes me suspicious !
If Chelsea and Spurs crash out of Europe it opens up GW32 & 33, and if Spurs crash out of FA Cup their blank could go to 32, 33, 36, 37 (I think)
I also suspect the broadcasters want the games spread out to have an ‘exclusive’ time slot, which is tricky if all the games are lumped into Tue/Wed in 34?
There may be a twist yet 🙂
But UEFA does not want domestic fixtures to be played during Champ League dates? I'll start a new topic. Thanks for now.. 😉
Not an issue with that this season
I have dcl do in my team still. should I keep him for gw31 even though he isn't starting every game. also stuck in a dilemma for pogba to Mahrez this gw or bring in two for gw31 next week
Surely the RMT is tailor made for situations like this
If a player has a RMT score higher than 5 he is worth a hit if no value is lost in the transfer, if you want him out right after than a RMT of 8 is something you want.
So basically hits won't be worth it bar the LIV assets if u dont have them
Not really, RMT is an average and we are trying to pick the player that will push the points ceiling, very difficult with a single pick and a bit like trying to pick your Captain (except Salah)
IMO, only Salah is worth taking a hit for.
Firmino is fair enough, but even that has 50% chance of failing for me.
Mane a bit more risky than Firmino but again, probably fair enough.
Wouldn't even entertain the idea of taking a hit for anyone else.
One should also take into account differential gains in GW30 and possibly GW32 if wild carding in GW33. E.g. expected returns of Schindler (SWA, CRY, new) vs. Dunk (eve, LEI).
I am getting Walcott for Ibe this weekend for a free hit. He then has 2 fixtures. You mention his lack of apparent goal mouth action, but I watched the game at Burnley. He had 2 huge chances and set up another huge chance. The eye test tells me he is worth getting in at home against Brighton then Stoke away. Both poor teams who he could score against. Plus, historically Walcott has hauled big. I know he is only playing for Everton now, but still. The problem with all these stats are they are open to interpretation. I am likely to take a hit to get 11 players and a keeper in for gw31.
Actually it was more his accuracy. No problem with shooting from close range...just few are on target.
I actually think getting in Walcott this week for a hit could pay off just as BRI is a great home fixture for him. Also as you say he has two fixt to reward you.
A reasonable assumption that many non-free hitting top 10k sides could only go in with max. 10 players anyway.
Current keeper ownership for playing keepers GW31 is very low.(possibly under 10% have a potential starting keeper atm).
Doubtful worth taking a hit for a keeper.
Top 1k. i'm on 6 + Quaner. Going to probably bring in a Huddersfield defender for Ogbonna this week then look at having max. of 3 transfer next week for a -4. Have to bring in Shaqiri just to cover the ownership. Then at least one will Bournemouth mid/forward.
On the whole still think I'd rather a drop 8-10pts to rivals in GW31 in order to save FH for GW35 (or later depending on FA cup).
Mark absolutely schooled you on this one Jonty. You’re dismissing Walcott’s inability to haul based on a small sample size. The stats are clearly there and if we extrapolate the data the likelihood of him performing is much higher than others you have suggested. I hope your dismissive attitude towards Everton comes back to bite you.
Ha, got to admire your club loyalty in the face of adversity 🙂
I actually argued that Niasse looked more worthy of a hit than Walcott - as even though not nailed he has hauled more often. Note he also has not started that often.
Also see above for specific Walcott stats - he's had plenty of time and away from home has been poor.
Tongue in cheek of course. Would love to have a wager with you to see who finishes higher this season between our respective clubs!
It's quite the tussle for mid table obscurity 🙂
Which for Brighton is incredible given I can still remember us needing an equaliser against Hereford to stay in the football league!
Jonty is doing a much better job on Walcott than Diane Abbott. We should all be thankful.
cheers Jezza. 🙂
For those without a WC, the other thing worth considering is whether these players will require a hit to remove in GW32/33 since you will likely want to replace them with someone with a double in GW34.
E.g. Arnie -> Walcott (-4) in GW31 will likely also mean Walcott -> Arnie (-4) for GW34.
Given we're on a maths hype, looking at the inequality of whether Keep Arnie < Get Walcott.
[0 in GW31] + [Arnie in GW34] < [Walcott in GW31] + [Arnie in GW34] - 8
if, and only if
8 < [Walcott in GW31]
I.e. FOR THOSE WITHOUT WC, the Arnie/Walcott swap & reverse swap only pays off if Walcott scores 9+ in GW31.
The double figure theory is one of the worst I’ve heard, a hit is worth 4 pts not 10. A player who scores 8 pts every week would have no double digit hauls yet still be worth spending 4 pts on.
As a few have said any player with an expected points greater than 4 is worth getting in on a one week score basis. Most average attacking midfielders on average teams manage to average over 4 pts a game and have ok enough fixture to be expected to score more than 4 pts in 31.
If you are not WC 32/33 and need to drop the player using FT or another hit before 34 then it’s probably not worth the 4 pts in and potentially another 4 pts out as few players average or can be expected to score more than 8 pts in a single game week.
Completely agreed. I've really struggled to hold back from commenting on this thread 😈
It's the same as captaincy decisions. Many choose players with a 'high ceiling' after 2 hauls in 10 games, whereas the better picks (who may or may not have reached double figures) have outscored them during that period.
Every week you make a decision based on poor logic is an opportunity for probability to stuff you.
This is pretty much spot on.
unfortunately that logic doesn't appear to suit the narrative that is presented. Just by reading responses to any critique you can see that confirmation bias is prominent once again.
No doubt after game week 31 is completed selection bias will be evident to justify original point.
The fault lies not in the preaching but in the preacher
Anyone considering doubling on Huddersfield defence this week? Everyone will probably go for 1 defender and if they haul in either of the weeks could be a nice differential.
Would have been interesting if one of the scoutcast was using and planning the free hit gw 31 too see the difference against the others Andy os usually brave enough too take a punt or a gamble good episode last night lads thanks
To partially answer the question, I'll ask the question I posed to the Scouts yesterday:
"For those planning to take a hit (or hits) next week, is it worth spreading the cost by making the transfer(s) this week instead - or is it too risky due to injury/lack of nailedess?
E.g. I could buy Baines for a hit this week and bench Shaqiri (MCI). I then need Baines to get 2 points more than Shaqiri this week, and then 2 points more than a blank player next week (aka 2pts total next week). However, there's no guarantee that Baines will play in GW31; whereas next week, I would be able to make a more informed decision."
I'm actually weighing up the sale of Zeegelaar, Wilshere and Vardy to buy Baines, Theo and Mounie for -4 this week. You could choose to put the hit on one of the upgrades, Baines or Walcott, share it between them or across all three. But the reason I'm doing it this week is to pay for the -4 with the points gained from playing the new players this week instead of Shaqiri and Wilshere. Hopefully that shall leave me with less than 4 points to pay back in GW31 with the "expected value" that Jasong123 talks about, plus extras.
With any luck, it will pay iteslef back this week, but it's always possible that Vardy wil be a hero and make me look stupid.
Interesting points, but am I the only one looking past GW31? Taking a hit for Walcott, which lets say returns 10 points (6 net). He then has Man city, Liverpool and swansea away (who have conceded 2 goals at home in their last 4). Do you want Walcott for these games? Or will people be taking another hit to bring him out.
Looking past GW31 - Liverpool, some obvious choices that I want go into.
- Huds - Mooy if fit
- Watford - Mariappa at 4.2 is easily fitted in most teams, Doucoure is now outscoring Richarlison
- Bourn - decent fixtures up till GW34.
I think making 2ft or even taking a hit for players just because they don't have a blank in GW31, needs to be done for longer term benefit than just 1 week.
very few would still have him for those games, it would either be immediate transfer or WC
Is it worth bringing someone in for 1 GW? Very rarely IMO.
Well that’s a different point
You need to look at those teama who playing really well and fighting for survival. CRystal Palace have been brilliant and if it was not tough sides, benteke and pva would have hauled. One of cpl looks good differential for gw31
Agreed and their post GW31 run isn't too bad.
Of course, if you take a hit this week to play them in 30 and 31 it means the hit is lessened even more.
Sadly for me, I’d want to play my likely taking-out-for-a-hit candidates in GW30!
this hot topic is sucking the fun out of fpl. All that expected goal stuff may be instructive in the long run but in a single game week it realy means close to f all. I'd be more than happy to get Theo in for a hit, fail and you lose 2-3 points whoop de do. Buy yourself a ticket for the raffle and ok you might end up with a key ring and a cd of Susan Boyle's greatest hits but there's that all inclusive holiday in Barbados up for grabs, good luck fellow punters, let the stat gurus freeze on the sideline as we make our way to the departure gate
How can it suck the fun out of it if you don't bother with it yourself?
I guess you like doing it your way but you're worried that this HT is going to help people actually get more points than you?
you do like to take things literally don't you. And no I'm not worried just voicing my opinion that hits are worth it
You weren't being literal? What did you actually mean by, "this hot topic is sucking the fun out of fpl"?
I'm looking at bringing 3 31's in next week with 2 frees and a hit. I don't know which player is really the hit so can blend it into all 3.
Kenny out >
A) Francis
B) Tomkins
C) someone else
4.4 to spend, need the player coming in to not have a blank next GW
Alexander-Arnold?
Got 3 Pool, good shout otherwise cheers.
Anyone getting him Huddersfield defenders this week?
Might get Lössl.
I brought in Kongolo, couldn't afford some of the more expensive defenders!
Schindler is on my list.
Hehe
Maybe Lossl. No one else.
Mee, Mkhykitarian to Zanka, Sigurdsson for -4 this week?
I wouldn't.
Let’s agree on a few things. There are different ways of playing the game.
Primarily we are all trying to second guess what will happen. And on every weekend there are players who surprise us. It is by no means certain that the obvious players will be the ones to haul in the points.
The luck element is high and unquantifiable. The skill element is also quite high. The good strategic planners will usually do better. Playing cards at what is perceived to be the right time doesn’t always work.
I see many players use their wildcards and make things worse!
So shouting at each other is pretty daft! Each to their own. And are those at the top lucky? You bet they are.
Yeah sounds about right
If you have all your chips left and you want you use BB and TC in both DGW 34 & 37..when would you use your Free hit chip other than GW31?Will the other blank GW involve a few more decent teams and therefore better saving until then?
Yep, 35.
Going to be a blank in GW35 - worth using it then, better teams available to pick from.
Cheers both
Best captain option?
A) Aguero
B) Salah
C) Alonso
Obviously I don't have Kane... Currently sitting on C.
Kün.
B
B or C, Kun been non-existent away from home, made that mistake with the Burnley game
Bench Mee (West Ham away) or Shaquiri (City at home) for this gameweek?
Mee.
Bench Ben Mee and sell him asap
Reckon I can roll the FT over this GW and do 2 on GW31? Will have to wait for CL to see what to do with Son and Sterling.
Pope
VVD Davies Otamendi
Salah Sterling Son Walcott
Kane (C) Firmino Quaner
Elliot Ogbonna Duffy Shaqiri
1FT 0.5itb
Good plan
I think I've cracked it! Still got wildcard, FH, BB and 2FT's.
Foster/Adrian
Alonso VvD Vertonghen
Salah Mane Sterling Gross Walcott
Kane Wilson
Mee Quaner Mariappa
GW30 Mee>Bauer + roll FT
BGW31 Gross>Shaqiri + Kane>Benteke
Gives me an X1 for free;
Foster
VvD Mariappa Bauer
Salah(c) Mane Shaqiri Walcott
Quaner Benteke Wilson
GW32 Benteke> Kane or wildcard
🙂
Have 2FT's this week
A: Keep De Gea for this week, transfer in Coleman & VLP
B: Get Lossl in for De Gea and one of Coleman or VLP (transfer other one next week)
A
Which option for the next 2 GWs?
A) Depoitre to King
B) Ransey to Shaqiri
C) Son to Siggy
D) Naughton to Coleman
Don't like Bournemouth or Stoke's fixtures for GW 30 so between C & D id probably say D
Why would you get rid of Ramsey when they will be home to Watford which will be a goal fest as Arse will want to make sure Deeney eats his words and I see plenty of goals........Ramsey was rested ( much to my frustration this week ) so that he can play Thursday in silly EUFA cup.......I wouldn't sell and I would wait till after Thursday game to make sure no injuries..........ARSEN OUT !!!!!!................COYI
he said ransey
Save FT for next week? Was going to take Mkhi or Sterling out but obviously both might play this week.
Two questions please:
1) Teams that have no game in GW 31 will have two games in GW 34 right?
2) Do we know which teams will have a game in GW 35?
We do not know when they will be arranged for yet, but expect it to be GW34.
Yes we do
Can't wait for the GW 31 disappointment after all these articles and planning. I remember experiencing a similar blank GW trap last season
The disappointment is not in GW31 itself, rather the lost alternative points this GW (e.g. Mahrez vs. Walcott )
Looking to ship out Jones, any suggestions on defenders that play GW31? Or maybe do Ogbonna, and keep Jones.
My team: 2ft, 1.4itb
Pope
Jones* Maguire Ward
Salah Sterling Mahrez Walcott
Kane Firmino Wilson
Pereira Duffy Milivojevic Ogbonna
Ogbonna > Schindler
Robertson
First time in a ffs cup. How do I find out what's going on?
Going to have 9 players for GW31 with a -4 (3 transfers on GW31), hoping it doesnt hurt my current 76K OR too much, all chips in tact.
Butland
VVD Bauer
Salah Walcott Milivojevic Stanislas
Firmino Quaner
