To see less ads Register or Login ----- Daily Fantasy Sports games 18+

FPL Discovery 2014/15 review

A Fantasy Football forum for news on fantasy football games run by the Premierleague (FPL).
Post Reply
User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

FPL Discovery 2014/15 review

Post by Mav3rick »

I'm still reading through a lot of it but there are some great stats and observations in the FPL Discovery 2014/15 review of the top 10k teams. See article here:

https://fpldiscovery.wordpress.com/summary_14_15/

The statements about wildcards certainly ring true to me and I felt their effect in my own teams ranking post DGW34:
FPL Discovery wrote:Wildcards were weird this season. Keeping your wildcard till GW34 was a good idea if you were doing well without playing it in the first half of the season. In fact, DGW 34 was by far the most popular time for playing a wildcard for the top 10K samples: 18% played them that GW vs. only 10% in GW4, which was the second by popularity. This is very different to last season: only 4.5% wildcards were played in top 10K in the gameweek that was most popular for wildcarding in the second half of that season.
Also interesting was that the top 10k turned over less than last season due to points gaps being bigger, even though there were less points in FPL overall. I'm not sure what that means, it could be that the lack of a Suarez type of captain increased the skill levels, but I've always thought of the captain as a way to reduce points gaps by increasing randomness in scores.
FPL Discovery wrote:Even though this season was not as high scoring as the previous one, gaps in points between teams were greater.

For example, at the end of the first half of the season distances from the 10,000th team to 1000th, 100th and the 1st one were 54, 93, and 178 points. Last season, the corresponding numbers were only 44, 79, and 143 points respectively.

During the second half, this difference between the two seasons only deepened: after the season finished, these gaps were larger than a year before by 22, 27, and 39 points respectively.

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

Just to add a few of my own comments on other data in the FPL discovery stats:
  • It seems that by season end, only about 20% of teams were active, so that's about 650k active managers meaning that a top 10k finish puts you in the top 1.5%. Obviously you'd need to finish approximately top 6.5k to be in the top 1% of active managers.
  • It looks like there was an average of about 9 points hits over the season for top 10k managers, so roughly one hit in every 4 GWs. I know personally that I took 4 hits, of which two were for DGWs so I would suggest that we can take off at least a couple of them as lower risk hits for extra players. Even so, 7 additional hits seems quite a bit to me. It would be interesting to see how that was distributed in the top 10k (i.e. were there half the teams that took loads of hits and half that took none?).
  • It looks to me that most January widcards were played in the penultimate and the last week of it being available.
  • The general wildcard was played mostly in GW34, with the second most popular time being for GW4 in the traditional money making window. I think that maybe the slow price rises stopped players using their wildcards in GW4 as the potential for manking cash wasn't there. With such low squad values (105.8 compared to 110.6 last season) the money factor could go two ways. Either the extra cash is not enough to buy higher bracket players, or the extra cash is so valuable that the 1.0 or so extra you might be able to get by playing an early wildcard is decisive. With Kane around this year, money was never a factor, but next year that could well be different. Right now I'd favour holding onto a wildcard next year rather than playing it early since I don't know that the additional squad value will go far enough to make a difference.
  • There's an interesting point on homogeneity where the top 10k peaked at owning 7 out of the 10 most owned players. That still leaves 4 spots for the lesser owned players and is probably where Spiderm4tt's idea of making sure that you own better differentials than your competitors comes in.
  • Going with the most popular captain scored around 6 points per game with the armband. That's not bad IMO and gives you consistency, but the main reason to go with the crowd is avoiding big red arrows. The one thing against this viewpoint is that only 3 times (out of 19) did a very well supported captain hit a score that would be likely to see a big red arrow if you'd not captained him. As a side note, on average the top 10k's captain scored 6.2 ppg so you see that picking a top voted captain (when over 50%) correlates closely to this figure, even though only 19GWs had such a candidate available. That would suggest that the other 19 GWs (when there was no clear choice) also averaged around 6ppg.
  • Forwards were again the dominant force in terms of PPG (including captaincy), while keepers perhaps slightly surprisingly outscore the defenders on a PPG basis, although not by enough that you shouldn't continue to consider the keeper and defenders as a unit.
  • Despite what The Rounded Keeper might have you believe, there are not thousands of lucky managers getting good points off the bench when their players no-show. Indeed the average off the bench score was less than appearance points at 1.36ppg. This suggests that it's mostly defenders coming off the bench and shows that you simply should not tolerate any serious rotation in your backline if you have a "normal" bench.
  • The issue about points gaps being bigger this year between top 10k, top 1k and top 100 is interesting. The number of same players in a team hasn't really changed from last season so you can't say that it's FISO/FFS making teams all the same that's the problem. I'm not sure why the points gaps have grown, any ideas?

User avatar
thekrumcake
FISOhead
Posts: 835
Joined: 25 Sep 2012, 16:15

Re: FPL Discovery 2014/15 review

Post by thekrumcake »

Great commentary!

User avatar
thekrumcake
FISOhead
Posts: 835
Joined: 25 Sep 2012, 16:15

Re: FPL Discovery 2014/15 review

Post by thekrumcake »

My theory is that the big points gap has happened by chance, because some opportunities have payed off way, way above expectations, eg. holding on to Costa for the first injury scare, the infamous Agüero vs. Kane captaincy GW, captaining Benteke/Austin for the DGW, etc.

User avatar
liquidfootball2
Dumbledore
Posts: 8672
Joined: 05 Jan 2012, 16:14
FS Record: Best fpl finish 233 in 14/15

Re: FPL Discovery 2014/15 review

Post by liquidfootball2 »

Thanks Mav3rick i appreciate that summary, very useful. I think i was pretty much typical of the top 10 K in that i generally just captained whoever came top of the FISO/FFS or latterly transfer hub polls. I also used about 8 hits which was higher than i intended but fits in well with the 10 K stats.

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

It's a fantastic blog that, it really is. And superb summaries too Mav, kudos. :D
Mav3rick wrote: [*]The issue about points gaps being bigger this year between top 10k, top 1k and top 100 is interesting. The number of same players in a team hasn't really changed from last season so you can't say that it's FISO/FFS making teams all the same that's the problem. I'm not sure why the points gaps have grown, any ideas?[/list]
My vague uneducated guess would be one of three possible factors:

1) The new slow price rise system and resulting low squad values maybe made it harder to maintain a coherent squad with high quality in all areas for large spells of the season, meaning that skill was possibly a bigger factor as the 'better' players who could manage this could accelerate away quicker. It would be interesting to see if there was a particular spell of the season when the top 10k average accelerated away from the rest of the field faster from the rest than normal or whether it was a season long thing (possible the pre-Kane spell). Problem is, you could just as easily say that the skillful players didn't get their reward in terms of squad value which may have made it harder for them to pull away as quick as normal - but possibly the fact that squad value wasn't needed at the big end of season GW34-38 bonanza due to Leicester players etc meant something, but not sure what. It could just be that the DGW at the end (and surrounding weeks of consistency from some of the players involved) gave a huge boost to the top 10k compared to normal as it was the more unusual players (Leicester/Hull) who got all the points instead of the player that more regular players would have lingering in their teams anyway.

2) The captains/'must have' premiums were possibly a bit more in and out of the team than normal (Hazard/Sanchez were poor at the start, Costa was in and out through injury etc). Last season basically Suarez/Aguero had your captaincy sorted all season, for everyone.

3) It was a bit of an odd year imo in that some of the cheaper players seemed to be much more consistent than normal, especially cheap defenders. Maybe the 'better' players were quicker to adapt and ditch the known quantities for these new gems than the masses, who would possibly stick for longer with under-performing tried and tested big assets like United players, Yaya etc. Or maybe the opposite, maybe the masses would jump onto cheap bandwagons too quickly because when big guns were underachieving whilst the 'better' players either kept the faith or choose a cheap bandwagon more wisely - it's much harder to decide between the midprice/cheaper players than the premium ones imo.

Bottom line, not really a clue - it's a bit odd. 8-) :lol:

User avatar
Carlos Kickaball
Dumbledore
Posts: 7801
Joined: 04 Sep 2013, 18:02

Re: FPL Discovery 2014/15 review

Post by Carlos Kickaball »

Thanks for the write up Mav. :)

Increased decisions will just increase the variance in points driven by BOTH skill and luck, particularly right at the top of the game. Yes it will give good players the chance to differentiate themselves more from average players, but it will also drive a larger difference between similarly skilled players who always had a good chance of ending in the top 10k.

Think about it, you're not seriously suggesting that the reason the top 100 managers got more distance ahead of Triggerlips and Ville (in the 101-1000 category) and all of those got further ahead of Rasmus Sundman and Ulrik (in the 1001-10k category) is because these managers in the 101-1k and 1001-10k category are unable to perform good fantasy management?

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

I don't really know what you're saying there (and won't be drawn into taking about the 'l' word if at all possible ;) ). Are you saying that your guess at the reason for the 'spread anomaly' is that there have been an increased number of decisions made this year? Why do you think that has been the case? The transfer numbers are pretty static compared to last season suggesting managers were in fact making roughly the same number of decisions, no?

One thing I was kindof implying with a suggestion of possible premium-player inconsistency or the possible need for having to choose between the multitude cheap players is that maybe decisions were in general slightly more difficult this season. That may explain it. But more decisions? I just don't think that has been the case. :?

User avatar
Carlos Kickaball
Dumbledore
Posts: 7801
Joined: 04 Sep 2013, 18:02

Re: FPL Discovery 2014/15 review

Post by Carlos Kickaball »

What I mean is more meaningful decisions that actually split opinion (and it's not just transfers, often it's the absence of a permacaptain as you alluded to).

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

Carlos Kickaball wrote:What I mean is more meaningful decisions that actually split opinion (and it's not just transfers, often it's the absence of a permacaptain as you alluded to).
Ah, so you agreed that increased decision difficulty could be a factor in this spread anomaly thing. Then, what is this all about?:
Carlos Kickaball wrote: Think about it, you're not seriously suggesting that the reason the top 100 managers got more distance ahead of Triggerlips and Ville (in the 101-1000 category) and all of those got further ahead of Rasmus Sundman and Ulrik (in the 1001-10k category) is because these managers in the 101-1k and 1001-10k category are unable to perform good fantasy management?
No, I am not suggesting that. Obviously. Don't make me use the 's' word Mr-serial-complainer-about-apparent-'s'-word-arguments... :wink:

User avatar
Archy
Grumpy Old Man
Posts: 3210
Joined: 22 Oct 2011, 10:09
FS Record: Maybe all is not lost after all

Re: FPL Discovery 2014/15 review

Post by Archy »

Top analysis Mav
It seems that by season end, only about 20% of teams were active, so that's about 650k active managers meaning that a top 10k finish puts you in the top 1.5%. Obviously you'd need to finish approximately top 6.5k to be in the top 1% of active managers.
I disagree with the view (not saying it is your view) that inactive players should be excluded from the analysis of one's performance. If someone wins a marathon, do they say they beat the number of people who started the marathon, or the number of people who finished it. Silkworm has said he's given up by Xmas in previous seasons where he got off to a bad start. If he gives up next Xmas, would he not count as someone you've managed to beat next season?
Postby thekrumcake » 04 Jun 2015, 11:46
My theory is that the big points gap has happened by chance, because some opportunities have payed off way, way above expectations, eg. holding on to Costa for the first injury scare, the infamous Agüero vs. Kane captaincy GW, captaining Benteke/Austin for the DGW, etc.
I totally agree with this. The remarkable performance of Leicester players following their DGW can also be added to the list.

User avatar
Carlos Kickaball
Dumbledore
Posts: 7801
Joined: 04 Sep 2013, 18:02

Re: FPL Discovery 2014/15 review

Post by Carlos Kickaball »

No Stem it's obviously not a straw man, as I asked if you were suggesting that (it even has a question mark :lol: ).

It was to remind you that the idea that the spread between positions like 100 and 1k is to mostly do with skill rather than variance, is slightly ludicrous when the managers like Ville, Trigger, Ulrick and Rasmus all finished 100-10k.

I'm in agreement with Archy and Krumcake that the spread at the elite end of the game is mostly down to chance.
Last edited by Carlos Kickaball on 04 Jun 2015, 15:03, edited 1 time in total.

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

Archy wrote:
It seems that by season end, only about 20% of teams were active, so that's about 650k active managers meaning that a top 10k finish puts you in the top 1.5%. Obviously you'd need to finish approximately top 6.5k to be in the top 1% of active managers.
I disagree with the view (not saying it is your view) that inactive players should be excluded from the analysis of one's performance. If someone wins a marathon, do they say they beat the number of people who started the marathon, or the number of people who finished it. Silkworm has said he's given up by Xmas in previous seasons where he got off to a bad start. If he gives up next Xmas, would he not count as someone you've managed to beat next season?
I think that's probably fair to some extent, though I don't think it really matters about what percentile you finish in anyway. What I would say is that I don't particularly take any pleasure from finishing above someone who hasn't bothered to finish the season, as unlike a marathon it seems a pretty basic and easy thing to do.

Welcome back btw Archy!

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

Carlos Kickaball wrote:It was to remind you that the idea that the spread between positions like 100 and 1k is to mostly do with skill rather than variance, is slightly ludicrous when the managers like Ville, Trigger, Ulrick and Rasmus all finished 1k-10k.

I'm in agreement with Archy and Krumcake that the spread at the elite end of the game is mostly down to chance.
CK - I think I started out agreeing with you, but you're closing in on luck vs skill again with the above which is a different topic for that legendary thread....

Anyway, in this particular case I don't think that the skill vs luck question comes into it since we're comparing like-for-like skill/luck level - this seasons top 10k to last seasons top 10k.

You'd expect that as groups they would all be equally affected by the same combination of luck and skill, so both factors can be ignored. The issue is what caused the extra 22 points spread between this years and last years equally skilled/lucky players (the actual points spread from 1k to 10k was something like 100 points and that's down mostly to skill IMO, all I'm looking at is that last season the gap was around 80 points despite there being more points overall - what's changed)?

If a correct captaincy decision was worth 4 points (going on the average top 10k captaincy score of 6 points) and there were 19 GWs where there was no candidate with >50% captaincy rate, then all you have to do is get lucky and nail the captaincy call 15 times instead of 9 to generate a 24 point gap.

I'm not sure if that would just increase the gaps in the top 10k or push different more lucky managers into the top 10k however. Since it would be a completely skill-less decision in my eyes (we are literally doing a captaincy coin-toss) it would seem likely to me that it just increases everyones score by the same factor, thus inflating the gaps.

Happy for a better theory to be put forward, however if we end up agreeing with the concept that it's the extra captaincy decisions that caused the increased gaps then this supports the idea of going with the same captain as everyone else to reduce the gaps and give your skill the best chance to come through.

User avatar
Mo Bot
Dumbledore
Posts: 7208
Joined: 15 Oct 2005, 13:59
FS Record: Inaugural Last Man Standing. FISO Fantasy Eurovision champ 2012. World #1 in UFPL for a whole week

Re: FPL Discovery 2014/15 review

Post by Mo Bot »

Postby thekrumcake » 04 Jun 2015, 11:46
My theory is that the big points gap has happened by chance, because some opportunities have payed off way, way above expectations, eg. holding on to Costa for the first injury scare, the infamous Agüero vs. Kane captaincy GW, captaining Benteke/Austin for the DGW, etc.
I got all three of these wrong which I now count as 120 lost points (Costa for GW3 and 4) and was the difference between 9503rd place and 758th.

Getting rid of injured Costa was a sound decision but the "not listening to Jose" rule was employed thereafter.
Aguero over Kane was a judgement call and I still believe was the right one.
Not captaining Austin for the DGW was just stupid....especially as I went with Guzan.
3 key decisions that took me from being a top 1k player to just inside the top 10k.

User avatar
liquidfootball2
Dumbledore
Posts: 8672
Joined: 05 Jan 2012, 16:14
FS Record: Best fpl finish 233 in 14/15

Re: FPL Discovery 2014/15 review

Post by liquidfootball2 »

Archy wrote:
I totally agree with this. The remarkable performance of Leicester players following their DGW can also be added to the list.
Welcome back Archy,

The Leicester surge was a massive bonus for late wildcarders and one where it was difficult for others to share.

You could also add surprise performances from the apparent minnows in the late dgwks - Hull and Sunderland, although I did manage to double up on the Wearsiders defence.

User avatar
SuperGrover
Grumpy Old Man
Posts: 1540
Joined: 12 Sep 2012, 15:38
Location: Chicago, USA

Re: FPL Discovery 2014/15 review

Post by SuperGrover »

"Forwards were again the dominant force in terms of PPG (including captaincy), while keepers perhaps slightly surprisingly outscore the defenders on a PPG basis, although not by enough that you shouldn't continue to consider the keeper and defenders as a unit."

I don't find that surprising at all. The amount of defender attacking points, when spread across typically 4 defenders per team, is very small. On the other hand, keep points for saves are regular with the occasional PK save adding to the total. To me, it's no surprise that GKers outscore defenders on average. it's also not surprising that the top scoring defenders well outpace Gkers on a PPG basis.

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

The interesting thing is that its not a figure of overall ppg for all defenders, its ppg for defenders selected in FPL teams. Since top 10k teams aren't selecting poor players (you'd hope) then that figure includes a Terry\Ivanovic type plus perhaps a Saints defender and then the best of the cheapies.

I find it surprising that our keeper selections outscored the average of those. Unless I mis-interpreted the FPL Discovery numbers.

smarty_pants
Cheetah
Posts: 3
Joined: 06 Jun 2015, 14:27
FS Record: FPL

Re: FPL Discovery 2014/15 review

Post by smarty_pants »

Nice to see my stats get some proper analysis, great thread ;)
Mav3rick wrote:The interesting thing is that its not a figure of overall ppg for all defenders, its ppg for defenders selected in FPL teams. Since top 10k teams aren't selecting poor players (you'd hope) then that figure includes a Terry\Ivanovic type plus perhaps a Saints defender and then the best of the cheapies.

I find it surprising that our keeper selections outscored the average of those. Unless I mis-interpreted the FPL Discovery numbers.
Nah, I think your interpretation is correct. My view is that
1) The difference is too insignificant (0.08 points). Scrap DGW 37 when a goalkeeper averaged 9.76 points and a defender averaged 5.98, and this difference is gone.
2) There are not as many attacking points for defenders as one could imagine. Even if you got an assist for one of your defenders each week (well, did you? ;) ), it would mean only 1 attacking point per average defender (if you play 3 defenders), a goalkeeper needs to make just 3 saves to match that (and that's a pretty standard scenario). And now that FPL have changed their bonus system, goalkeepers were more or less just as likely to get bonus points as defenders (therefore the difference compared to last season).

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

Great to see you posting here @smarty_pants. I and many other members of the forum find your site a hugely valuable resource - thank you for taking the time to maintain it and to provide the graphs and analysis.

FPL Discovery is one of the few FPL sites I check every week and I'm glad we can thank you in person for it!

I think I still find it surprising that the GK out averages the best three defenders you can field each week. I take the point about attacking returns not being worth much but there are also the cleansheets to consider... Most of us take our keepers from cheap teams so I still find it counter intuitive that they would be equal points scorers on a ppg basis with your three best defenders.

It must mean that the cheap defenders are a huge drag on the overall defender performance in a team, despite our attempts to rotate that position around.

I don't think it will change my selection policy in defence next year but maybe it will persuade me to pay more attention to my keeper transfer strategy rather that viewing it as a set and forget until wildcard strategy.

User avatar
Carlos Kickaball
Dumbledore
Posts: 7801
Joined: 04 Sep 2013, 18:02

Re: FPL Discovery 2014/15 review

Post by Carlos Kickaball »

Yes, thanks Smarty Pants! I was sure you had posted here before, but it says you only have one post, if this is your first one welcome. :D

I was hoping your scanning of the selection of top teams in FPL could help me find Mav's team link next year actually. ;) [joking obvs]

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

Here, here. Yes, congrats smarty_pants on a really superb blog. I was hoping leaving a comment there might lure you over. :p

Hopefully you will stay around and post during the season proper as it would be great to have a bit of extra bonus commentary on your stats/graphics. I can't praise the blog enough to be honest - invaluable stats. :D


We've had a nice week or so of 'fpl celebs' on FISO with smarty_pants, triggerlips and ChrisA (FFS HoF & RMT) to add to regulars spiderm4tt and nigel (of FPL Statistics). All we need is Ville to pop his head in now...

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

Stemania wrote:We've had a nice week or so of 'fpl celebs' on FISO with smarty_pants, triggerlips and ChrisA (FFS HoF & RMT) to add to regulars spiderm4tt and nigel (of FPL Statistics). All we need is Ville to pop his head in now...
I was just thinking that. I'm hoping that now we have celebrities posting I might finally get to advise Scarlett Johansson on her front line.

smarty_pants
Cheetah
Posts: 3
Joined: 06 Jun 2015, 14:27
FS Record: FPL

Re: FPL Discovery 2014/15 review

Post by smarty_pants »

Carlos Kickaball wrote:Yes, thanks Smarty Pants! I was sure you had posted here before, but it says you only have one post, if this is your first one welcome. :D
When I registered today it said to me that the nickname "Smarty Pants" had already been taken, so this may well be the case. But I don't remember it :)
(however, I remember trying to register a year ago to say thanks for the invaluable data that I had found in the archives of this site for my post about FPL history, but there was some problem with activating an account back then)

User avatar
Billy Bongo
FISO Knight
Posts: 12000
Joined: 21 Jul 2013, 22:18

Re: FPL Discovery 2014/15 review

Post by Billy Bongo »

liquidfootball2 wrote:
Archy wrote:
I totally agree with this. The remarkable performance of Leicester players following their DGW can also be added to the list.
Welcome back Archy,

The Leicester surge was a massive bonus for late wildcarders and one where it was difficult for others to share.

You could also add surprise performances from the apparent minnows in the late dgwks - Hull and Sunderland, although I did manage to double up on the Wearsiders defence.

Why? One transfer and a spend you had 2 Leicester players. Plus they were cheap so no need to reorganise your team. Wildcard had nothing to do with it, thats just the excuse those that were to slow to jump on use.

Also, how are they deciding on dead teams? No transfers for 2 weeks, 4 , 6 whatever? 650 k seems very low

User avatar
liquidfootball2
Dumbledore
Posts: 8672
Joined: 05 Jan 2012, 16:14
FS Record: Best fpl finish 233 in 14/15

Re: FPL Discovery 2014/15 review

Post by liquidfootball2 »

Yes Billy i agree it is an excuse for some poor play. I was too preoccupied with getting Arsenal/Liverpool double gwk players and missed the obvious source of points.

Almost like not seeing the wood for the trees and placing too much of a priority on future points and not enough on what's right in front of me.

User avatar
Mav3rick
FISO Jedi Knight
Posts: 20858
Joined: 20 Jul 2009, 20:35
FS Record: FPL: 1082, 1201, 1800, 10203

The stats are dark and full of errors.

Re: FPL Discovery 2014/15 review

Post by Mav3rick »

FPL Discovery now have some more excellent stats regarding the top 10k

https://fpldiscovery.wordpress.com/2015 ... #more-3375

I'm especially captivated by the points hits vs average rank graphs, which seem to show a general trend of worsening average rank starting from 3 or 4 hits up to 22 hits (where the data becomes harder to interpret due to the reducing number of teams taking more than 22 hits). There really isn't a great deal of difference between 1 hit and 10 hits though, just the "optimal" trough at 3/4 hits and everything else is the 1-10 range is pretty equal.

There's an interesting note that there is also a worsening average rank for teams taking 2, 1 and 0 hits on average than by taking 3 hits (which has the lowest average rank, apart from outliers).

Obviously that's not a hard and fast rule and even the winner took no hits this season, but it does go to show that refusing to take any hits is not what you could call an optimal behaviour in most teams and that all hits need to be considered on their own basis. It's probably quite an obvious thing to say, but the right hits will aid your team while taking more than you need, on average, seems to be worse for your chances of a better rank.

User avatar
liquidfootball2
Dumbledore
Posts: 8672
Joined: 05 Jan 2012, 16:14
FS Record: Best fpl finish 233 in 14/15

Re: FPL Discovery 2014/15 review

Post by liquidfootball2 »

Great point Mav3rick. FPLdiscovery is a cracking site and their analyses well worth a read, seem to learn something with every one.

I'm going to try and heed their findings and cut down on hits next season, took eight last season and although injuries may have forced my hand, probably too many.

User avatar
Stemania
FISO Jedi Knight
Posts: 20448
Joined: 27 Aug 2006, 11:54
Location: On the Iron Throne of xG, the seat of The Crown Prince of the Stat Perverts. Or if not, in the STC!
FS Record: Best: TFF 321st. FPL 129th. FFS Career HoF Peak 2nd (Live 1st). Ability since lost.

Re: FPL Discovery 2014/15 review

Post by Stemania »

Absolutely top draw stuff again smarty_pants. :D

Really interesting. especially the bit about hits which Mav has expertly described and analysed. :mrgreen:

I also like the observation that "75% of the final top 10K teams were inside 25K by the season equator". I wonder if generally speaking this should signal 25k should be a midpoint marker if you're aiming for a top 10k. I do actually find it quite surprising as I personally always aim for the higher 35k by the slightly later time of when I play the January wildcard as I always feel that puts me well on course for a top 10k. (This may sound suspicious considering I was 34,714 this year at that point - as well as bizarrely 35,306 at the half way point - but I assure you it's a genuine annual precise aim of mine, and only an amusing and satisfying coincidence - basically that's where I was for my first ever top 10k finish).

And in Rob., another really outstanding record highlighted - lowest rank 2579 in 7 years is crazy!

View Latest: 1 Day View Your posts
Post Reply

Return to “Fantasy PremierLeague.com (FPL)”