Wednesday, 5 October 2011

Stats So Far


BEST XI (3-4-3)
De Gea 92
Williams 62.5
Rangel 56
Jones 54.5
Nani 83.5
Silva 82
Young 77.5
Dempsey 77.5
Rooney 135
Aguero 115
Dzeko 95.5
TOTAL 931
OTHER TOP SCORERS
Van Persie 86
Hart 77.5
Suarez 69
Vorm 68.5
Lampard 67.5
Di Santo 65
Given 65
M.Petrov 63.5
Van Der Vaart 63
Anderson 60.5
Agbonlahor 60
AVG 10+ (MIN 3 GAMES)
Rooney 22.5
Aguero 16
Dzeko 15.92
De Gea 15.33
Sturridge 15
Zamora 14.75
Adebayor 13.75
Young 12.92
Van Der Vaart 12.6
Van Persie 12.29
Nani 11.93
Silva 11.71
Lampard 11.25
Bramble 11.17
Hart 11.07
Dempsey 10.36
Ba 10

This post has been compiled without any real aim in mind, other than to see if we can draw anything from early season form. Just to get you thinking really, but do any of the these stats mean anything in terms of possible patterns?

We can see that two Swansea players are ripping shit up when it comes to defenders, and many keepers are having a great time - DDG in particular, but also Vorm, Hart & Given - which is very different to last season when most keepers struggled.

It's clear that dropping Silva or Aguero early would've been a mistake at this point, you can see defenders score far less than attacking players (der!) which seems to enforce the reason so many of us spend light on our back three. Premium midfielders are scoring well too, but strikers really are dominating the points thus far, with Rooney & Aguero ahead by miles (despite a zero each already).

Of course, we need to take "price per point" in to consideration before we can make any definitive analysis. It would be great if we had a reporting/stats systems of early values/points to see true patterns of price per point since Week 1, because player price fluctuations are so important in YFF, but we can't without some serious nerdy manual work (and I'm not sure the player values shown on the "weekly" individual player stats pages are 100% accurate anyway), so from now on we need to look to the future based on current prices.

One thing I would think is ... IF you hold DDG, Silva, Aguero, Adebayor & Rooney at their early prices of 6, 8, 6, 6 & 17 (ish) respectively, and they keep their fitness & anything close to their early form (i.e. average points), then you could have a big advantage over other managers in the next few months. But that really is a big IF, and I may be clutching at straws (because I have all four out of the five of them!) :)

Anyway, I'll be back with some more on this tomorrow, but I'd like to hear your thoughts on anything you think you've picked up on.

Cheers,
AM

26 comments:

capiTeno said...

it would be nice to have the option to see and sort by all the scoring stats like it was about 4/5 seasons ago
i always liked to have a winger in my team for successful cross and corner won 1pt, m.pet has 30 of 63.5 pts this way (0.5pt more than VdV !)
i think might be that nerdy manual worker ;-)

Assistant Manager said...

Completely agree Capt - sorting by Assists/Successful Cross was so handy, but they were replaced by Corners Won/Fouls Won 4 seasons ago. Makes no sense for such a detailed game; in fact it is actually crazy that such key stats aren't available to sort by.

Data Head said...

Nik -

This is just my sort of thing, and since I keep track of all the player points anyway, it took me little over an hour to create a spreadsheet showing the profitability of each player you listed above (in the same order). I uploaded it as a Google Document here. Anyone can view it, and I think download it as well to manipulate the data. A few key points to make:

- I took the base price to be what a player had cost at the deadline for week 1. For this reason, Rooney, for example, is listed at 19.54 million, and I calculated his profitability on that value. If you have held him at 17.xx since before the deadline to week 1, then that obviously makes him more profitable.

- I color-coded the whole thing based loosely on traffic signals. Red = less than 50% return on the original investment (Frank Lampard was the only one), Yellow = 50-99% return, and Green = return of 100%+

- If a player did not play, then the box for that week is blank. Following up on this, I did not start calculating player value until that player had played his first match. It would not make sense to take the cost of Adebayor and multiply it by 7 when he has only been playing since week 4 (when any serious manager might have picked him up).

You know, I think I have rubbed a lot of your readers the wrong way by leaving comments that redirect people to different sites (often my own), but sharing this sort of information is all I have ever wanted to do. I hope anyone who reads the document finds it interesting and can extrapolate something worthwhile from it.

By the way, this season has been a harsh one for me. Just when I thought I had things figured out...

Cheers,
D.H.

P.S. Have you have checked out Chatango and Cbox as possible alternatives to the underwhelming Chatroll?

Bitsketchy said...

Hi Nik, I'm running loads of stats at the mo and the scary one is that to date you had a 50/50 chance of scoring zero or below on any given keeper!

Coyotedelmar said...

I was looking at some of this earlier in the week, and one thing I did notice is, so far, there are quite a few players high up on the average points led mainly by one or two big games, and very little the rest of the time. Have to expand on some of it when I get home later though.

De Gea being top isn't too shocking though, Manchester United is actually give up 97 shots so far, tied with Bolton for most allowed. Of course, unlike bolton, the shots are being saved and United is scoring, so you get a lot more points then you would with Jussi!

Zonal Marking gave their opinion on why there are so many shots against Utd. which was kind of interesting.

Nicky said...

Data Head thank you that is very interesting. The ratio works for all players that have stayed above their base price from the very start but Lamp's price has gone down to 17.XX at times in the season so you'd get that investment back, meaning his total investment is less than 140; I imagine he's the exception as most of those others have remained above base

Assistant Manager said...

Great stuff D.H, thanks mate. I'll start checking out some of the Chat Alternatives today!

Coyotedelmar said...

Kind of simple look (and not the way I was planning to go) on the best XI (working on the top 10). Its basically trying to see what stats are main point getters for the top players. Lot of them are obvious but it helps right? =P

De Gea: 69.6% of points from saves (64 pts of 92); 21.7% from Wins

A. Williams: 38.4% from CS's (24); 32% from Blocked Shots (20)
Rangel: 42.9% from CS's (24); 19.6 from Passes Intercepted (11)
P. Jones: 33% from CS's (18)

Nani: 25.1% from Goals (21); 18% from SOT (15)
Silva: 32.9% from SOT (27); 24.4% from Assists (20)
BFAY: 27.1% from SOT (21); 25.8% from Assists (20)
Dempsey: 45.5% from SOT (33); 19.3% from Goals (14)

Rooney: 46.7% from Goals (63); 37.8% from SOT (51)
Aguero: 50.2% from Goals (56); 29.6% from SOT (33)
Dzeko: 44% from Goals (42); 34.6% from SOT (33)

Still needs a lot of work granted, but it seems like it was mostly straight forward. Forwards get most of their points from Goals and shots on goal (minus RvP and Suarez), Defenders get it from Blocked shots and CS's, Goalkeepers get it from Saves, CS, and wins.

Data Head said...

@Nicky - You're right about Lampard, and I did think about that as I put the chart together. I also used the Adebayor price from week 1, for example, which was something like 0.12 million below what it was a couple weeks later. But in the end, I decided to stick to the KISS principle: Keep It Simple, Stupid! That is, it would have been impossible to know exactly when or at what prices managers had bought or BD'd players at, since prices can fluctuate a couple times each weekend. But this spreadsheet was a general guide that I thought would complement Nik's article nicely.

@Nik - I'm impressed by the vocal community you were able to create surrounding YFF. The dialogues, both below the articles and in the chat room, are fantastic. I hope that once the "blog-omerate" comes together, the communities from the respective sites (realizing that there is probably quite a bit of overlap) will come together, and that some sort of chat room will become integral to that site, as well. Great work!

- D.H.

Bitsketchy said...

Hi everyone,

Does anyone actually have a full database or spreadsheet of a seasons data for yff? there's loads of things I want to try but we've only got 7 weeks data so far and I don't really want to have to wait until after the season to get the full story....

ahmadkhairulnizar said...

@nik.. so, you think what is the best pick from the players name that you mentioned above..that can save us during the 2nd round of Blog Cup.??

Eldwin said...

You know I've always loved reading all these stats, but how effective they are is yet to be proven? (Or someone has? :P) What I'm trying to say here is that say even if you get average for each player, this doesn't take into consideration for the fact that in a particular week where such player score low and you pick another which scores higher, seems like that would pay off more. And really the average only applies to your team if you hold that particular for most of the weeks, or at least weeks that they actually score.

Anyway, even though they are not 100% correct, oddly enough you have stats in real life say, one team has never won against a team in a particular ground for X amount of games or years or so. If you think about it, even though over the period the stats have been accumulated managers or players have changed (on both teams), they still stand true! Something to ponder about! :)

Bitsketchy said...

Datahead, is it possible to get your spreadsheet of players values over time?

Data Head said...

@Bitsketchy - I kept those stats for a while last season, but it was a tedious process because Yahoo! kept booting me out for hours at a time when I would click through the week-by-week player stats. (I guess the system took me for some sort of malicious data-scrubbing program.) After a while, I just gave up and deleted everything. I was also calculating value for each player across five weeks, based on the then current price, but it was taking too much time (especially considering that I was doing this for both YFF and YFF Superliga). Now, I just track team and individual home/away points and performance across five weeks, along with some of the key point-earning actions, such as goals and SOTs for outfield players, clean sheets for defenders, and saves, goals allowed, wins, and losses for goalkeepers. I do keep all these data in a master spreadsheet and take pictures each week to post on my blog.

I have long hoped to meet other "data heads" such as myself to get together and create a larger stat-tracking project. By myself, just maintaining everything is taking 50+ hours a week. With a team, we could send one another the same spreadsheet and have different people add different numbers. For example, one person could be responsible for points, one for prices, one for the sortable actions, and even someone to update the spreadsheet for those nitty-gritty actions that are no longer sortable, such as successful crosses, cards, etc. (With enough people, the last task could probably use a division of labor itself... with 5 people, each one could be responsible for 4 teams a week.) The larger the data-gathering and -sharing community, the better or more comprehensive the stats would be, and the faster they could be shared.

Having said all of which, I have to take Eldwin's point above into account. So far, gathering a lot of stats hasn't done much for my own team, for example. But I guess I feel some sense of accomplishment when someone who is doing well lets me know that my work is appreciated, that it helped them think about players or issues they might not have. I guess I will leave things there for now. If anyone is interested in building a larger stat-sharing community, let me know. I certainly have a lot to contribute!

- D.H.

Ken M said...

“The quality of a person's life is in direct proportion to their commitment to excellence, regardless of their chosen field of endeavor.”

Vince Lombardi.

I am in awe of the commitment displayed here and the willingness to openly share all this hard work with others.

Bravo..

Bitsketchy said...

Data head, I'm in. Intact, I might have a shortcut using external data queries in access using some visual basic code I've got. That should automate it I think. Let me have a look tomorrow. If you need to talk off this site you can contact me at rallen@imawinner.co.uk

billthegrunt said...

Thanks Nik for starting a stats discussion, and to Data Head and Coyotedelmar for your good work. I collected players' weekly stats for all of last season and am thinking about doing it again this year. It takes a little time to get started but after the initial prep work it only takes a few seconds each week. The hard part is the double weeks, and I wish YFF would break those down by game instead of adding them up.

I think there's a lot to be learned from the stats, since even the best players can go dormant for weeks at a time before having a great game. That's why I'm not a great fan of the average (ppg) in YFF, it smooths out the highs and lows too much. The statistical mean is more helpful, since it takes a players high-scoring games into account. The best YFF players, IMO, are the ones whose mean is higher than their average. Then if that player hasn't had a big game in several weeks, I'd consider him "due."

I guess I'm a YFF anorak, but I'd love to know what numbers the U.S. owners at Liverpool are crunching. "Moneyball" comes to the UK.

Ken M said...

Not sure if this fits into the "stat" conversation but..

I see many of the top teams having players week in & week out getting zeros. Not intentional zeros and not just one. Sometimes 2-4 zeros and still producing above average weeks. This is far from isolated,and I only have the one season to draw from, but this appears to be much higher then what I saw last year.

Even though I agree it is hard to determine what is "intentional" it just seems like for some of the top players there has been a shift in strategy. High risk greater reward sort of thing.

Is this just from more teams being "loaded" with high quality players and rotation being part of the reason or, has there been a new approach to playing the game.

One example: Managers picking an entire squad, say all City players one week then going all Arsenal the next. Not sure I saw much of that last season.

Ken M said...

O.K.

A stat of my own. Please excuse and small errors in calculations.

Out of the top 10 teams on the leader board this week. Obviously, not the top (point) earner's as I am sure they had few zeros, but overall point total.

110 players fielded.

32 players earning zeros.

Ken M said...

Sorry..last one.

@Data head

I see how this can be addicting.

Total points scored by those 10 teams was 723.50 or an average of over 72 points per game. Yes..low, but STILL out performed my team by a nice margin.

Bitsketchy said...

Ok, me and data head have worked out a xls spreadsheet that automatically draws all of the data from yff, happy to share? Although, if Nik allows I'll happily run the stats and supply/explain the key details?

kwyjibo said...

@Ken M - Yes, it confirms the fact that, while from week to week there are players who get some big numbers, overall, it's more important to be consistent. The top teams rarely, if ever, have a top score for the week.

If you can average around 100 per week, you'll be up near the top. I don't remember how many teams averaged 100 per week last season, but it certainly wasn't the whole top 50. The team that won averaged 106.9.

I still haven't mastered the consistency part. I was close to averaging 100 per week after the first five weeks, but after the collapse of the last two weeks am now 113 points behind. I need one of those 200 point double weeks!

Assistant Manager said...

Hello mate, thanks to you and D.H. for all your hard work. I'd be happy to host it here within it's own post (giving you boys full credit of course) but would understand if you would both rather put it up on D.H's site, I'd still happily link to it. Otherwise feel free just to share the document link here/add key details in the comments sections - thanks again lads, it's stuff like this that makes the community so awesome!

Assistant Manager said...

Ken, I agree with kwyjibo, an average of 100 per week will get you top 50, currently the top guys are averaging 115-120!

It's been a freaky season; a lot of the obvious big names (De, Gea, Dzeko, Aguero, Young & Rooney to name 5) went crazy big in the first few weeks so anyone who picked them up for Week 1 (and a lot of teams had all 5 of them) and hung on for a few weeks stormed away at the top, and it has skewed the standings somewhat. I'm still convinced once injuries/ suspensions/ loss-of-form of the big players starts happening and everyone needs to start thinking outside the box that things will change a lot. It's been tough, but I still think it's going to be a really great season!

Eldwin said...

Definitely Nik. My aim week in week out is to maintain the average! Although this season's average is skewed so far, I'm guessing that it will go down as the season proceeds back to the 100 mark which, technically is realistic and reasonable with the fact that we do only have 100mil to spend every week. After all, it's all maths and statistics isn't it? :D

Having said that though, all these stats will help. I've always had weeks where I need to pick a player or two or something to fill up the roster. If there are no specific players that come to mind, what I'd so is search by points, then SOTs, then SCs etc. (Which is what most will do). Definitely with a more comprehensive stats with average, that would help massively. :)

Nobanno said...

These stats are really effective. The start of the NFL season is right around the corner and you know what that means. I am now gathjering the fantasy football cheat sheets for upcoming leagues.

Post a Comment