Soccer/Football random discussion & other leagues thread

[kungfuhustler]

problem is, with alot of my player and team analysis, i am contracted not to give this information away for free, so how much i can contribute will depend on what exactly we are doing

[kungfuhustler]

btw, an excellent sports data company is opta sports data (optasportsdata.com) that may be worth a look, although you do have to buy the products they offer.

They did an analysis of the premier league 2007/2008 season and based on the criteria they chose this is the team of the season

--------James-------------

Sagna--Vidic--Hyypia--Clichy

Ronaldo--Essien--Fabregas--Gerrard

-----Rooney----Adeybayor---


some interesting stats that decided this

James kept 16 clean sheets and made 138 saves
Vidic made 544 defensive clearences
Hyypia successfully found team mate with 81% of passes, when he played Liverpool won 63% of games, when he didn't they won 36%
Ronaldo scored 31, had 131 shots and embarked on 191 dribbles

[ThaSaltCracka]

yeah thats a pretty damn good team

[Entee]

Originally Posted by kungfuhustler

btw, an excellent sports data company is opta sports data (optasportsdata.com) that may be worth a look, although you do have to buy the products they offer.

They did an analysis of the premier league 2007/2008 season and based on the criteria they chose this is the team of the season

--------James-------------

Sagna--Vidic--Hyypia--Clichy

Ronaldo--Essien--Fabregas--Gerrard

-----Rooney----Adeybayor---


some interesting stats that decided this

James kept 16 clean sheets and made 138 saves
Vidic made 544 defensive clearences
Hyypia successfully found team mate with 81% of passes, when he played Liverpool won 63% of games, when he didn't they won 36%
Ronaldo scored 31, had 131 shots and embarked on 191 dribbles

Actually I thought he played the best season I've seen. No one that I talked to agreed. There's the data that proves i was RIGHT! He definitely looked more intense and more offensive(?!) than before. Good to see a fellow finn looking THAT good. Not much football success for us usually

[kungfuhustler]

Originally Posted by Entee

Actually I thought he played the best season I've seen. No one that I talked to agreed. There's the data that proves i was RIGHT! He definitely looked more intense and more offensive(?!) than before. Good to see a fellow finn looking THAT good. Not much football success for us usually

I've always rated Hyypia very highly so i don't know if this proves it is his best season, just he is still a superb defender. He reminds me of Shearer in that neither were ever very quick (nor did they need to be) so they didn't really lose much as they grew older, both still being very strong, so they didn't reallly decline physically, but just improved with experience, especially both with great positioning

Also, no country that has Jari Litmanen can complain about not producing quality players

[mmbt0ne]

Originally Posted by kungfuhustler

problem is, with alot of my player and team analysis, i am contracted not to give this information away for free, so how much i can contribute will depend on what exactly we are doing

What if you weren't involved in any of the data collection, just the analysis of it? Or even just the theory behind the analysis?

[valenzuela]

Chile is 4th on the Conmebol qualifiers !!!!!! SouthAfrica here we ****ing come!!!!!!!! We defeated Venezuela with a last minute and the entire country exploded!!
Our next match is agaisnt Brazil and we are ahead of them of the table so far!! ( brazil would destroy whatever team finished 4th in the concacaf if the qualifiers ended today, we would qulaify along with colombia, paraguay, argentina)
SouthAmerican qualifiers are awesome, Dunga is probably going to get fired, its true that we are only 6 games into the league but being out of the direct qualification zone is unacceptable fro them.

[JaredL]

To clarify what I had in mind and get some of these thoughts out of my head and down somewhere, here's what I've been thinking.

Club team ratings:
ELO which would start last year or maybe the year before and include all leagues that are eligible for the Champions League and UEFA Cup. Going back several years it is very tough to get good data, particularly from smaller leagues. If that were available, then going back would be fine except for the extra time commitment. One could probably include lower-tier teams which would hopefully have lower rankings based on getting beaten by upper-tier teams in domestic competitions. That's one issue that would need to be resolved - counting cup wins over teams that are in very low divisions.

Eventually the same could and should be done with the North and South American teams. Because they have international competitions such as the Copa Libertadores the same should be possible. I don't think there are matches between European clubs and those from the Americas where both sides are motivated, so I don't think a combined Elo-type ranking would say anything.

Some new stats:
Players: +/- type stuff (goals for and against when on the pitch), scoring stats adjusted for the situation (so a goal to go up 4-0 late counts less than a goal to go up 1-0 late), other things based on availability of video replay of matches and surely there are a couple things I'm forgetting here.

Teams: Goal stats in different game states (up 1, tied, maybe playing with ten men etc.), some kind of expected points

Regular stats, tables and information
The stuff you can see at sky sports or whatever your favorite sports news site.

Deeper content for bigger clubs
Because there are some teams where it is possible to watch all of their league matches, it's possible to go a lot deeper for them. I want to focus on objective stats like clearances (when someone on the opposing team has a touch that goes into the penalty area and the defender sends it out of the penalty area), assists (having a touch that goes to a teammate who then scores a goal), secondary assists (after the last touch the opposing team, other than the goalkeeper, gets at most x touches before a goal is scored) as well as fairly objective stats like tackles, pass steals and the like. Mainly I want to avoid things like chances and even shots which are fairly subjective.

The same could be done with big competitions for which all the matches are televised or available to view online. I'm honestly not sure if this includes the Champions League as I would have to see whether the UEFA pay-per-view allows you to view replays of all matches from the group stage onward. I suspect that this isn't the case. I don't think that it's possible to get every single premiership game either. It may be possible to get FA Cup matches from pretty far out on, but I suspect it's only the last 8 at best. So this would have to be the World Cup and European Championship finals only it seems.

Predictions

I'm still working on a system to predict the outcome of matches based on data available before the match takes place. Every week, perhaps after placing the bets , I would post this for whatever leagues the site will cover as well as league predictions done via simulations.

Discussion Articles:

I would take a look at things like which league is the best overall, whether a 2-0 lead is somehow dangerous, whether or not scoring just before half time is extra important, whether a direct-passing approach is better, whether some teams are actually set up to be tournament teams or if it's just variance, and whether taking long shots is a good idea.

As far as I can tell there really isn't a site out there with this type of stuff for soccer. There are for baseball, basketball, American football and hockey, but I haven't seen something similar for soccer, at least for free. The reason I asked about interest here is that 2+2ers are going to be the type of people that would like this stuff. I'm not sure how much interest there is since soccer fans tend to be resistant to a lot of this type of stuff. Obviously taking all of this on at once would be extremely ambitious. With nobody's help, I would probably work on some of the articles, the prediction algorithm, and the new stats for a few of the bigger leagues while very slowly getting the Elo thing done. In addition to whatever new ideas people would have, having people help would likely allow more leagues and some of the deeper stuff as well as getting the Elo ratings, and really everything, done at a faster time.

[JaredL]

Oh, yeah. I meant to add that if I actually get off my ass and get this going I won't use something like the pay for data services. They seem pretty convenient, but most of the stuff I have in mind uses public data that is easy to obtain and I don't want to have to deal with any sort of issues around putting up info that came directly or indirectly from them.

[Bjørn]

I would be interested as well though probably not in pure number crunshing which I'm sure there are some people a lot better at but I'd definitely be willing to give my take on most things related to football.

I do have some natural scepticism regarding how fully this game can be turned into numbers or rather how relevant a lot of those numbers would be.
The key obviously I think would be finding stats that actually apply and tells a true story. I've been thinking about some since reading this and it's a pretty hard thing to do I think.
Mainly because it's not how you've been used to thinking.
However since getting into poker and baseball stats slowly but surely as well, I haven't quite been able to escape the thought, so I'm just gonna throw some in there that comes to mind:

Passing statistics I think CAN be good but as they're often shown now are flawed or misguided.
When a team or certain players have a higher completíon it's almost entirely to do with either the style and game plan of the team or specific role the player is asked to play.
When Spain have a much higher completion than Sweden it really just tells you that one team plays a short passing game and the other a direct passing game. Either one can be depending on the circumstances good. It doesn't mean one team or the other is the better one or the more effective.
To be able to see that you need to look deeper which would be a cool thing to do.

Makalele is someone who will usually have a very high completion but that doesn't really make him the best passer.
It just has to do with the type of passes his role on the field dictates.
Pirlo play a similar position but a very different role meaning his completion percentage will naturally be lower but maybe he is still more effective?
Those things would actually be interesting looking into.
Maybe a differentation between different passes and their specific intent.
Sort of like tennis where you have forced and unforced errors.

Other interesting in game stats would be how goals are created or chances made.
How often is it the result of the ball being lost by the other team, with say an "unforced error pass" and depending on where it's lost, what big of a chance does it lead to, or goal even.

Risk and reward ratios when it comes to passing is something I find extremely interesting and is something that's probably deciding a lot of games.
Hugely relevant I think.

Also things like how many players a team have behind the ball when they concede a goal I would be interested in.
Then with that you can go further in many ways and really find different teams different defensive values.

Someone questioning Luca Toni had me thinking of striker effectiveness too.
My theory would be that a lot his chances a lot of other strikers don't even get or create if you will.
How much value is that? How much does it hurt that value or his team when he hasn't scored on say nine of them?
What different type of situations are different type of strikers able to create and how much value are those (different) type of situations.
That can then be compared to perhaps other positions and other players. The type of situations they can create. How do the different strenghts of players on the same team go together. How effective are they together in terms of value.

Maybe it can sort of be compared to different baseball factors.
Fly ball pitchers or ground ball pitchers in different parks and with defences behind them.
A very ground ball pitcher with a bad infield is shooting yourself in the foot.
Is that similar in anyway to if Jan Koller was the lone striker for Spain?

Well those are just thoughts thrown in there but basically what I find fascinating, but which also seems like a gigantic task, is finding values for different teams and different players.

edit: I just realised some of this to various degrees is wandering into subjective fields but I just have a feeling that in order to find some of the most relevant numbers that's something you might have to do, if that makes sense.

[ThaSaltCracka]

if it was possible to measure certain individual player it would be very interesting to see, however due to the fluidity of the game and the subjectiveness in regards to certain things, I think its really really hard to rate players. At least statistically rate them.

One thing that I have seen in the CL games is stats showing distance ran and passing completion %, or total shots/shots on target.

However, even those stats are misleading because someone could run a lot and not do much. Someone could pass along the back line a bunch (such as a DM) and really not do much. Someone could shoot a ton of speculative shots outside the box, or someone's shoots on target could be right at the keeper. Its so freaking hard to asses what these guys are really doing unless you watch every game and track every play, and assign every play into clearly defined groups of effectiveness and ineffectiveness.

I think it would be a good idea to first assess teams and identify the trends of each team, offensively/defensively, by halfs, by certain scores, etc. Then look at what effectiveness of those trends, etc.

I guess I am rambling. You guys get the idea.

[borovoselo]

brief thoughts;

1-extremely non-linear/fluid game when compared to any other sport
2-scoring is very low, 1-4 goals per game, that are as often products of unusual circumstances as they are not

[ThaSaltCracka]

I actually think its comparable to basketball in game flow, and probably even closer to hockey.

I think some sort of statistical analysis is possible, we just need to know where to start.

[borovoselo]

i couldnt disagree more with the comparison of the two sports

[vaagsel]

Steve McClaren new coach of FC Twente, who effectively ended 2nd in the Dutch league last year and are therefore playing CL prelims. Is he any good?