Stats. Everyone is obsessed with them. Why?

[Hibsclaret]

We win with very poor passing stats. Blah blah blah.

Everyone is obsessed with stats these days and tbh I think a lot of it is bs.

The only stat that matters is Cardiff 1 Burnley 2.

There’s all this expected goals bs and quality of actual chances blah blah blah.

It’s all nonsense tbh.

[Devils_Advocate]

Totally agree now tell me again what the stats say about how we perform with Ward in the team compared to Taylor

[MRG]

I agree but the issue is that the stats can tell a bit more about the story. Most other teams in the division would of given us a hammering with 50 touches in the penalty area. The stats highlight that we need to fix this regardless of the score

[CrosspoolClarets]

A fair enough opinion Hibs but I make a good living by advising organisations on stats related issues and I firmly disagree.

The shots conceded today (a stat) was more than we conceded at either Arsenal or City last season, yet expected goals (another stat) was only 1.49 (round down to 1, the same as actual goals conceded). Ours was only 0.11.

Analysing just those two stats tells me, even if I hadn’t watched the game, that we restricted them quite well and scored with what were not even half chances. Sometimes when we watch the game we get so caught up in it we don’t see the truth, and the stats are useful when thinking about it afterwards. I thought we were hammered, but on reflection, maybe not.

If we had the full Opta range (sprints etc) that clubs pay for, we would have even more info on who has done well.

I’m assuming you aren’t a fan of Fantasy Football or Moneyball?

[Hibsclaret]

I agree but the issue is that the stats can tell a bit more about the story. Most other teams in the division would of given us a hammering with 50 touches in the penalty area. The stats highlight that we need to fix this regardless of the score

How many times has that happened with Ward playing?

[Hibsclaret]

Totally agree now tell me again what the stats say about how we perform with Ward in the team compared to Taylor

No idea. Don’t agree with stats

[MRG]

How many times has that happened with Ward playing?

He was a good un in his day, there’s no doubting that!

[bfcjg]

89.5 % on here agree interestingly 9% don't 1.5% don't have an opinion.

[Hibsclaret]

A fair enough opinion Hibs but I make a good living by advising organisations on stats related issues and I firmly disagree.

The shots conceded today (a stat) was more than we conceded at either Arsenal or City last season, yet expected goals (another stat) was only 1.49 (round down to 1, the same as actual goals conceded). Ours was only 0.11.

Analysing just those two stats tells me, even if I hadn’t watched the game, that we restricted them quite well and scored with what were not even half chances. Sometimes when we watch the game we get so caught up in it we don’t see the truth, and the stats are useful when thinking about it afterwards. I thought we were hammered, but on reflection, maybe not.

If we had the full Opta range (sprints etc) that clubs pay for, we would have even more info on who has done well.

I’m assuming you aren’t a fan of Fantasy Football or Moneyball?

So the expected goals conceded was one. If Wardy was in the team does this mean the expected goals conceded would be nil as he would have closed the gap from where the goal was created?

The stats tell you nothing. Last week we scored 4 and I’m guessing the expected was 1.

[Quickenthetempo]

Sean Dyche bases his management career around them.

[willsclarets]

They can be misleading, but 56% passing accuracy summed up our first half performance pretty neatly.

[SGr]

They can be used to back up arguments, but not form the basis of them. There is no substitute for actually watching the game.

[SGr]

They can be misleading, but 56% passing accuracy summed up our first half performance pretty neatly.

In fairness, on that first half I think 56% was flattering

[Imploding Turtle]

If you weren't interested in statistics then you wouldn't be interested in results.

[FactualFrank]

Statistics help peform a step. An algorithm is a set of steps. Stats are just maths lumped together and you can't forecast anything without those stats. If it was nonsense, then bookies wouldn't exist.

[watsonsclarets]

They can be used to back up arguments, but not form the basis of them. There is no substitute for actually watching the game.

Watching a game you mean watching the ball because watching a game can mean so much different to different people.

They form the basis of how people play and who they play against and how watching the game a bit like many on this forum to many unconscious bias issues take place.

My main issue this season is the gap between the lines which is currently a lot larger than it used to be and we are giving the ball away 10 yards closer to our goal than we did last year and teams are not closing us down as much as they did.

Patterns of Play and why goals are scored and conceded are really fun topics but the eye can only tell you what your watching and watching the ball does not tell you everything.

Good book is Moneyball.

[RocketLawnChair]

Today’s Goal Stats- Cardiff City 33.3333 % Burnley 66.6666 %

No, call me old fashioned but Cardiff City 1 Burnley 2 sounds better.

[BennyD]

There's only 1 stat that matters.

[dsr]

A fair enough opinion Hibs but I make a good living by advising organisations on stats related issues and I firmly disagree.

The shots conceded today (a stat) was more than we conceded at either Arsenal or City last season, yet expected goals (another stat) was only 1.49 (round down to 1, the same as actual goals conceded). Ours was only 0.11.

Analysing just those two stats tells me, even if I hadn’t watched the game, that we restricted them quite well and scored with what were not even half chances. Sometimes when we watch the game we get so caught up in it we don’t see the truth, and the stats are useful when thinking about it afterwards. I thought we were hammered, but on reflection, maybe not.

If we had the full Opta range (sprints etc) that clubs pay for, we would have even more info on who has done well.

I’m assuming you aren’t a fan of Fantasy Football or Moneyball?

How was it only 0.11? At what point in the move is it calculated? Because when Vokes took his header, he certainly had a better than 1 in 10 chance of scoring. I suspect when Vydra was about to cross, the chances were better than 1 in 10 even then.

[tiger76]

We only had 2 efforts on target and both resulted in goals,that's either very lucky or very clinical.

[Wile E Coyote]

its american bull**** again that's why.

[Claretto]

The stats tell you nothing. Last week we scored 4 and I’m guessing the expected was 1.

What do you expect? Of course the actual score will be a better indicator of the actual score than the xG...

The xG is a useful tool, but not gospel. We are somewhat of an xG anomaly given the number of players we get behind the ball and the number of blocks we achieve.

You sound intimidated because you don't understand it.

[Longsidebogs]

Burnley had 3 shots in 94 minutes of football.....a very poor stat.
Burnley had 2 shots on target from a total of 3 shots in 94 minutes of football and scored 2 goals.....a happier stat.
But overall we are bobbins and if things don’t change soon we will be relegated.....a realistic stat

[Hibsclaret]

What do you expect? Of course the actual score will be a better indicator of the actual score than the xG...

The xG is a useful tool, but not gospel. We are somewhat of an xG anomaly given the number of players we get behind the ball and the number of blocks we achieve.

You sound intimidated because you don't understand it.

I am. I am so intimidated by stats. Tool.

How many times are xg’s actually correct?

As for stats with reference to betting, that is all designed to help us lose our hard earned.

[Hibsclaret]

How was it only 0.11? At what point in the move is it calculated? Because when Vokes took his header, he certainly had a better than 1 in 10 chance of scoring. I suspect when Vydra was about to cross, the chances were better than 1 in 10 even then.

So we have expected goals of 0.11 in an away game.

The previous away games we had scored a sequence of 0, 2, 0. The xg has to be at least 0.66. If you factor in playing against a team that conceded 3 and 5 in their previous home games, it should have increased to around the 2 we actually scored....

[thatdberight]

So we have expected goals of 0.11 in an away game.

The previous away games we had scored a sequence of 0, 2, 0. The xg has to be at least 0.66. If you factor in playing against a team that conceded 3 and 5 in their previous home games, it should have increased to around the 2 we actually scored....

You really do hate stats. So much you've not even bothered to find out what this one means. Fair enough.

[Hibsclaret]

You really do hate stats. So much you've not even bothered to find out what this one means. Fair enough.

True

[CrosspoolClarets]

How was it only 0.11? At what point in the move is it calculated? Because when Vokes took his header, he certainly had a better than 1 in 10 chance of scoring. I suspect when Vydra was about to cross, the chances were better than 1 in 10 even then.

The two goals were given (by Understat) a 4% and 5% chance of being scored, and the Vydra one into outer space a 2% chance.

Remember it is judged as the chance, Vokes getting his head on it is part of taking the chance, as the cross came over the most likely outcome was the defender would clear it or there would be no clean contact. Even afterwards the keeper would have been a favourite from most headers. If Understat were rubbish at judging this they wouldn’t be so highly regarded but I agree I thought the goals were slightly better chances than that

I’m amused at whether Hibs is fishing or not, or genuinely doesn’t understand it (which would be fair enough). Expected goals can never be exactly the same as real goals, unless in life everything that happens is exactly what would happen the average time. How boring would that be? It’s a useful measure of quality of chance creation and quality of finishing.

[Hibsclaret]

I start to understand things by reading the Einsteinesque ramblings of people on here.

0.11 sounds about right then. It is all damn lies and statistics. It shows we have the most clinical forward line in the league then. Wouldn’t swap them for the Man City players as they are way more wasteful

[SalisburyClaret]

The stats are primarily used to fill air time. It provides the pundits with the basis for a discussion. The idea was originally taken from how American Football an Baseball are televised- nothing happens for long periods and you cant just fill the time with ads

Having said that we've scored 6 goals from our last 7 attempts so there is no excuse for anything other than all out attack versus Huddersfield

[StuffyClaret]

This might help: Statistically speaking, the average human being has less than two legs

[houseboy]

We win with very poor passing stats. Blah blah blah.

Everyone is obsessed with stats these days and tbh I think a lot of it is bs.

The only stat that matters is Cardiff 1 Burnley 2.

There’s all this expected goals bs and quality of actual chances blah blah blah.

It’s all nonsense tbh.

Totally agree. There was a time when you went to a game and only the score mattered - you might argue the toss about whether the win or loss was 'deserved' but in the end that was that. Stats have been imported from America (American football seems to exist only for statisticians) in order to give the 'experts' on TV something to talk about. I think they are used now to give managers and some fans a way of explaining how they should have won a game they've just lost 5-0.
Statistics can 'prove' anything you want as long as you present them in the right way.

[houseboy]

To quote that great American one-line comedian Steve Wright - '87.5% of statistics are made up'.

And: '99% of lawyers make the other 1% look bad'.

[Hipper]

This might help: Statistically speaking, the average human being has less than two legs

That's where the term 'mean' comes in.

[thatdberight]

That's where the term 'mean' comes in.

The mean is less than 2.

[Hipper]

Totally agree. There was a time when you went to a game and only the score mattered...

Surely that's never been true. You look how individual players have played, how the team plays etc. etc..

Stats can add insight into each performance. As with all stats though you have to understand how to use them. If you don't want to or can't then they are useless.

[Lancasterclaret]

Every podcast on football I listen too says exactly the same thing, according to the stats on stuff like shots, Burnley should be losing more games than we do.

We don't because we are also high on blocks, saves, running, covering etc etc etc

They are useful, but like all stats only up to a point.

There can't be a stat for the "strong jaw" that we have for example.

[Hipper]

The mean is less than 2.

Have the meanings changed or have I forgotten this stuff?

What is it called when you use the most common number - two in the case of legs on humans?

[houseboy]

Surely that's never been true. You look how individual players have played, how the team plays etc. etc..

Stats can add insight into each performance. As with all stats though you have to understand how to use them. If you don't want to or can't then they are useless.

Two points:
First I DID say that you would argue over the pros and cons of the game (who had the most corners and whether you should have had a penalty, all the usual stuff) but we didn't go into percetages and become obsessed with it all, that is the difference.
Second: As someone who, unfortunately, has had to work with stats quite a lot in my work I can say that many, if not most, are open to interpretation, in the sense, as I stated, that they can prove almost anything depending on how they are presented. An example being if you have, for instance, a defensive unit that has a record of the most blocks in a season. Is that a 'good' stat or a 'bad one? It's good from a defensive point of view but in the overall scheme of things it's bad because it probably means you spent most of the season being outplayed.
Stats are all okay and can be a useful tool in many areas but I think in football they just give more airtime to pundits who otherwise would have very little of interest to say.
Having said that if you are interested in that side of things that's okay, I just find them a bit dull and uninteresting.

One of my favourite quotes: There are three kinds of lies: lies, damn lies and statistics. (Attributed to Benjamin Disraeli but oft quoted by others, including Mark Twain).

[aggi]

I did a bit of work with a company that was providing stats to a betting company and the level of stats that they provide (and that are also provided to clubs, etc) are way beyond what we see.

Looking at passing for instance, as well as the basic pass completion there'll also be stats for length of pass, direction, if it resulted in a shot further down the line, what other passes were available, etc

[deanothedino]

Stats are all okay and can be a useful tool in many areas but I think in football they just give more airtime to pundits who otherwise would have very little of interest to say.

I think the main of what you are saying here is the problem. The stats in football are useful tools, to the right people. For the mainstream watching at home they are irrelevant. Yes, some people are interested (including myself), but to most people it's just noise.

For managers and coaches stats are a no brainer to me, provided you are using the right ones.

There were some interesting stats about us defensively last season around how we limited the quality of the oppositions chances, but an observant, unbiased, viewer of the game could have told you that as they'd see with their own eyes that we didn't give away many clear cut chances.

[Hibsclaret]

I did a bit of work with a company that was providing stats to a betting company and the level of stats that they provide (and that are also provided to clubs, etc) are way beyond what we see.

Looking at passing for instance, as well as the basic pass completion there'll also be stats for length of pass, direction, if it resulted in a shot further down the line, what other passes were available, etc

So in a nutshell stats are for bookies to take more money off us

[Tall Paul]

So in a nutshell stats are for bookies to take more money off us

How can they use them to make money if they're nonsense and don't matter?

[Hibsclaret]

They are nonsense to the punter but matter to the bookies. It helps them price the games clearly.

[houseboy]

I think the main of what you are saying here is the problem. The stats in football are useful tools, to the right people. For the mainstream watching at home they are irrelevant. Yes, some people are interested (including myself), but to most people it's just noise.

For managers and coaches stats are a no brainer to me, provided you are using the right ones.

There were some interesting stats about us defensively last season around how we limited the quality of the oppositions chances, but an observant, unbiased, viewer of the game could have told you that as they'd see with their own eyes that we didn't give away many clear cut chances.

You're probably pretty spot on there (statistically speaking ). I think, as you say, that many people find them mildly of interest but it all gets a but over the top. MOTD could be a good half hour shorter if it weren't for the endless analysis between games (time to top up the wine glass). I suppose the pundits have to justify the ridiculous amounts of money they get for mostly stating the obvious, so enter the 'stats' to pad out the programme.
Some stats are okay, I'm thinking for instance the number of consecutive passes in a good team goal (25 last season for Burnley at Everton) but I'm not the least bit interested in how many miles a player has run during a game. And then there are the classicly misleading (even though true) stats, as per the 35 year unbeaten run for Rovers against us. Looks impressive on paper until you work out it was only 11 games, 4 were draws and two were FA cup ties, leaving 9 league games, 33% of which were draws. All of a sudden it's not much to write home about.
As they say in cookery, it's all about presentation.

[Hibsclaret]

Absolutely.

If you look at goals to shot ratio in our last 2 games we are probably the best team in Europe.

If you listen to most on here we can’t pass wind...

[TVC15]

Dyche has said several times that statistics are instrumental to him and how he manages the team.
You ever heard him debate with people who question why he brings everyone back for corners ? His answer is statistical evidence.

Very simplistic to say the only statistic that matters is the score - very clearly it is not to Dyche and many clubs now. They would argue that one of the reasons for the result was the statistical analysis they use to decide how the team set up, which players are picked, the decisions they make on the pitch etc.

I find expected goals one of the better statistics - far better than shots attempted. Perfect example is the home game v Liverpool when we returned to the Premier League and won 2-0. Lots of talk after that game about the number of shots Liverpool had and the amount of possession...but a lot of their possession was useless passing sideways across the back line which we allowed them to do and vast majority of their shots were wild attempts from outside the area. Don't think expected goals were around then but if they were they would have told a much more accurate story of the game.

It would be a strong guess that a team that has a high number of expected goals but not scoring those goals would be more likely to eventually start scoring than a team with low expected goals over a number of games. Even if you do not have especially clinical strikers like ourselves or Cardiff then even a less clinical striker is still more likely to score with an easier chance than he is with a half chance or shot from outside the area.

[thatdberight]

Have the meanings changed or have I forgotten this stuff?

What is it called when you use the most common number - two in the case of legs on humans?

Mean = Total legs / total humans.
Mode = Most common number of legs for a human to have.
Median = If you listed the number of legs of every human in numerical order, the number of legs in the middle item of the list.

So; 1.9something (my estimate), 2, 2.

[joey13]

100% Colin is still a ****** and Craig Bellamy is a Dick