Stats and stuff

Where fellow sufferers gather to share the pain, longing and unrequited transfer requests that make being a Wanderer what it is...

Moderator: Zulus Thousand of em

Post Reply
User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Stats and stuff

Post by Dave Sutton's barnet » Thu Sep 28, 2017 1:12 pm

TBH I don't know if this will work as a topic, as it may not be everyone's cup of tea, but I thought it might be useful to have a thread for statistical stuff.

Specifically, on the Bristol City thread we were discussing Gary Madine's attacking effectiveness, or lack thereof, according to your taste. Starts about here (although there's a Madine heatmap further up), so before it disappears into the Match Threads subforum I thought I'd link it onto a new thread...

In that discussion, PT makes the point that shots attempted may not be a decent metric if your shots are shit bobblers from miles away. That's fair enough, and it's kind of why there's now a lot of talk about Expected Goals (xG), including on Match of the Day and suchlike. I'm not an enormous fan of it because it introduces that judgement but it's basically giving a rating to each chance for how likely you are to score from it. (For a fuller description with examples, see this BBC explainer.)

I mention this because on his ever-interesting Experimental361 blog, Ben Mayhew has done the maths on all the Championship teams individually.
https://experimental361.com/2017/09/28/ ... -sep-2017/

Here's ours. (It's somehow darkly amusing that we've had more different scorers in the League Cup – not shown here – than in the League.)

2017-09-27-bolton.png
2017-09-27-bolton.png (26.6 KiB) Viewed 5366 times


If strikers are in the top white bit, they're in the top 10% for finishing efficiency; if they're in the bottom white bit, they're in the bottom 10%. Now, it's a (laughably) small sample but it's perhaps notable that Gaz is closer to adequacy than ALF or Arma. It's also notable that he should be scoring about 0.28 goals per game; if you remember, he's having 2.9 efforts per game, so he should only be scoring about one in 10 of those. In other words, he's not missing easy chances, and while PT might be right that some of those attempts are low-probability, I'm not sure that's a stick to beat him with – unless he's taking a ridiculous potshot and ignoring players in better positions, which I think we all know isn't happening. Seems to me the bloke's doing his best and trying to make something happen. Hopefully the reversion to 4-2-3-1 and the return of Sammy and then Vela will give him supporting options.

For the record, and while we're at it, Armstrong has had 11 attempts in 381 minutes, or one every 34.6 minutes. Madine's had 29 in 866 mins (a shot every 29.8 min), ALF 10 in 347 (a shot every 34.7 minutes). Basically they're all having one effort every half-hour; shame none of them are going in. It's worth having a flick through other team's graphs on the link above; Ipswich's is ridiculous.

Or you could peruse this updated scatter graph: https://experimental361.com/2017/09/28/ ... -sep-2017/

Screen Shot 2017-09-28 at 13.10.26.png
Screen Shot 2017-09-28 at 13.10.26.png (260.74 KiB) Viewed 5366 times


The *state* of Burton and Ipswich...

User avatar
Lost Leopard Spot
Immortal
Immortal
Posts: 18436
Joined: Wed May 09, 2012 11:14 am
Location: In the long grass, hunting for a watering hole.

Re: Stats and stuff

Post by Lost Leopard Spot » Thu Sep 28, 2017 4:33 pm

The problem with stats is that, fascinating as they are and undoubtedly 'correct', they are simply meaningless. So what, if the average male dies at 79.8 in the UK, does it help when you die at 43.
I have a mate who works on statistical probability at Manchester University. He tells me that driving a white car at night is statistically less likely to end in a collision than driving a blue one. He's no doubt right but I've been driving non-white cars for more than four decades and not had a collision at night yet. As I say, correct but realistically meaningless.
That's not a leopard!
頑張ってください

User avatar
TonyDomingos
Passionate
Passionate
Posts: 2751
Joined: Fri Jul 13, 2007 9:27 pm
Location: Sarf East London

Re: Stats and stuff

Post by TonyDomingos » Thu Sep 28, 2017 4:54 pm

Well, I for one applaud DSB for starting up this thread! :D

I take your point LLS that observing averages, rates and frequencies doesn’t always translate into what actually happens. But, for example, if you look at our “goals for” over recent games, it suggests strongly that our next result will be “- Bolton 0”. Let’s review again at 5pm on Saturday. :P
Às armas, às armas!
Sobre a terra, sobre o mar,
Às armas, às armas!
Pela Pátria lutar!
Contra os canhões marchar, marchar!

User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Re: Stats and stuff

Post by Dave Sutton's barnet » Thu Sep 28, 2017 5:23 pm

Like I say, won't be everyone's cup of tea. That's fine. If I see a thread, or board, that doesn't interest me, I don't go on it...

User avatar
Lost Leopard Spot
Immortal
Immortal
Posts: 18436
Joined: Wed May 09, 2012 11:14 am
Location: In the long grass, hunting for a watering hole.

Re: Stats and stuff

Post by Lost Leopard Spot » Thu Sep 28, 2017 5:28 pm

Dave Sutton's barnet wrote:
Thu Sep 28, 2017 5:23 pm
Like I say, won't be everyone's cup of tea. That's fine. If I see a thread, or board, that doesn't interest me, I don't go on it...
Did I say it didn't interest me? Just pointed out an opinion thank you. Why does everyone have such thin skin these days? To counter in the same vein as your answer, why bother starting up a thread if you don't want replies to it?
That's not a leopard!
頑張ってください

nelson66
Promising
Promising
Posts: 489
Joined: Sun Aug 12, 2012 9:44 pm

Re: Stats and stuff

Post by nelson66 » Thu Sep 28, 2017 6:50 pm

i think its a great thread - I see Gary Madine as being a major reason why we are not scoring goals - DSBs statistics back up my opinion
The Whites Are Going Up 2021 :pissed: :grin:

bristol_Wanderer3
Dedicated
Dedicated
Posts: 1713
Joined: Tue Apr 10, 2007 8:53 pm

Re: Stats and stuff

Post by bristol_Wanderer3 » Thu Sep 28, 2017 8:28 pm

Great thread. Aren't our league scorers this season Morais, ALF, Madine? The first graphic replaces Morais with Armstrong?

User avatar
Harry Genshaw
Legend
Legend
Posts: 9097
Joined: Sun Nov 13, 2005 10:47 pm
Location: Half dead in Panama

Re: Stats and stuff

Post by Harry Genshaw » Thu Sep 28, 2017 9:17 pm

It all demonstrates to me that we're fecked and in the unlikely event Carlton Cole pitches up, things ain't getting any better! It doesn't seem long ago I was appalled that we were linked with Cameron Jerome - I'd give him a piggy back to the macron now
"Get your feet off the furniture you Oxbridge tw*t. You're not on a feckin punt now you know"

User avatar
TonyDomingos
Passionate
Passionate
Posts: 2751
Joined: Fri Jul 13, 2007 9:27 pm
Location: Sarf East London

Re: Stats and stuff

Post by TonyDomingos » Thu Sep 28, 2017 9:19 pm

I need to be careful on here. The last set of stats I was involved in earned a rebuke from the government’s UK Statistics Authority yesterday!
Às armas, às armas!
Sobre a terra, sobre o mar,
Às armas, às armas!
Pela Pátria lutar!
Contra os canhões marchar, marchar!

User avatar
Hoboh
Immortal
Immortal
Posts: 13303
Joined: Wed May 31, 2006 8:19 am

Re: Stats and stuff

Post by Hoboh » Thu Sep 28, 2017 10:30 pm

Needs renaming this,

The Worthy thread!

User avatar
Abdoulaye's Twin
Legend
Legend
Posts: 9167
Joined: Tue Apr 17, 2007 1:27 pm
Location: Skye high

Re: Stats and stuff

Post by Abdoulaye's Twin » Fri Sep 29, 2017 5:19 am

Yes, but we're not worthy :)

User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Re: Stats and stuff

Post by Dave Sutton's barnet » Fri Sep 29, 2017 10:21 am

Lost Leopard Spot wrote:
Thu Sep 28, 2017 5:28 pm
Dave Sutton's barnet wrote:
Thu Sep 28, 2017 5:23 pm
Like I say, won't be everyone's cup of tea. That's fine. If I see a thread, or board, that doesn't interest me, I don't go on it...
Did I say it didn't interest me? Just pointed out an opinion thank you. Why does everyone have such thin skin these days? To counter in the same vein as your answer, why bother starting up a thread if you don't want replies to it?
Seems quite a thin-skinned response for a leopard… :-D

I apologise if I’ve rattled you, especially if it’s by misunderstanding you. So you’re not uninterested (or even disinterested) in statistics, you just think they are “meaningless”. Seems to me that’s all in the usage of them, some of which is good, some of which is bad. Yes, the average human has one testicle and one ovary, but that’s reductio ad absurdum. But I find it genuinely fascinating that Brentford (second-bottom before they had the good fortune to play us) were averaging almost double the number of attempts on goal that had shot Ipswich into the play-off zone.

To say stats “are simply meaningless” is a surprising statement from someone I’ve had down as a thoughtful poster, for two reasons: it’s such a sweeping generalisation and it’s fairly obviously nonsense. There are many good, interesting and even vital uses of statistics. For a start, it keeps your mate busy at Manchester University. More importantly, without statistical analysis John Snow wouldn’t have identified the source of a Soho cholera outbreak and thus helped to found modern epidemiology, which has saved millions of lives. Meaningless? Really?

Down at the much less important level of kickball, it depends which datasets are studied. I doubt there’s much point analysing whether goalscorers are more likely to have surnames starting A-M than N-Z, but there’s probably some useful information in whether the presence or absence of certain players increases or reduces your likelihood of winning.

And that is, in the end, one of the main reasons why I post things on here: trying to work out why we’re winning, or not. To quote the Beta Band, “I don’t have too many answers, but I got a whole heap of questions I won’t hide from you.” As usual, I postulate these things to start or maintain the conversation, so of course I want replies – it’s a forum. But I’ll be completely honest, I find it tiring to read a response which is effectively “this is all shit” – it seems to me an attempt to close down the topic. Call me thin-skinned if you like – I've been called much worse, and these things depend on how much one cares – but it depends whether you want conversation or confrontation. If this place becomes more about the latter, it’s simply not worth the effort, and the loudest voices will be braying in an empty room.

User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Re: Stats and stuff

Post by Dave Sutton's barnet » Fri Sep 29, 2017 10:26 am

bristol_Wanderer3 wrote:
Thu Sep 28, 2017 8:28 pm
Great thread. Aren't our league scorers this season Morais, ALF, Madine? The first graphic replaces Morais with Armstrong?
You're right but these graphs don't list scorers, they list "a club’s main attacking players: those who have:
• Featured for at least a third of their total pitch minutes in the league this season, and
• Taken an average of at least one shot per game."

Also: "Both axes exclude penalties, as those can massively skew a player’s contribution away from the threat they pose from open play" – which is why both Arma and ALF (who scored from the spot v Leeds, nine league games and two team goals ago) are averaging precisely zero goals per 90 minutes (Y axis).

User avatar
Lost Leopard Spot
Immortal
Immortal
Posts: 18436
Joined: Wed May 09, 2012 11:14 am
Location: In the long grass, hunting for a watering hole.

Re: Stats and stuff

Post by Lost Leopard Spot » Fri Sep 29, 2017 1:31 pm

Dave Sutton's barnet wrote:
Fri Sep 29, 2017 10:21 am
Lost Leopard Spot wrote:
Thu Sep 28, 2017 5:28 pm
Dave Sutton's barnet wrote:
Thu Sep 28, 2017 5:23 pm
Like I say, won't be everyone's cup of tea. That's fine. If I see a thread, or board, that doesn't interest me, I don't go on it...
Did I say it didn't interest me? Just pointed out an opinion thank you. Why does everyone have such thin skin these days? To counter in the same vein as your answer, why bother starting up a thread if you don't want replies to it?
Seems quite a thin-skinned response for a leopard… :-D

I apologise if I’ve rattled you, especially if it’s by misunderstanding you. So you’re not uninterested (or even disinterested) in statistics, you just think they are “meaningless”. Seems to me that’s all in the usage of them, some of which is good, some of which is bad. Yes, the average human has one testicle and one ovary, but that’s reductio ad absurdum. But I find it genuinely fascinating that Brentford (second-bottom before they had the good fortune to play us) were averaging almost double the number of attempts on goal that had shot Ipswich into the play-off zone.

To say stats “are simply meaningless” is a surprising statement from someone I’ve had down as a thoughtful poster, for two reasons: it’s such a sweeping generalisation and it’s fairly obviously nonsense. There are many good, interesting and even vital uses of statistics. For a start, it keeps your mate busy at Manchester University. More importantly, without statistical analysis John Snow wouldn’t have identified the source of a Soho cholera outbreak and thus helped to found modern epidemiology, which has saved millions of lives. Meaningless? Really?

Down at the much less important level of kickball, it depends which datasets are studied. I doubt there’s much point analysing whether goalscorers are more likely to have surnames starting A-M than N-Z, but there’s probably some useful information in whether the presence or absence of certain players increases or reduces your likelihood of winning.

And that is, in the end, one of the main reasons why I post things on here: trying to work out why we’re winning, or not. To quote the Beta Band, “I don’t have too many answers, but I got a whole heap of questions I won’t hide from you.” As usual, I postulate these things to start or maintain the conversation, so of course I want replies – it’s a forum. But I’ll be completely honest, I find it tiring to read a response which is effectively “this is all shit” – it seems to me an attempt to close down the topic. Call me thin-skinned if you like – I've been called much worse, and these things depend on how much one cares – but it depends whether you want conversation or confrontation. If this place becomes more about the latter, it’s simply not worth the effort, and the loudest voices will be braying in an empty room.
Apology accepted. I still stick with my original contention that statistics are meaningless. You use the example of John Snow and the Broad Street water pump: that was a triumph of forensic epidemiology, aided by a use of statistics. The statistics themselves were meaningless until mapped. The mapping itself was meaningless until its relationship to the distribution of water pumps was also mapped onto the same map. And even then that combined map was only meaningful if one had already worked out a potential relationship between water pumps and death by cholera. John Snow didn't work out the source of cholera using statistics, he worked that out by dismissing the miasma theory and postulating water as the source of cholera. It's a valuable lesson that statistics are only viable if you apply them using the correct theory. Quite often (in fact more often than not) statistics are utilised in a raw fashion without reference to any theory, or are used to support a theory that has no real evidence whatsoever.
That's what I mean by "statistics are meaningless". You need a theory backed up by evidence before applying statistics to support that theory, only then can stats impart their own backing to a theory. Unfortunately though stats are wielded as the be all and end all of an argument without any other evidence of any kind, and lead into false presumptions.
I would summarise by saying that pure maths gives proof by its own standards, science gives proof that is conditional, and statistics provides a handle that can either enable a mathematic proof or point towards a scientific theory, but used alone statistics are meaningless.
I hope this qualifies my opening remarks such that you understand my position.
As for the thread itself, I welcome it. I'd just prefer that when the stats are produced they are used to back up a theory and not for the mere meaningless production of stats as an exercise in itself - not that I'm accusing you of doing that. :D
That's not a leopard!
頑張ってください

User avatar
BWFC_Insane
Immortal
Immortal
Posts: 36010
Joined: Mon Jul 09, 2007 4:07 pm

Re: Stats and stuff

Post by BWFC_Insane » Fri Sep 29, 2017 1:51 pm

Lost Leopard Spot wrote:
Fri Sep 29, 2017 1:31 pm
Dave Sutton's barnet wrote:
Fri Sep 29, 2017 10:21 am
Lost Leopard Spot wrote:
Thu Sep 28, 2017 5:28 pm
Dave Sutton's barnet wrote:
Thu Sep 28, 2017 5:23 pm
Like I say, won't be everyone's cup of tea. That's fine. If I see a thread, or board, that doesn't interest me, I don't go on it...
Did I say it didn't interest me? Just pointed out an opinion thank you. Why does everyone have such thin skin these days? To counter in the same vein as your answer, why bother starting up a thread if you don't want replies to it?
Seems quite a thin-skinned response for a leopard… :-D

I apologise if I’ve rattled you, especially if it’s by misunderstanding you. So you’re not uninterested (or even disinterested) in statistics, you just think they are “meaningless”. Seems to me that’s all in the usage of them, some of which is good, some of which is bad. Yes, the average human has one testicle and one ovary, but that’s reductio ad absurdum. But I find it genuinely fascinating that Brentford (second-bottom before they had the good fortune to play us) were averaging almost double the number of attempts on goal that had shot Ipswich into the play-off zone.

To say stats “are simply meaningless” is a surprising statement from someone I’ve had down as a thoughtful poster, for two reasons: it’s such a sweeping generalisation and it’s fairly obviously nonsense. There are many good, interesting and even vital uses of statistics. For a start, it keeps your mate busy at Manchester University. More importantly, without statistical analysis John Snow wouldn’t have identified the source of a Soho cholera outbreak and thus helped to found modern epidemiology, which has saved millions of lives. Meaningless? Really?

Down at the much less important level of kickball, it depends which datasets are studied. I doubt there’s much point analysing whether goalscorers are more likely to have surnames starting A-M than N-Z, but there’s probably some useful information in whether the presence or absence of certain players increases or reduces your likelihood of winning.

And that is, in the end, one of the main reasons why I post things on here: trying to work out why we’re winning, or not. To quote the Beta Band, “I don’t have too many answers, but I got a whole heap of questions I won’t hide from you.” As usual, I postulate these things to start or maintain the conversation, so of course I want replies – it’s a forum. But I’ll be completely honest, I find it tiring to read a response which is effectively “this is all shit” – it seems to me an attempt to close down the topic. Call me thin-skinned if you like – I've been called much worse, and these things depend on how much one cares – but it depends whether you want conversation or confrontation. If this place becomes more about the latter, it’s simply not worth the effort, and the loudest voices will be braying in an empty room.
Apology accepted. I still stick with my original contention that statistics are meaningless. You use the example of John Snow and the Broad Street water pump: that was a triumph of forensic epidemiology, aided by a use of statistics. The statistics themselves were meaningless until mapped. The mapping itself was meaningless until its relationship to the distribution of water pumps was also mapped onto the same map. And even then that combined map was only meaningful if one had already worked out a potential relationship between water pumps and death by cholera. John Snow didn't work out the source of cholera using statistics, he worked that out by dismissing the miasma theory and postulating water as the source of cholera. It's a valuable lesson that statistics are only viable if you apply them using the correct theory. Quite often (in fact more often than not) statistics are utilised in a raw fashion without reference to any theory, or are used to support a theory that has no real evidence whatsoever.
That's what I mean by "statistics are meaningless". You need a theory backed up by evidence before applying statistics to support that theory, only then can stats impart their own backing to a theory. Unfortunately though stats are wielded as the be all and end all of an argument without any other evidence of any kind, and lead into false presumptions.
I would summarise by saying that pure maths gives proof by its own standards, science gives proof that is conditional, and statistics provides a handle that can either enable a mathematic proof or point towards a scientific theory, but used alone statistics are meaningless.
I hope this qualifies my opening remarks such that you understand my position.
As for the thread itself, I welcome it. I'd just prefer that when the stats are produced they are used to back up a theory and not for the mere meaningless production of stats as an exercise in itself - not that I'm accusing you of doing that. :D
Statistics (or data as really we're discussing) can also be used to develop theories and ideas though - in that they can reveal patterns or correlations that previously were unknown.

I broadly agree that looking at statistics can simply become a form of confirmation bias. Bolton don't score enough...we don't have many shots...can mean whatever you want it to. And those that think the issue is the manager will use a stat to back that up. But there is no real context or meaning to goals scored and shots taken. So is without some extra context a fairly meaningless statistic.

However, there was a stat early doors this season that said we'd made fewer tackles than any other team in the league. At the time honestly, that was a slight surprise (someone will be on to say it was obvious blah, blah, blah) - but it possibly pointed to the relative weakness of our midfield. Certainly tackles made are quite a good indicator for how competitive you are especially as a struggling team near the bottom. On its own it doesn't entirely tell a story but I think is a useful indicator of where our problems were...or some of them at least.

User avatar
Lost Leopard Spot
Immortal
Immortal
Posts: 18436
Joined: Wed May 09, 2012 11:14 am
Location: In the long grass, hunting for a watering hole.

Re: Stats and stuff

Post by Lost Leopard Spot » Fri Sep 29, 2017 2:01 pm

^
Yes, stats can be used to develop theories and ideas. However, to use the example you quoted, you already needed a theory (that we were losing games because our midfield was weak) in order to give meaning to the stat. The very same stat if applied in a different context, say for example Man City have made fewer tackles than any other team in their division, can be construed to back the proposition that Man City were dominant because their midfield kept possession and didn't need to make tackles... See what I mean, stats are only meaningful when used to back up a theory. Stats are correct but meaningless. Theories are meaningful but can be wrong or (provisionally) right.
That's not a leopard!
頑張ってください

User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Re: Stats and stuff

Post by Dave Sutton's barnet » Fri Sep 29, 2017 2:14 pm

LLS: Cool, let's crack on.
BWFC_Insane wrote:
Fri Sep 29, 2017 1:51 pm
However, there was a stat early doors this season that said we'd made fewer tackles than any other team in the league. At the time honestly, that was a slight surprise (someone will be on to say it was obvious blah, blah, blah) - but it possibly pointed to the relative weakness of our midfield. Certainly tackles made are quite a good indicator for how competitive you are especially as a struggling team near the bottom. On its own it doesn't entirely tell a story but I think is a useful indicator of where our problems were...or some of them at least.
Yeah, I remember that stat/thought. For the record, here are the current stats on that (per WhoScored). We're pretty much bang-on mid-table - 11th fewest, 3.6 more per game than the tackle-shy Brummies, 2.6 less than the in-yer-face Dogheads.
Screen Shot 2017-09-29 at 13.56.56.png
Screen Shot 2017-09-29 at 13.56.56.png (119.72 KiB) Viewed 5109 times
That said, there seems so little correlation between tackles and league position that it would be a large stretch to consider it causation. Here's that list again with those teams' league position in the final column (instead of WhoScored's hoky algorithmic "score"):
Screen Shot 2.png
Screen Shot 2.png (126.27 KiB) Viewed 5109 times
As you can see, we've got the same average tackle rate as the Blades, who are second; leaders Cardiff are only fractionally more than 21st-placed Barnsley. That doesn't mean it's not part of the problem. As a related aside, only three teams (Preston, Millwall and QPR) have given away more fouls per game than our 13.9; maybe that suggests we're just that fraction too slow, but to assess that I'd have to do a deeper dive than I have time to do.

User avatar
BWFC_Insane
Immortal
Immortal
Posts: 36010
Joined: Mon Jul 09, 2007 4:07 pm

Re: Stats and stuff

Post by BWFC_Insane » Fri Sep 29, 2017 2:15 pm

Lost Leopard Spot wrote:
Fri Sep 29, 2017 2:01 pm
^
Yes, stats can be used to develop theories and ideas. However, to use the example you quoted, you already needed a theory (that we were losing games because our midfield was weak) in order to give meaning to the stat. The very same stat if applied in a different context, say for example Man City have made fewer tackles than any other team in their division, can be construed to back the proposition that Man City were dominant because their midfield kept possession and didn't need to make tackles... See what I mean, stats are only meaningful when used to back up a theory. Stats are correct but meaningless. Theories are meaningful but can be wrong or (provisionally) right.
Indeed. But you could take that stat without the theory and investigate it, interrogate it and develop a theory. Sure you can't do that with the one stat in isolation. But sometimes without that stat you'd never have investigated or interrogated what was happening.

User avatar
Dave Sutton's barnet
Immortal
Immortal
Posts: 28435
Joined: Sun May 14, 2006 4:00 pm
Location: Hanging on in quiet desperation
Contact:

Re: Stats and stuff

Post by Dave Sutton's barnet » Fri Sep 29, 2017 2:38 pm

Lost Leopard Spot wrote:
Fri Sep 29, 2017 2:01 pm
say for example Man City have made fewer tackles than any other team in their division, can be construed to back the proposition that Man City were dominant because their midfield kept possession and didn't need to make tackles...
Indeed they have (12.3pg). Then it's West Brom (12.7), Burnley (12.8 ) and Man U (13), which backs up LLS's point while also suggesting that things can happen for a number of reasons.

User avatar
Lost Leopard Spot
Immortal
Immortal
Posts: 18436
Joined: Wed May 09, 2012 11:14 am
Location: In the long grass, hunting for a watering hole.

Re: Stats and stuff

Post by Lost Leopard Spot » Fri Sep 29, 2017 2:44 pm

BWFC_Insane wrote:
Fri Sep 29, 2017 2:15 pm
Lost Leopard Spot wrote:
Fri Sep 29, 2017 2:01 pm
^
Yes, stats can be used to develop theories and ideas. However, to use the example you quoted, you already needed a theory (that we were losing games because our midfield was weak) in order to give meaning to the stat. The very same stat if applied in a different context, say for example Man City have made fewer tackles than any other team in their division, can be construed to back the proposition that Man City were dominant because their midfield kept possession and didn't need to make tackles... See what I mean, stats are only meaningful when used to back up a theory. Stats are correct but meaningless. Theories are meaningful but can be wrong or (provisionally) right.
Indeed. But you could take that stat without the theory and investigate it, interrogate it and develop a theory. Sure you can't do that with the one stat in isolation. But sometimes without that stat you'd never have investigated or interrogated what was happening.
I agree entirely. Stats are a fact. They can lead to theory. My contention is simple: theory is meaningful, statistics aren't. I think maybe we aren't disagreeing, just not agreeing on what is meaningful: Factuality does not equal meaningfulness!
That's not a leopard!
頑張ってください

Who is online

Users browsing this forum: Baidu [Spider] and 73 guests