"Expected Goals" (xG)

[Carnero]

Proof that the "expected goals" statistic and subsequent tables are a load of old Betty Swollocks:

[RadioactiveWaste]

3 minutes ago, Carnero said:

Proof that the "expected goals" statistic and subsequent tables are a load of old Betty Swollocks:

I've no problem with statistics, it's just fancy ways of comparing some groups of numbers to others. I get insanely irritated with the interpretation of them and those who would bestow upon them heavenly qualities they do not merit.

xG is probably the worst.

 

[Carnero]

We had far more clear chances than PNE today, yet xG will I presume hand us a 1-2 defeat ?

Garbage.

[Srg]

That is really shocking to be honest. No one would have blinked an eye if we were 3 up at half time. 

[Shuff264]

I think they've counted the disallowed goal? The huge spike just before 60mins

[Tamworthram]

For me it just beggers believe that there must be people being paid and making money out of producing pointless statistics. They really should get a real job. 

[LB_DCFC]

8 minutes ago, Tamworthram said:

For me it just beggers believe that there must be people being paid and making money out of producing pointless statistics. They really should get a real job. 

Yeah I know. The recruitment team at Brentford, which is all built on statistical profiling of players are really doing a pointless and awful job. Not as if the best and most consistent recruitment team in the division have the most involvement of stats guys out of any club in the league or anything.

[DarkFruitsRam7]

Those who live and die by stats are fools. But so too are those who completely disregard them. 

[RadioactiveWaste]

17 minutes ago, LB_DCFC said:

Yeah I know. The recruitment team at Brentford, which is all built on statistical profiling of players are really doing a pointless and awful job. Not as if the best and most consistent recruitment team in the division have the most involvement of stats guys out of any club in the league or anything.

Its the interpretation of what the numbers mean that counts.

Brentford recruitment appear to make a good job of it. The xG stats are often to contrary of what people can see with their own eyes.

Its not the statistic, that's just a number. The xG statistic being a particularly pointless one.

You could calculate the mean position of the ball throughout the match (probably somewhere in the middle of the pitch) for example and say that tells you who really won the balance of play. You could generate tables and charts, you could have pundits tell you it's insightful beyond what a fan might make of the match they sww, you could give it a fancy name. But would it really mean anything?

Everyone else will look at the score and hum a bit if they thought one side on the other was a big lucky or deserved.

 

 

[EtoileSportiveDeDerby]

1-0 the only stat that matter

[SamUltraRam]

More examples of pointless 'artificial' jobs being created to replace 'real' jobs that have been lost.

It amazes me how we survive 

[JfR]

2 hours ago, Shuff264 said:

I think they've counted the disallowed goal? The huge spike just before 60mins

That's actually Ben Davies's shot immediately before the disallowed goal and he probably will be disappointed to have missed from here:

[Millenniumram]

2 hours ago, DarkFruitsRam7 said:

Those who live and die by stats are fools. But so too are those who completely disregard them. 

This. Stats are absolutely vital for giving us an idea about trends, and also as evidence to back up our own ideas about trends. Honestly the amount of the world that revolves around statistics is unbelievable really, you’d have to be an idiot to not see them as being useful. But, you can’t use statistics to overrule what you can see with your own eyes. Like today, don’t think anyone outside a Preston forum would argue that we didn’t deserve the 3 points, so on this occasion we can dismiss that stat based on what we saw. But other times, it may be very useful.

FWIW the XG stat is quite an interesting concept, in terms of being able to give a quantitative interpretation of how a match has played out overall, which is something (aside from the scoreline of course!) that we haven’t really had in the past. And, in fairness, when looked at over long periods of time it’s often fairly accurate- I’ve noticed teams scoring more than they’re expected to by XG often drop away later in a season, and vice versa. But it can obviously get skewed within a single game quite significantly as today shows, so it’s far from perfect and the calculation system of it needs changing. Until then, for a single match it’s not a great stat to look at.

[desirelines]

xG reckons 30-yard screamers shouldn't count as goals 

[Tamworthram]

11 hours ago, LB_DCFC said:

Yeah I know. The recruitment team at Brentford, which is all built on statistical profiling of players are really doing a pointless and awful job. Not as if the best and most consistent recruitment team in the division have the most involvement of stats guys out of any club in the league or anything.

Fair point. Still not a real job though. Makes you wonder though, if that's the route of their recruitment success, why others don't or can't adopt it more effectively. 

[RadioactiveWaste]

8 hours ago, desirelines said:

xG reckons 30-yard screamers shouldn't count as goals 

Apparently there's a modelling process that deduces what chance each shot has of being a goal.

However, like any maths based model, it's prone to both GIGO "garbage in, garbage out" and the fact it's just doing some sums - do those sums mean anything, do they reflect what's happening in the real world?

You'd hope that the masters of xG would refine their models based on real events, but as the xG routinely doesn't reflect what people can see, I don't pay it much attention.

What's better than xG, IMO, is have a look at what can be directly measured (i.e. Shots, shots on target, passes completed, distance covered etc) and use your own mind to decide if you think the score reflected the game, on one side was lucky.

 

[Sith Happens]

In light of recent matches, can we have an 'Expected Injury that goes unpunished' stat for the next match?

[richinspain]

 So, I've been looking at that graph and trying to work it out. The only conclusion I can come up with is that we (or Waghorn) should have scored 1.25 goals, and they should have scored 1.75 goals. So we should have lost. 
 I don't have a problem with statistics (although I do agree with Jimmy Sirrell, "statistics are like a mini skirt, very nice but they hide what is really important"), but this statistic is just complete and utter garbage. Expected goals, what the * are expected goals? Is there an expected save stat? Or an egfac (expected goalkeeper flapping at crosses) stat? An expected to win stat, as supplied by @B4ev6is or by @Curtains? (Sorry Curtains, not picking on you just a user I thought of).

 Statistics, as in tangible, countable things, are very useful. However, "expected" stats as far as I'm concerned are useless. Is the  person shooting is taken into account? Would say Rooney have the same xG as Sammon? Show me the distance covered by each player, passes made, %pass completion. Anything you like, but not expected stats. I expect us to win every game (look at my predictions), but I know that we won't.
Anyway, what exactly does xG mean?

[tomonred]

xG is a cracking tool  to be able to measure the quality of chances created and how teams are performing. It shouldn’t just be used in isolation to tell you anything.

Each attempt is allocated a value from 0 to 1 based on the likelihood of that chance being converted into a goal. The values are allocated based on 1000’s and 1000’s of previous shot data so it is a really good method for seeing if teams/players are over or under performing based on their xG stats v actual goal stats.

Angle, pressure on ball, head/foot etc is all taken into account. It just depends on the accuracy of the model being used. The best ones are by the big stats companies like Opta where you have to pay money to be able to access.

Lots of people make good money from betting/trading using these stats as a basis to enter markets when their eyes are correlating with what the stats are saying

 

[richinspain]

17 minutes ago, tomonred said:

xG is a cracking tool  to be able to measure the quality of chances created and how teams are performing. It shouldn’t just be used in isolation to tell you anything.

Each attempt is allocated a value from 0 to 1 based on the likelihood of that chance being converted into a goal. The values are allocated based on 1000’s and 1000’s of previous shot data so it is a really good method for seeing if teams/players are over or under performing based on their xG stats v actual goal stats.

Angle, pressure on ball, head/foot etc is all taken into account. It just depends on the accuracy of the model being used. The best ones are by the big stats companies like Opta where you have to pay money to be able to access.

Lots of people make good money from betting/trading using these stats as a basis to enter markets when their eyes are correlating with what the stats are saying

 

So thousands of examples are used to say whether a goal "should" have been scored? I suppose that statistically it has got something behind it, but practically I think that our eyes are much better judges, and from what I saw yesterday we were by far the better team and with a bit more luck/composure could have won by more. Sorry, but for me these are stats for stat's sake.