Opta and statistics

SeantheMan

I'm sure most people have seen the statistics for games from Opta and wondered how exactly they are collected etc etc

Video on "numberphile" shows the amount of data they collect , what they use the data for, and their next video will show 'how' they collect the data.

Thought this may interest some of you

https://www.youtube.com/watch?v=YJuHC7xXsGA

BhoscaCapall

Always wondered how they collect it Do they have loads of people watching the games each person to check for each thing (passes etc)? I presume they go to the actual games live. And they must have multiple people for each role because there's bound to be errors.

Guess I'll find out.

SeantheMan

They finally released Part 2 ! Enjoy

How they actually collect the data

https://www.youtube.com/watch?v=GyN-qpVfOWA

upandcumming

Great vids, would love to know how they do the running distance over a game!

roanoke

upandcumming wrote: »

Great vids, would love to know how they do the running distance over a game!

I heard it was calculated based on stride count (rather than measurement of actual distance run) but I'm readily prepared to be corrected on that.

eagle eye

Their stats are awful though. Bradley Orr is high up their list for the Championship and he is awful.

Vurnon San Benito

eagle eye wrote: »

Their stats are awful though. Bradley Orr is high up their list for the Championship and he is awful.

And you're prepared to argue against statistics?

inmyday

roanoke wrote: »

I heard it was calculated based on stride count (rather than measurement of actual distance run) but I'm readily prepared to be corrected on that.

Is it not just a chip???
Like the ones that runners are given in marathons? And the more advanced ones arent in the shape of a microchip, its more like a piece of string on the back of your number that you pin to your tshirt.

Doubt its stride count, cause many times I have saw players being substituted and it would say "distance covered 8.94km". No way could it be precise.

COYVB

eagle eye wrote: »

Their stats are awful though

How can stats be awful? They're a record of things that actually happened...

Pro. F

Doocey wrote: »

And you're prepared to argue against statistics?

They are not statistics. Not what a statistician would recognise anyway. What Opta and the like provide is collections of data points with the most simple of mathematical functions applied. They also have yet to prove any results over traditional scouting.

SeantheMan

Pro. F wrote: »

They are not statistics. Not what a statistician would recognise anyway. What Opta and the like provide is collections of data points with the most simple of mathematical functions applied. They also have yet to prove any results over traditional scouting.

Were they set up as an alternative to scouting though, to me they are just a fun fact for half/full time or during the match.

I doubt many managers put much stock by the statistics.

Vurnon San Benito

SeantheMan wrote: »

Were they set up as an alternative to scouting though, to me they are just a fun fact for half/full time or during the match.

I doubt many managers put much stock by the statistics.

Personally, I think managers are coming more and more involved with statistics. Brendan Rodgers is a big man for them as far as I know, especially passing rate, distance etc.

Pro. F wrote: »

They are not statistics. Not what a statistician would recognise anyway. What Opta and the like provide is collections of data points with the most simple of mathematical functions applied. They also have yet to prove any results over traditional scouting.

Stats are stats - mathematical facts, is there really any room for argument there?

BOHtox

He said it went to the keeper's left but it went to the keeper's right!

https://www.youtube.com/watch?v=xKc_6X0q11w

Eirebear

Doocey wrote: »

And you're prepared to argue against statistics?

Oh no.
Are you going to tell us that the statistics make the player?

Pro. F

Doocey wrote: »

Stats are stats - mathematical facts, is there really any room for argument there?

It's the reliability of the collection and the analysis of those mathematical facts, the science of statistics, that makes them useful or not. Opta does a lot of collecting data and they perform some basic mathematical functions on those statistics. But they show no evidence of the reliability of their data collection and they don't do any proper statistical investigation at all - from what I have seen anyway. So their use of the word 'statistics' gives the completely false impression of scientific authority to their findings.

SeantheMan wrote: »

Were they set up as an alternative to scouting though, to me they are just a fun fact for half/full time or during the match.

I doubt many managers put much stock by the statistics.

At the moment Opta and the other companies seem to be pushing themselves as a useful addition to scouting, which is fair enough. But the problem is they sound a lot more authoritative than they actually are.

batistuta9

Doocey wrote: »

Stats are stats - mathematical facts, is there really any room for argument there?

If there were any mathematical facts/formulas behind it then there wouldn't be different values appearing form different sources - there's the argument

possession being the biggest disparity from them i think

during the euro's over in the betting forum i got the OPTA stat's for some guy who could find them, their possession stats & the ones on the UEFA site were miles apart
But what mattered there was which one the bookie used

it's like Pro. F says above.