Antonio Conte - officially NOT the coach of THFC

[parklane1]

I think you will see stats and analytics play an ever more important role in football. There is a ton of data and, once teams figure out the right way to analyze it, it will probably show more truth than what the naked eye sees in a game. Not in one individual game (even though that could be possible too as analytics evolve), but over a period of games. I'm posting a bit from a 2017 article about Brentford, who are an extremely well-managed team and heavily reliant on stats and analytics, which I think points to what the future of football will be. Spurs sorely need as good an analytics dept as Brentford. And since we're talking about how we feel Conte has underachieved, I'm underlining below a passage that I found really interesting, but that makes a lot of sense.

Brentford measures success differently, not by the league table, but by Ankersen's "table of justice," a variant of the expected goals model. This is informed by the belief that, because football is such a low-scoring sport—the average game has 2.7 goals, compared to over 200 points in basketball—luck is very important. As The Numbers Game notes, favourites win only 65 percent of football matches but 80 percent of basketball games.

Brentford have the confidence to take a long-term view and not allow football's inherent randomness to cloud the club's analysis.

"We don't look so much at the league table position when we evaluate performance. What we look at are the underlying metrics, which we believe are a better indication of where we are going and how we've done," Ankersen says. "We know how we measure performance; we don't overreact to those swings in results that you see that's largely down to randomness, which you see in football because it's a low-scoring sport.

"Telling people that the league table lies is like telling people that the earth is flat. All their preconceptions are being challenged, and the media won't accept it because they rely on having tragedies and triumphs, so it's a difficult thing to say—especially when you are underachieving."

But Brentford do not merely use the "table of justice" as an excuse. In 2015, Mark Warburton was not offered a new contract as manager after Brentford came fifth. The club believed the team had been lucky, that their performances had not merited such impressive results. Warburton was also less enamoured with analytics than other senior figures.

I have worked for years in football and used stats all the time, however i repeat they are a tool to use and consider but they are NOT the be all and end all and should not be seen that way ( imo), where the trouble with them comes in ( particuly among supporters) is where they are given as a absolute and enable some to make the point they are trying to make and is not generally correct.

But each to their own i guess.

 

[K.D.D.D.D.Soc]

I don't know enough about how the magical stats are produced, but do they reflect the strength and form of teams involved, weather conditions, travel time to games, times of k.o. there are many factors which effect players and teams.

 

[Grays_1890]

Just watching our games and ignoring the stats, it doesn't feel like we are over achieving.

Exactly this, to over achieve we would have been busting a gasket playing a better standard then last year and clinging on to results we shouldn't have the right to. Anyone with eyes in their head can see we have not been anywhere near the same standard as last year, nor improved. I wouldn't say we have overachieved or Conte is overreaching on any basis, there are many reasons we are where we are, in part with others underachievement.

People often want to push the overachieving point because they want to highlight the board and money spent, another stat that doesn't tell the whole story.

I would say we have had a par year with scope each way, if someone sid below par I would agree, and 4th despite that NOT because we are punching above our weight and performing well

Sent from my SM-A127F using Fapatalk

 

[alekaras]

I don't know enough about how the magical stats are produced, but do they reflect the strength and form of teams involved, weather conditions, travel time to games, times of k.o. there are many factors which effect players and teams.

Not all stats do, because we don't have enough data points for some the things you outlined. Once we do have enough data points, I see no reason why these can't be modeled out too.

 

[Bedfordspurs]

I don't know enough about how the magical stats are produced, but do they reflect the strength and form of teams involved, weather conditions, travel time to games, times of k.o. there are many factors which effect players and teams.

Nope
They don’t allow for many variables
And football is so full of them
Crowd noise
History of fixture
State of pitch
3 random ones which affect teams too

 

[K.D.D.D.D.Soc]

Nope
They don’t allow for many variables
And football is so full of them
Crowd noise
History of fixture
State of pitch
3 random ones which affect teams too

As I thought, I've never been a fan of "stats" as I had to provide performance and volume stats monthly for my teams at work, they were pointless as they didn't reflect seasonal variations, staffing levels or details of jobs.

 

[Raziel]

I think you will see stats and analytics play an ever more important role in football. There is a ton of data and, once teams figure out the right way to analyze it, it will probably show more truth than what the naked eye sees in a game. Not in one individual game (even though that could be possible too as analytics evolve), but over a period of games. I'm posting a bit from a 2017 article about Brentford, who are an extremely well-managed team and heavily reliant on stats and analytics, which I think points to what the future of football will be. Spurs sorely need as good an analytics dept as Brentford. And since we're talking about how we feel Conte has underachieved, I'm underlining below a passage that I found really interesting, but that makes a lot of sense.

Brentford measures success differently, not by the league table, but by Ankersen's "table of justice," a variant of the expected goals model. This is informed by the belief that, because football is such a low-scoring sport—the average game has 2.7 goals, compared to over 200 points in basketball—luck is very important. As The Numbers Game notes, favourites win only 65 percent of football matches but 80 percent of basketball games.

Brentford have the confidence to take a long-term view and not allow football's inherent randomness to cloud the club's analysis.

"We don't look so much at the league table position when we evaluate performance. What we look at are the underlying metrics, which we believe are a better indication of where we are going and how we've done," Ankersen says. "We know how we measure performance; we don't overreact to those swings in results that you see that's largely down to randomness, which you see in football because it's a low-scoring sport.

"Telling people that the league table lies is like telling people that the earth is flat. All their preconceptions are being challenged, and the media won't accept it because they rely on having tragedies and triumphs, so it's a difficult thing to say—especially when you are underachieving."

But Brentford do not merely use the "table of justice" as an excuse. In 2015, Mark Warburton was not offered a new contract as manager after Brentford came fifth. The club believed the team had been lucky, that their performances had not merited such impressive results. Warburton was also less enamoured with analytics than other senior figures.

Thing is, these models are great for middle of the road teams.

Spurs under BMJ proved how "easy" it was to push into best of the rest with being smart

Anything above that become exponentially harder and playing the odds isn't enough

 

[billyiddo]

The teams at the top generally have the best of both worlds, the money to attract the best players/coaches and also having the analytics nailed down off the field.

So middling teams that are smart can take advantage of underperforming big teams and other sides around their level, but don't have the clout to match the rich teams that are run well off the field.

 

[Bedfordspurs]

I think you will see stats and analytics play an ever more important role in football. There is a ton of data and, once teams figure out the right way to analyze it, it will probably show more truth than what the naked eye sees in a game. Not in one individual game (even though that could be possible too as analytics evolve), but over a period of games. I'm posting a bit from a 2017 article about Brentford, who are an extremely well-managed team and heavily reliant on stats and analytics, which I think points to what the future of football will be. Spurs sorely need as good an analytics dept as Brentford. And since we're talking about how we feel Conte has underachieved, I'm underlining below a passage that I found really interesting, but that makes a lot of sense.

Brentford measures success differently, not by the league table, but by Ankersen's "table of justice," a variant of the expected goals model. This is informed by the belief that, because football is such a low-scoring sport—the average game has 2.7 goals, compared to over 200 points in basketball—luck is very important. As The Numbers Game notes, favourites win only 65 percent of football matches but 80 percent of basketball games.

Brentford have the confidence to take a long-term view and not allow football's inherent randomness to cloud the club's analysis.

"We don't look so much at the league table position when we evaluate performance. What we look at are the underlying metrics, which we believe are a better indication of where we are going and how we've done," Ankersen says. "We know how we measure performance; we don't overreact to those swings in results that you see that's largely down to randomness, which you see in football because it's a low-scoring sport.

"Telling people that the league table lies is like telling people that the earth is flat. All their preconceptions are being challenged, and the media won't accept it because they rely on having tragedies and triumphs, so it's a difficult thing to say—especially when you are underachieving."

But Brentford do not merely use the "table of justice" as an excuse. In 2015, Mark Warburton was not offered a new contract as manager after Brentford came fifth. The club believed the team had been lucky, that their performances had not merited such impressive results. Warburton was also less enamoured with analytics than other senior figures.

That’s the American way on pretty much everything
Use data to understand cause and effect, then improve
You need so many variables for data to work and yet the human element still cannot be modelled, just analysed after

 

[Lilbaz]

The teams at the top generally have the best of both worlds, the money to attract the best players/coaches and also having the analytics nailed down off the field.

So middling teams that are smart can take advantage of underperforming big teams and other sides around their level, but don't have the clout to match the rich teams that are run well off the field.

Also if a player/coach doesn't work they can easily replace.

 

[alekaras]

As I thought, I've never been a fan of "stats" as I had to provide performance and volume stats monthly for my teams at work, they were pointless as they didn't reflect seasonal variations, staffing levels or details of jobs.

And you didn't adjust for these variables to show normalized numbers?! No wonder you're not a fan. You're not using them correctly.

 

[Lilbaz]

And you didn't adjust for these variables to show normalized numbers?! No wonder you're not a fan. You're not using them correctly.

How for football? Chaos theory and the butterfly effect come into play. A footballer stubs his toe or has an arguument with his wife. Before a big game. Does brick crowd turns on them. Becomes dele ali for the rest of the season.

The beginning of every season we have "super computer predicts league". 10 different super computers. 10 different predictions.

You cannot put in all the variables into a computer. At least not for the forseable.

 

[alekaras]

Nope
They don’t allow for many variables
And football is so full of them
Crowd noise
History of fixture
State of pitch
3 random ones which affect teams too

That’s the American way on pretty much everything
Use data to understand cause and effect, then improve
You need so many variables for data to work and yet the human element still cannot be modelled, just analysed after

As I said, this doesn't work for one game, as it is not a statistically significant sample size. But over the course of a season or two, it can show you how well your team is performing relative to how it should perform.

And the models will evolve to factor in the human element. I wouldn't be surprised if in the future a players' backgrounds are analyzed from their socioeconomic status growing up, whether in a single-parent or two parent family, the level of competition they faced when they were 6, 7, 8 years old and so on. These are just example data points that will be combined with dozens of others to come up with profiles. Teams might even start having all players take the Myers-Briggs test to see what personality types they are and factor that into the equation too. Collecting all these data won't be easy and identifying the correct patterns even more so. But observing for all these variables and modeling them out correctly might tell you which player is more likely to succeed in which type of team, what types of games he will thrive in, what systems he is better suited for, and the list goes on. Again, as this is statistics you should not expect an exact correlation. A 75-80% degree of confidence would be excellent. And this also doesn't mean that there won't be outliers that you will have to be shrewd enough to identify.

We are still a long way from these types of analyses. The current ones are still pretty basic. But we will get there eventually. It's amazing what data can tell us, even about things that we believe to be random and not data-driven.

 

[alekaras]

How for football? Chaos theory and the butterfly effect come into play. A footballer stubs his toe or has an arguument with his wife. Before a big game. Does brick crowd turns on them. Becomes dele ali for the rest of the season.

The beginning of every season we have "super computer predicts league". 10 different super computers. 10 different predictions.

You cannot put in all the variables into a computer. At least not for the forseable.

As I mentioned in the post after yours, this type of in-depth analysis is in its nascent phase and the degree of confidence is not particularly high, due to the lack of data points. So yes, in the foreseeable future it won't be as accurate, but it will get there, and in the meantime it is better than not doing it at all. And I am not talking about predicting the league standings. I am talking about post mortem analysis where you evaluate if, controlling for a player stubbing his toe or having an argument with his wife or getting injured, your season went well. That's what normalizing is for. And based on that you plan for what you need to improve on the next season.

 

[Lilbaz]

As I mentioned in the post after yours, this type of in-depth analysis is in its nascent phase and the degree of confidence is not particularly high, due to the lack of data points. So yes, in the foreseeable future it won't be as accurate, but it will get there, and in the meantime it is better than not doing it at all. And I am not talking about predicting the league standings. I am talking about post mortem analysis where you evaluate if, controlling for a player stubbing his toe or having an argument with his wife or getting injured, your season went well. That's what normalizing is for. And based on that you plan for what you need to improve on the next season.

I'm not saying the analytics aren't useful. But they are still a tool. A human still inputs the information that the computer will need to anylase or at least what information it should analyse. We still need to have a human to decide if that analysis is right or wrong.

At the end of the day though football isn't solely results based. It is entertainment. It relies on giving the fans moments of joy. Which yes is largely down to results or goals or trophies but can be more than that. Fans don't get enough of those moments, they'll start to turn. At the end of the day they are the customers.

 

[K.D.D.D.D.Soc]

And you didn't adjust for these variables to show normalized numbers?! No wonder you're not a fan. You're not using them correctly.

No that's not they were for, they were to show the board what a marvellous job someone was doing, they were a complete waste of my time.

 

[alekaras]

I'm not saying the analytics aren't useful. But they are still a tool. A human still inputs the information that the computer will need to anylase or at least what information it should analyse. We still need to have a human to decide if that analysis is right or wrong.

At the end of the day though football isn't solely results based. It is entertainment. It relies on giving the fans moments of joy. Which yes is largely down to results or goals or trophies but can be more than that. Fans don't get enough of those moments, they'll start to turn. At the end of the day they are the customers.

I get what you're saying about the entertainment value being more than results or trophies. But you have to play some modicum of decent football over the course of a season to get the results, so you'll probably offer some "beautiful play" entertainment value beyond just the results. And let's face it, it really is 95% results and 5% everything else. I don't think the Leicester fans minded the agricultural football they were seeing every weekend, as long as the team was winning and on its way to a PL title. Stadiums will be packed if the team is progressing in a cup competition, or challenging for the title, no matter the football played. No team is going to play expansive football in this day and age just to "entertain" and regularly score 3-4 goals a game while conceding 5-6.

The bottom line is that for every club it is a results business: get to the top division, stay in the top division, finish higher in the table, challenge for a European spot, challenge for the title. That's where the money is and that's what the fans really want, even if they're complaining at times that their team is not playing free-flowing football. Give them 38 1-0 wins to win the title playing catenaccio and they will bite your hand off. These are the moments fans live for. You couple if with nice football, great. But even if you don't, I have yet to hear a fan whose team won a trophy say "I wish we had played better football, instead of winning the cup/title."

 

[Mr Gogolak]

But even if you don't, I have yet to hear a fan whose team won a trophy say "I wish we had played better football, instead of winning the cup/title."

It's the other way round: you appoint Mourinho and Conte and you say 'I wish we'd won a trophy to make up for watching this crap'.

 

[alekaras]

It's the other way round: you appoint Mourinho and Conte and you say 'I wish we'd won a trophy to make up for watching this crap'.

I agree, but your reply is taking what I said out of context.

 

[Mr Gogolak]

I agree, but your reply is taking what I said out of context.

Completely. It was a just a tongue-in-cheek comment and while I agree that the debate on stats and their use is an interesting one, I don't have enough time to add something meaningful!