Saints' Last 5 Seasons (according to Lawro)

[kwsaint]

I was discussing with my Dad earlier this afternoon whether Lawro actually has any idea what he's on about when predicting football results. This resulted in me doing some trawling through the internet to see where he predicted us to finish for the past 5 seasons (since we were in the PL and including this season). Here they are (in case anyone is interested )

 

2012/13:

2013/14

2014/15

2015/16

2016/17 (so far)

Edited 12 April, 2017 by kwsaint
Lack of computer skills

[Nolan]

He's only predicted we win 4 out of 31 this season? sheesh.

 

Sent from my Nexus 6P using Tapatalk

[saint si]

2015/16 table has some mistakes in it i think?

[Lets B Avenue]

2015/16 table has some mistakes in it i think?

 

For a start, we did finish 6th. He was only one point out as well, on the generous side.

[mrfahaji]

The caveat with these predictions is that they are done on a week by week basis. If I was predicting, say Liverpool to win against Burnley and they lost, I would probably predict them to bounce back and win the next one against Stoke. If they lost that one as well, I might think "I can't see them losing three on the bounce, so I'm going to predict they beat Everton."

 

I've predicted them to win 3/3, whereas if you had asked me at the start of the three games how many they'd win, I might only say 1 or 2. It's why you get extremes at either end of the table - I doubt he'd have predicted Hull only to win 2 games all season, or Liverpool to go undefeated if you'd asked for his predicted table at the start of the campaign.

 

Doesn't mean his predictions aren't silly though. I used to play a "predict Lawro's predictions" game with some colleagues, was much more fun, although also pretty easy...!

[OttawaSaint]

So apart from our first season back and this season he has been pretty much on the money regarding Saints.

[Turkish]

Do you really have so little going on in your life that you care so much what a bloke who stopped playing football before you were probably born thinks our score will be?

[OttawaSaint]

Do you really have so little going on in your life that you care so much what a bloke who stopped playing football before you were probably born thinks our score will be?

 

Do you really have so little going on in your life that you care so much to comment about what a bloke decided to post on a football forum?

 

INB4: well you just did

[CB Fry]

The caveat with these predictions is that they are done on a week by week basis. If I was predicting, say Liverpool to win against Burnley and they lost, I would probably predict them to bounce back and win the next one against Stoke. If they lost that one as well, I might think "I can't see them losing three on the bounce, so I'm going to predict they beat Everton."

 

I've predicted them to win 3/3, whereas if you had asked me at the start of the three games how many they'd win, I might only say 1 or 2. It's why you get extremes at either end of the table - I doubt he'd have predicted Hull only to win 2 games all season, or Liverpool to go undefeated if you'd asked for his predicted table at the start of the campaign.

 

While this is very true, and also very simple to understand, it will never be accepted by a small group of utter helmets that populate this forum.

 

Please say we'll win Lawro please say we'll win. Just say we'll win Lawro. Please say we'll win. Please Lawro. Lawro please. Please. Lawro. Please.

[OttawaSaint]

Also:

 

BBC ANTI SAINTS AGENDA #trollfacts

[Whitey Grandad]

Do you really have so little going on in your life that you care so much to comment about what a bloke decided to post on a football forum?

 

INB4: well you just did

 

Tu quoque

[OttawaSaint]

While this is very true, and also very simple to understand, it will never be accepted by a small group of utter helmets that populate this forum.

 

 

Likely the same helmets that were crying into their wrist protectors that we were getting relegated after going 1-0 down to Palace.

[Saint_clark]

While this is very true, and also very simple to understand, it will never be accepted by a small group of utter helmets that populate this forum.

 

Please say we'll win Lawro please say we'll win. Just say we'll win Lawro. Please say we'll win. Please Lawro. Lawro please. Please. Lawro. Please.

 

It is all a bit weird, however the fact he's predicted us as only getting 4 wins this season compared to his predictions in previous seasons is a little odd!

[SKD]

Do you really have so little going on in your life that you care so much what a bloke who stopped playing football before you were probably born thinks our score will be?

 

I think this every week when I see people getting aggy when he predicts us to lose

[OttawaSaint]

Tu quoque

 

Croque? As in bite?

[Professor]

I don't enjoy Lawrenson on TV, whilst as a 'football expert' I'm not sure what he's done since he stopped playing to justify the description, except to talk about football. If that makes him an expert, everyone on here is equally qualified. But having now seen his predictions coming so close to reflecting Saints' actual finishing positions in the table suppose he has to be given some credit where credit is due. Perhaps he's not smug, self-important and boring after all.

[mrfahaji]

I don't enjoy Lawrenson on TV, whilst as a 'football expert' I'm not sure what he's done since he stopped playing to justify the description, except to talk about football. If that makes him an expert, everyone on here is equally qualified. But having now seen his predictions coming so close to reflecting Saints' actual finishing positions in the table suppose he has to be given some credit where credit is due. Perhaps he's not smug, self-important and boring after all.

 

Not sure if this is a sarcastic post... Anyone could predict on a week by week basis and end up with an almost perfect league table at the end of the season. A more (albeit only slightly...) interesting stat would be a league table of how many scorelines per team he got right over the season. Would you give him credit if he predicted every win as a loss and vice versa, yet still ended up with the same points/position?

[angelman]

So apart from this season he hasn't been too bad?? 1 place wrong save for our first season up, where it was 4 places out. If you really want to see how accurate he is, why not run something like Spearman's Rank Correlation Coefficient? That way, rather than forming an unscientific opinion, you can get an accurate one.

[angelman]

As its a bit quiet at work today, I seem to have averaged out the seasons standings, and got a correlation coefficient of 0.725 (where 1 would be Lawro being 100% correct, and -1 being 100% incorrect).

 

HTH.

[benjii]

It really does help a lot.

 

Thank you.

[saintmatt]

Do you really have so little going on in your life that you care so much what a bloke who stopped playing football before you were probably born thinks our score will be?

 

27,500 posts on an internet forum.

 

I'm not knocking it, but those in glass houses and all that - live and let live!

[Chris27687]

27,500 posts on an internet forum.

 

I'm not knocking it, but those in glass houses and all that - live and let live!

 

You could extrapolate further.

 

27,500 posts over 10 years.

 

Let's be generous and say 10 minutes on the site per post - comes to around 458 hours on here a year, 38 hours a month, 8.5 hours a week.

 

If you had spend the time working at minimum wage you could have had an extra £3435/year in your back pocket (minus taxes).

 

You could have purchased an industrial walk in fridge and had money left over for a season ticket. http://secondhand-catering-equipment.co.uk/cat/307/walk-in-fridges-and-freezers

[kwsaint]

Do you really have so little going on in your life that you care so much what a bloke who stopped playing football before you were probably born thinks our score will be?

 

Says the man who posts on average 7.5 times a day. Pot and kettle somewhat?

[Hokie]

Not Saints related, but he seems to generally believe safety is easier than it is. He has teams in 17th predicted, by year, as achieving 30, 36, 32, 34, and 20 points. For this year he predicted 9th place would be on 38 points! Is he still on 2 pts for a win?

[mrfahaji]

Not Saints related, but he seems to generally believe safety is easier than it is. He has teams in 17th predicted, by year, as achieving 30, 36, 32, 34, and 20 points. For this year he predicted 9th place would be on 38 points! Is he still on 2 pts for a win?

 

Oh FFS, this isn't a representation of how he thinks the table will/should end up! How difficult a concept is it to grasp?!

[LGTL]

The caveat with these predictions is that they are done on a week by week basis. If I was predicting, say Liverpool to win against Burnley and they lost, I would probably predict them to bounce back and win the next one against Stoke. If they lost that one as well, I might think "I can't see them losing three on the bounce, so I'm going to predict they beat Everton."

 

I've predicted them to win 3/3, whereas if you had asked me at the start of the three games how many they'd win, I might only say 1 or 2. It's why you get extremes at either end of the table - I doubt he'd have predicted Hull only to win 2 games all season, or Liverpool to go undefeated if you'd asked for his predicted table at the start of the campaign.

 

Doesn't mean his predictions aren't silly though. I used to play a "predict Lawro's predictions" game with some colleagues, was much more fun, although also pretty easy...!

 

And this is also why bookies clean up on football.

[Hokie]

Oh FFS, this isn't a representation of how he thinks the table will/should end up! How difficult a concept is it to grasp?!

 

I would certainly say it does. It's a very intelligent way of aggregating his predictions in a meaningful sense.

 

Being done on a week to week basis, I would actually expect it to be more accurate than the beginning of the season one as you have knowledge of form, injuries, suspensions, players bought and sold in January, etc, minus your point about expecting a team to "bounce back" after a loss. (Which if it truly is a real effect, if you are aware of previous performance would improve your prediction. Your case simply begs the question.)

 

Anyhow, the criticism over the years is that Lawro has favours the top teams. I ran this season's predictions (aggregated in table form) against actual points. The dashed line is perfection (y = x). Any team with higher predictions than actual performances lies below that line. Any team performing better than predictions is above it.

 

https://postimg.org/image/vrhcinto5/

 

FWIW, the slope of the predictive model from a regression I ran is 0.573 (rather than 1) with standard error of 0.092, which means its more than 4.5 standard deviations from a slope of 1. I can say with 99.999% confidence that the predictions are biased towards teams towards the top.

 

[Note: Anyhow, my I don't have a Ph.D. in stats, and likely have been picking up more lately at work than I am actually in full command of, so if anyone wants to critique the analysis, have at it.]

[mrfahaji]

I would certainly say it does. It's a very intelligent way of aggregating his predictions in a meaningful sense.

 

Being done on a week to week basis, I would actually expect it to be more accurate than the beginning of the season one as you have knowledge of form, injuries, suspensions, players bought and sold in January, etc, minus your point about expecting a team to "bounce back" after a loss. (Which if it truly is a real effect, if you are aware of previous performance would improve your prediction. Your case simply begs the question.)

 

Anyhow, the criticism over the years is that Lawro has favours the top teams. I ran this season's predictions (aggregated in table form) against actual points. The dashed line is perfection (y = x). Any team with higher predictions than actual performances lies below that line. Any team performing better than predictions is above it.

 

https://postimg.org/image/vrhcinto5/

 

FWIW, the slope of the predictive model from a regression I ran is 0.573 (rather than 1) with standard error of 0.092, which means its more than 4.5 standard deviations from a slope of 1. I can say with 99.999% confidence that the predictions are biased towards teams towards the top.

 

[Note: Anyhow, my I don't have a Ph.D. in stats, and likely have been picking up more lately at work than I am actually in full command of, so if anyone wants to critique the analysis, have at it.]

 

The only thing the table shows is how regularly teams perform better or worse than he thinks in a given week. This season, when he has looked at the week's fixtures, he clearly hasn't fancied us, and we've proved him wrong many times. I doubt he tipped us for relegation at the start of the season though. Your graph is good for showing the teams he over/under rates on a weekly basis, and maybe that is reflective of how he perceives them in general, but I think that's as far as it goes; you can't use the points from his table to assume he thinks 30pts is enough to avoid relegation...!

 

If Lawrenson's main objective was to get an accurate league table, all he'd have to do is predict based on what's needed to get closer to the actual. For example, "Spurs have won 14 games so far, I've only predicted them to win 11 so far, so I'll predict them to win this week." After a week or two this wouldn't give a perfect league table but after a while you'd start getting pretty close. So to somehow think that the league tables above are reflective of his "season" predictions completely misses the point, and is a perfect example of how statistics can be misused and misinterpreted. And that is coming from someone who bl00dy loves data and statsistics!

Edited 15 April, 2017 by mrfahaji