Maffs 'n' graffs - statistical porn

  • The Fighting Cock is a forum for fans of Tottenham Hotspur Football Club. Here you can discuss Spurs latest matches, our squad, tactics and any transfer news surrounding the club. Registration gives you access to all our forums (including 'Off Topic' discussion) and removes most of the adverts (you can remove them all via an account upgrade). You're here now, you might as well...

    Get involved!

Latest Spurs videos from Sky Sports

I've change the parameter of the last 'n' to be 10 in all competitions, which seems to be a fairer number which allows Santini to be considered
Right, I saw that just after I pressed "post reply".

Off the top of my head, I'd suggest this:

Considering the reasonable expectation that Levy has an itchy trigger finger, it seems fair to assume that that he reacts to local, catastrophic events. That's my suspicion w/ AVB (and, like you, I think AVB had a lot of agency vis-à-vis the departure). That's also the media's version of the Ramos story (two pts from eight games). But Ramos oversaw a dip in form over a longer spell of matches than just eight. But the proof is also in the pudding for just the final eight. That is, Ramos's goose may have been cooked in May, but if the following season had started with 24 points, he would have stayed. Redknapp's final ten might also be what sealed his demise. So looking at the final ten feels about right, at least for starters.

I'd then compare that win percentage (and goal differential and goals for, since I suspect those tell the tale a bit more starkly) to a similarly sized random sample over the manager's entire tenure at Spurs and to a randomly selected ten-match stretch during the manager's tenure at Spurs (or map a ten-match rolling average over his whole tenure—that's probably better). The former "should" look like the manager's average win percentage over his entire tenure, but if it doesn't, it'll give you a sense of how much variance is going on. The latter would help see how much correlation there is between matches. It makes a lot of sense to suggest that matches are heavily dependent upon one another, after all.

Since ten is a small number for a sample, there's going to be volatility, which is why I think it'd be useful to try to tame that volatility by providing some context via samples over the course of the tenure and via rolling averages.

I'd probably do it myself, since now I'm intrigued, but I'm terribly busy atm, and it seems like you've already started building the dataset.
 
Having thought about it for an infinity in internet time (seven minutes), I think a graph of the ten-match rolling averages of win percentage, GD, and GF would be very, very sexy.
 
Right, I saw that just after I pressed "post reply".

Off the top of my head, I'd suggest this:

Considering the reasonable expectation that Levy has an itchy trigger finger, it seems fair to assume that that he reacts to local, catastrophic events. That's my suspicion w/ AVB (and, like you, I think AVB had a lot of agency vis-à-vis the departure). That's also the media's version of the Ramos story (two pts from eight games). But Ramos oversaw a dip in form over a longer spell of matches than just eight. But the proof is also in the pudding for just the final eight. That is, Ramos's goose may have been cooked in May, but if the following season had started with 24 points, he would have stayed. Redknapp's final ten might also be what sealed his demise. So looking at the final ten feels about right, at least for starters.

I'd then compare that win percentage (and goal differential and goals for, since I suspect those tell the tale a bit more starkly) to a similarly sized random sample over the manager's entire tenure at Spurs and to a randomly selected ten-match stretch during the manager's tenure at Spurs (or map a ten-match rolling average over his whole tenure—that's probably better). The former "should" look like the manager's average win percentage over his entire tenure, but if it doesn't, it'll give you a sense of how much variance is going on. The latter would help see how much correlation there is between matches. It makes a lot of sense to suggest that matches are heavily dependent upon one another, after all.

Since ten is a small number for a sample, there's going to be volatility, which is why I think it'd be useful to try to tame that volatility by providing some context via samples over the course of the tenure and via rolling averages.

I'd probably do it myself, since now I'm intrigued, but I'm terribly busy atm, and it seems like you've already started building the dataset.


It's possible to drill a lot further, but not always completely necessary:
Redknapp was more than likely sacked for his arrogance and lack of loyalty - the last 10 games are skewed slightly as the final 4 were after Hodge got the England job, miraculously Harry started caring again and we didn't lose again. I don't think it was purely based on the final form.
AVB has confirmed in the last day or so that his departure was contractual (see AVB thread in general football) so that's not form based, IMO.
Santini left of his own accord
Graham was binned for being a gooner twat and, also, contractual breaches

That leaves 5 managers, 4 of which had guaranteed relegation form in their last 10 league matches, the only one that actually stands out as a little harsh is Christian Gross, but we know that was media pressure.

Funny how your mind plays tricks on you and you remember the managers being awful, when perhaps on paper they weren't. What stats can't tell you is if the football is 'exciting'
 
Thanks for the AVB tip. As I've been busy, I've not been reading here as much. That said, I prefer to leave AVB's comment as very ambiguous and intriguing, but I don't know about the dropping players contract part. Maybe we'll never know.

When I was writing the previous post, I was thinking of Santini in particular as a problematic example. As far as we know, his departure was entirely of his own design, so we can't use it to measure something like "what kind of form gets you the sack?"

But if we were in first and not playing Ade, then Levy would have never had the meeting that led to AVB's departure (I don't think). Similarly, if we had finished above ww in Redknapp's last year, I think it's reasonable to guess that he might have also stayed.

Which is to say that even though both managers' departures may have been locally not form-related, I hypothesise that the form occasioned an environment where other issues became sackable offences. Which is why I do think charting the stuff might be useful (even in the case of Santini!!!)
 
Got every match & score since start of '94/95 in a spreadsheet now.

Just a little tidbit of info from it

winrate.jpg


Harry's goal difference :harrylol:
 
Got every match & score since start of '94/95 in a spreadsheet now.

Just a little tidbit of info from it

winrate.jpg


Harry's goal difference :harrylol:
Harry had a tittle winning side in 2011/12, even his subs were of decent to good quality. I saw it as a farce that we finished 4th in that season.

Still enjoyed it the most out of recent seasons
 
So much for deciding I'm too busy…
tBe1Vdq.png

fBuyPCG.png

bHtS6Qt.png

Most of what we see here shouldn't be surprising, though it certainly underscores Harry's "they've never had it so good" comment. The start of his final season was absolutely amazing, and it will be one of my fondest memories of Spurs, despite the poundings we got at the start of that season from both Manchester clubs.

Also, AVB's first season was a study in sturdy consistency. Compare our goals per game rate over that season to any other era. The goal difference was also relatively stable… probably the stablest since Jol's first season. And once Harry's relegation stink washed off the average, our winning percentage was consistently reasonable.

The winning percentage table also suggests our growth as a club under Enic. Under Hoddle, we broke 50% once. Jol hit 60% twice. Harry spent almost half a season above 60% before throwing it all away on draws and losses, but we also spent a lot of time hovering around 40% under him. In contrast, AVB was establishing himself with an average in the mid 50s before getting sacked.

Christ, we were schizophrenic under Harry…
 
So much for deciding I'm too busy…
tBe1Vdq.png

fBuyPCG.png

bHtS6Qt.png

Most of what we see here shouldn't be surprising, though it certainly underscores Harry's "they've never had it so good" comment. The start of his final season was absolutely amazing, and it will be one of my fondest memories of Spurs, despite the poundings we got at the start of that season from both Manchester clubs.

Also, AVB's first season was a study in sturdy consistency. Compare our goals per game rate over that season to any other era. The goal difference was also relatively stable… probably the stablest since Jol's first season. And once Harry's relegation stink washed off the average, our winning percentage was consistently reasonable.

The winning percentage table also suggests our growth as a club under Enic. Under Hoddle, we broke 50% once. Jol hit 60% twice. Harry spent almost half a season above 60% before throwing it all away on draws and losses, but we also spent a lot of time hovering around 40% under him. In contrast, AVB was establishing himself with an average in the mid 50s before getting sacked.

Christ, we were schizophrenic under Harry…

For those looking blankly at those charts they indicate performance over last 15 games - e.g game 1 to 15, 2 to 16, 3 to 17 upto 24 to 38 then game 25 of previous season to game 1 of next season etc..
 
Got every match & score since start of '94/95 in a spreadsheet now.

Just a little tidbit of info from it

winrate.jpg


Harry's goal difference :harrylol:
Given the amount of games Jol's GD is surprisingly weak. We're we that bad at the back under him?

Enjoying this thread bye. Even if some of the discussion leaves my head spinning. I am a man of letters rather than numbers!
 
Firstly CJJ CJJ - I fuckin love you. Marry me?

I cook, clean, and give blowjobs for a small fee. Let me know.

Okay, little Messy with 10 Managers on there, might be best to 'zoom' in a bit to see the figures:


winrate2.jpg
This graph makes my cum boil. Superb work. Visual is always better than 146 lines of dross.

Pretty shitty no-ones managed an above 50% win average within first 20 games.
Apart from 2 of them!

Ermm.... Sorry André!

Another very interesting one. Last 20 games. Done for both league games and in all competitions (with points column used based on the league games)

Edited: Santini's stats were out, fixed...


winrate.jpg
It does bring into stark reality the harsh sacking of AVB, if based solely on winning games. We can't expect much better than that. Poor chap.

Keep 'em coming. This thread is my new hobby.
 
Firstly CJJ CJJ - I fuckin love you. Marry me?

I cook, clean, and give blowjobs for a small fee. Let me know.


This graph makes my cum boil. Superb work. Visual is always better than 146 lines of dross.


Apart from 2 of them!


It does bring into stark reality the harsh sacking of AVB, if based solely on winning games. We can't expect much better than that. Poor chap.

Keep 'em coming. This thread is my new hobby.


To be fair his 50% comment was right at the time, I missed a couple of FA Cup games (as above) and fixed the charts.

I know you like visual, but sometimes it doesn't work. I have noticed that the graph for first 20 shows that, disregarding Gross and an outlier, regardless of how well they start, they all converge into one similar stat. #Spursy

Given the amount of games Jol's GD is surprisingly weak. We're we that bad at the back under him?

Enjoying this thread bye. Even if some of the discussion leaves my head spinning. I am a man of letters rather than numbers!


He wasn't the worst. I've made a table for you though.

winrate.jpg


I've extrapolated the scores for each game and disregarded any clean sheets. This can give us a better idea of how leaky we were when we weren't leaking...

Final column is just indicative of the distribution of goals, how many games we conceded 'x' number in
 
To be fair his 50% comment was right at the time, I missed a couple of FA Cup games (as above) and fixed the charts.

I know you like visual, but sometimes it doesn't work. I have noticed that the graph for first 20 shows that, disregarding Gross and an outlier, regardless of how well they start, they all converge into one similar stat. #Spursy




He wasn't the worst. I've made a table for you though.

winrate.jpg


I've extrapolated the scores for each game and disregarded any clean sheets. This can give us a better idea of how leaky we were when we weren't leaking...

Final column is just indicative of the distribution of goals, how many games we conceded 'x' number in
Dare I ask, who did we concede 7 against under Gerry Francis? Bit before my time...
 
Similarly, for scoring (with a % spread at the end, calculated on goals scored in games where we score)

winrate.jpg


Notice how, despite Harry's "brand" of football and AVB's drought this season, both AVB and Ramos had a smaller misfire % (games where we failed to score) than any other manager
 
Here's a Cumulative Win Rate for the data I have (Games starting from Ossie's first match of 94/95)
The figure is reliable as a growth/improvement indicator rather than what our actual win rate is

Accidentally chopped the left off, scale is zoomed in from 35% to 50% or so. Labels are there anyhow

(just as a quick Edit, didn't realise I'd put the manager faces in roughly the right place, totally unintentional!)
cumu.jpg
 
Last edited:
Finally, last one for now, managerial 'progression' chart. If you look carefully, you'll see Mr Gross just in view. Santini and Sherwood get lost in the crowd on this one due to lack of games played

cumu.jpg
 
Back
Top Bottom