Got one that is far more plausible ... PL games with saturday 3PM KO. (Is there a new article in the making ... ?? ) .
Here are the PL data sets for the ENIC era managers (manager / games played / win ratio / GF ratio / GA ratio) : Antonio Conte 53 0.584906 2 1.11321 Nuno Espírito Santo 10 0.5 0.9 1.6 Tim Sherwood 22 0.590909 1.81818 1.36364 Mauricio Pochettino 202 0.559406 1.84653 1.02475 André Villas-Boas 54 0.537037 1.5 1.24074 Harry Redknapp 144 0.493056 1.58333 0.979167 José Mourinho 58 0.465517 1.67241 1.15517 Martin Jol 114 0.412281 1.48246 1.29825 Glenn Hoddle 89 0.359551 1.30337 1.58427 David Pleat 33 0.363636 1.27273 1.36364 Jacques Santini 11 0.272727 0.545455 0.727273 Juande Ramos 35 0.285714 1.51429 1.42857
So on Conte (done for discussion but acknowledging his total number of games is statistically insignificant) ... His win ratio, and GF ratio is at the top end of Spurs managers. His GA ratio is closer to the mean than it is to the best. Given that Conte : 1. allegedly has some great defensive nous 2. has a meh GA ratio 3. has a GA with 2/3 of this season gone that is only 4 less than the total for last season is the case made for or against him continuing ??
Executive summary (for Spurs managers in the ENIC era) ... Conte is doing as well as enyone else did, when it comes to win ratio and average GF per game (his teams win as many, score as many per game etc) But his average GA per game is unremarkable (hence my previous remarks) .
Is your mean GA weighted by number of matches? Santini has by far the best number but very few matches so will skew the answer very considerably. Conte's number is much less than the mean GA per match over ENIC's reign but in some ways that isn't surprising because of the big upward trend in performance (probably driven by acquiring better players as finances improved). Do you have the tools to take the trend out and see who outperformed most then?
All the ratios are totals / number of matches. No weightings have been (or have to be) applied. I have objectively (as I would do in my day job) analysed the histograms (ie a multi-dimensional analysis) . The numbers are in for Conte on GA. My questions remain ...
So what're the odds that, now we are pretty much guaranteed to play one game a week, we go on a run of form and qualify for the CL...and people call for Conte to be given an extension because we can hit form in the league as soon as we crash out of everything else?
I don't think that's the right treatment. The baseline against which to compare each manager should be the average per game across all managers, not the average of the average of each other manager.
I agree with that but you then compared Conte with the mean...and deduced he wasn't significantly different from it on GA. But that mean is heavily affected by Santini and is way below the true mean of all events. If the dataset contained 10 managers with 1 game and no goal conceded, 9 with 100 games and 200 goals conceded and 1 with 200 games and 200 goals conceded then the mean by your method is 0.95 and the last manager is worse than it but is absolutely clearly the best of all of them.
It is what it is. An analysis of 3 measures for all appropriate managers, even though only 3 of them (of which Conte is NOT one) have "source data" that would be deemed statistically significant enough by professionals to be used in a dara analyrics process.
Ornstein reckons a Poch return is unlikely and would be an uphill battle. Rahman Osman says Paratici prefers Paulo Fonseca to Poch. I’d highly doubt Fonseca would want our job after the way we mugged him off prior to appointing Nuno. I’d got massively on board with his appointment at the time too. Type of manager that could fit this new progressive coach mould that the likes of Potter, Frank, De Zerbi, Cooper etc are all in.