Ellafella Posted July 28, 2020 Share Posted July 28, 2020 8 hours ago, Jram said: I don’t know what an SPC analysis is but do I think it’s possible that in a 46 game period any team could receive 6 red cards? Yes, that’s completely feasible and I don’t think it necessarily points to a fundamental lack of discipline in a team Statistical analysis will show that Derby County are a significant outlier in the distribution. Only Derby received 6 straight red cards. Time series analysis would in all likelihood show that 6 Straight red cards In a season would put you well above the average for any season and QED that underlying this was an issue causing it to happen. If I had the time and inclination I would do the analysis ... but I don’t need to do it to know there’s a problem. Link to comment Share on other sites More sharing options...
Mihangel Posted July 28, 2020 Share Posted July 28, 2020 8 hours ago, Jram said: I don’t know what an SPC analysis is but do I think it’s possible that in a 46 game period any team could receive 6 red cards? Yes, that’s completely feasible and I don’t think it necessarily points to a fundamental lack of discipline in a team I couldn't agree more. Lies, damned lies and stats as they say - I'm a sad lover of statistics but, especially in sport, they're only ever meaningful with appropriate context. There's nowhere near enough data or spread among that data to reach a strong conclusion. Given the stats as they are, one could argue that Derby were exemplary in the 'normal' season yet lost the plot in the 'Covid' season which, again, would be subject argument about the impact of the break and the mini season we had. A couple of missed timed tackles and you have 1/3rd of the straight reds for the season, I'm not sure this points to this nebulous 'problem with discipline' that people talk of. Link to comment Share on other sites More sharing options...
RamNut Posted July 28, 2020 Share Posted July 28, 2020 One of the six, is sibley’s red that was rescinded. Link to comment Share on other sites More sharing options...
Ghost of Clough Posted July 28, 2020 Share Posted July 28, 2020 1 hour ago, RamNut said: Fulham aren’t a bad benchmark for where we need to get to..... their final 7 games of the season yielded 13 points more than we got from the same fixtures. And the run of games before that? Link to comment Share on other sites More sharing options...
Ghost of Clough Posted July 28, 2020 Share Posted July 28, 2020 1 hour ago, Sparkle said: If we are fair all three of them give the ball away more than they give it to our own players, as for Lawrence his work rate improved since Rooney came in at January but a liability he remains when the opponent has the ball I wish it was different but it isn’t in my opinion If we're fair we'll be truthful. Lawrence and Holmes give the ball to our own players much more often than the opposition. Marriott only just favours the other team (not too unusual for a forward like him). Marriott Accurate passes = 163 Inaccurate passes = 56 Fouls = 18 Offside = 5 Shots (minus goals) = 41 Unsuccessful dribbles = 7 Lost possession = 48 Does he give the ball away more than he gives it to our own? Yes (difference of 12 occasions) Holmes Accurate passes = 881 Inaccurate passes = 219 Fouls = 38 Offside = 2 Shots (minus goals) = 28 Unsuccessful dribbles = 28 Lost possession = 107 Does he give the ball away more than he gives it to our own? No (difference of 459 occasions) Lawrence Accurate passes = 968 Inaccurate passes = 220 Fouls = 58 Offside = 7 Shots (minus goals) = 80 Unsuccessful dribbles = 42 Lost possession = 148 Does he give the ball away more than he gives it to our own? No (difference of 413 occasions) Link to comment Share on other sites More sharing options...
Ellafella Posted July 28, 2020 Share Posted July 28, 2020 39 minutes ago, ShoreRam said: I couldn't agree more. Lies, damned lies and stats as they say - I'm a sad lover of statistics but, especially in sport, they're only ever meaningful with appropriate context. There's nowhere near enough data or spread among that data to reach a strong conclusion. Given the stats as they are, one could argue that Derby were exemplary in the 'normal' season yet lost the plot in the 'Covid' season which, again, would be subject argument about the impact of the break and the mini season we had. A couple of missed timed tackles and you have 1/3rd of the straight reds for the season, I'm not sure this points to this nebulous 'problem with discipline' that people talk of. See my earlier response to @Jram. Statistically we are an outlier. Significantly so. Time series analysis of averages over previous seasons will show that we are at the top end of the distribution for disciplinary points, probably by more than 2sd. We should not ignore the evidence as random chance. Link to comment Share on other sites More sharing options...
RamNut Posted July 28, 2020 Share Posted July 28, 2020 41 minutes ago, Ghost of Clough said: And the run of games before that? Don’t know. I haven’t looked. But they finished strongly and were 17 points ahead of us. Link to comment Share on other sites More sharing options...
Mihangel Posted July 28, 2020 Share Posted July 28, 2020 54 minutes ago, RamNut said: One of the six, is sibley’s red that was rescinded. ? Step one of using statistics - Ensure their veracity. Link to comment Share on other sites More sharing options...
Jram Posted July 28, 2020 Share Posted July 28, 2020 6 minutes ago, Ellafella said: See my earlier response to @Jram. Statistically we are an outlier. Significantly so. Time series analysis of averages over previous seasons will show that we are at the top end of the distribution for disciplinary points, probably by more than 2sd. We should not ignore the evidence as random chance. But the sample size is so unbelievably small, you can’t draw any conclusions from it This is the problem with data, you can take the ones which suit your narrative Take our season: people are making all sorts of assumptions about why we were rubbish to start with and had a good run after Christmas but it’s probably just random... we were always going to lose a certain amount of games and they happened to fall towards the start Link to comment Share on other sites More sharing options...
Ellafella Posted July 28, 2020 Share Posted July 28, 2020 2 minutes ago, Jram said: But the sample size is so unbelievably small, you can’t draw any conclusions from it This is the problem with data, you can take the ones which suit your narrative Take our season: people are making all sorts of assumptions about why we were rubbish to start with and had a good run after Christmas but it’s probably just random... we were always going to lose a certain amount of games and they happened to fall towards the start Believe me, as an analyst working in health analytics we work on “never events” on a regular basis which involves small numbers of observations and there are statistical procedures and underlying theory which enables us to pinpoint statistical variations from normality. As an analyst of some 30 years I don’t need to justify this. It’s quite clear that we are an outlier and the Club really need to address it. If we look at historical data we will see a moderate but significant relationship between success and low/moderate disciplinary scores. It suggests a lack of professional discipline within the team over the course of a season which will detract from our overall aim of winning games. Common sense really. Would you climb aboard an aeroplane which had crashed 3 times in 2 years compared with 20 others which had never crashed or would you say the sample size is too small to conclude? Small numbers are not a barrier to statistical analysis. Link to comment Share on other sites More sharing options...
nottingram Posted July 28, 2020 Share Posted July 28, 2020 2 hours ago, RamNut said: Fulham aren’t a bad benchmark for where we need to get to..... their final 7 games of the season yielded 13 points more than we got from the same fixtures. They’re probably not a very fair benchmark though given their squad cost about 3x ours and our ability to invest in the squad is very much hindered by our financial situation, be that rules based or simply that Mel won’t put more cash in than he needs to anymore which is entirely justifiable. To compete with Fulham we are going to have to do things very differently to them. This season was an excellent step in that direction - three players came through the academy that look to be excellent and may be able to fetch us eight figure sums if we are sensible with them and if they don’t help us to promotion first. Basically for this reason I am happy with the season. Anyone sensible knew it would be transitional with ups and downs. The only blight for me was the well publicised off pitch issue that put paid to the DCFC careers of two useful to excellent players for us. Link to comment Share on other sites More sharing options...
Mihangel Posted July 28, 2020 Share Posted July 28, 2020 30 minutes ago, Jram said: But the sample size is so unbelievably small, you can’t draw any conclusions from it This is the problem with data, you can take the ones which suit your narrative Take our season: people are making all sorts of assumptions about why we were rubbish to start with and had a good run after Christmas but it’s probably just random... we were always going to lose a certain amount of games and they happened to fall towards the start You are, of course, entirely correct - You cannot just look at a (small) dataset and draw a conclusion like this. It's a nonsense to think you can. Link to comment Share on other sites More sharing options...
Jram Posted July 28, 2020 Share Posted July 28, 2020 19 minutes ago, Ellafella said: Believe me, as an analyst working in health analytics we work on “never events” on a regular basis which involves small numbers of observations and there are statistical procedures and underlying theory which enables us to pinpoint statistical variations from normality. As an analyst of some 30 years I don’t need to justify this. It’s quite clear that we are an outlier and the Club really need to address it. If we look at historical data we will see a moderate but significant relationship between success and low/moderate disciplinary scores. It suggests a lack of professional discipline within the team over the course of a season which will detract from our overall aim of winning games. Common sense really. Would you climb aboard an aeroplane which had crashed 3 times in 2 years compared with 20 others which had never crashed or would you say the sample size is too small to conclude? Small numbers are not a barrier to statistical analysis. It sounds like your job gives you a certain level of expertise and I have none so I’m inclined to believe you but i just can’t ? If you can find me some evidence that in the last ten years Derby have had significantly more red and yellow cards than other teams, I’d be willing to draw some conclusions about culture at the club or whatever but, football being the random game that it is, I’m not willing to accept that 6 red cards over such a tiny sample of games points to anything significant Link to comment Share on other sites More sharing options...
Ghost of Clough Posted July 28, 2020 Share Posted July 28, 2020 11 minutes ago, Ellafella said: Believe me, as an analyst working in health analytics we work on “never events” on a regular basis which involves small numbers of observations and there are statistical procedures and underlying theory which enables us to pinpoint statistical variations from normality. As an analyst of some 30 years I don’t need to justify this. It’s quite clear that we are an outlier and the Club really need to address it. If we look at historical data we will see a moderate but significant relationship between success and low/moderate disciplinary scores. It suggests a lack of professional discipline within the team over the course of a season which will detract from our overall aim of winning games. Common sense really. Would you climb aboard an aeroplane which had crashed 3 times in 2 years compared with 20 others which had never crashed or would you say the sample size is too small to conclude? Small numbers are not a barrier to statistical analysis. One metric to measure discipline is the number of cards received. Obviously, we're one of the worst. However, another measurement is the number of fouls committed, putting us 9th overall. We committed just under 10% more than the best team in the league, yet nearly 30% less than the worst team. Maybe not actually that ill-disciplined? In fact, taking the first half of games into isolation, we're only 9 fouls worse than 2nd placed Bristol. As games go on, we're more likely to foul - in the final 15 we're only 9 fouls better than 3rd worst Cardiff. This seems to coincide with us pushing late on for a goal, getting hit on the counter, then taking tactical yellows as a result. Link to comment Share on other sites More sharing options...
Tyler Durden Posted July 28, 2020 Share Posted July 28, 2020 10 minutes ago, ShoreRam said: You are, of course, entirely correct - You cannot just look at a (small) dataset and draw a conclusion like this. It's a nonsense to think you can. What sampling interval would you then propose would be of an acceptable size to draw robust conclusions from then. Are you saying we are just unlucky then as all of the other teams in our division who haven't received our number of red cards in the season avoided them due to good fortune, biased officiating etc? Link to comment Share on other sites More sharing options...
nottingram Posted July 28, 2020 Share Posted July 28, 2020 Malone - last man foul, probably slightly on the harsh side as there was no indication it was deliberate, certainly no more deliberate than Forsyths in the same game but definitely not ‘ill discipline,’ just clumsiness. Bielik - rubbish, mistimed tackle. Red was fair but I wouldn’t really consider it as ill discipline. Lowe - hands on ref, definitely poor discipline as he lost his head Lawrence - handbags with Miazga. Definitely ill discipline as he allowed himself to be goaded, but Miazga had two swipes at him before he reacted. Not sure many would walk away from that. Waghorn - poor tackle, clumsy, strikers challenge. No malice in it, he wasn’t even looking at Yates which is a shame because I’d have loved it if he’d meant it. Not really ill discipline. Sibley - rescinded, nothing more needs to be said. One of the worst red card decisions I’ve ever seen. Link to comment Share on other sites More sharing options...
YorkshireRam Posted July 28, 2020 Share Posted July 28, 2020 3 hours ago, Gritty said: The problem with this analysis IMO is that it doesn't take into account we are already massively loss making and - were it not for stadium sale - would have obliterated FFP. So the wage savings you highlight won't free up cash for new signings, they'll reduce our losses and hopefully bring us back into compliance with FFP. But I don't expect we'll be paying anything other than nominal fees for players unless there are significant outgoings. And players coming in will be on much lower wages That wasn't my point, it was that the net savings from departures are bigger than net spending on Rooney; the figures don't need to be wholly accurate with such large margins. Funds for transfers are not usually the same as wages anyway, hence why I mentioned adopting a sell-to-buy policy which you later agreed with? 13 hours ago, RamNut said: Don’t know where you getting those salaries from. If 32red money = 30k per week then we could be missing Another 30-50k per week for Rooney. Doesn't matter that they're not totally accurate with a £61k difference, that's enough to disprove your assertion that the money saved on departures is ''offset'' by Rooney's wages which was the only point. Link to comment Share on other sites More sharing options...
RoyMac5 Posted July 28, 2020 Share Posted July 28, 2020 47 minutes ago, Ghost of Clough said: One metric to measure discipline is the number of cards received. Obviously, we're one of the worst. However, another measurement is the number of fouls committed, putting us 9th overall. We committed just under 10% more than the best team in the league, yet nearly 30% less than the worst team. Maybe not actually that ill-disciplined? In fact, taking the first half of games into isolation, we're only 9 fouls worse than 2nd placed Bristol. As games go on, we're more likely to foul - in the final 15 we're only 9 fouls better than 3rd worst Cardiff. This seems to coincide with us pushing late on for a goal, getting hit on the counter, then taking tactical yellows as a result. This seems a much more valuable use of the stats. I don't know if the conclusion is correct or not, but we have scored late goals. Link to comment Share on other sites More sharing options...
brady1993 Posted July 28, 2020 Share Posted July 28, 2020 1 hour ago, Ellafella said: Believe me, as an analyst working in health analytics we work on “never events” on a regular basis which involves small numbers of observations and there are statistical procedures and underlying theory which enables us to pinpoint statistical variations from normality. As an analyst of some 30 years I don’t need to justify this. It’s quite clear that we are an outlier and the Club really need to address it. If we look at historical data we will see a moderate but significant relationship between success and low/moderate disciplinary scores. It suggests a lack of professional discipline within the team over the course of a season which will detract from our overall aim of winning games. Common sense really. Would you climb aboard an aeroplane which had crashed 3 times in 2 years compared with 20 others which had never crashed or would you say the sample size is too small to conclude? Small numbers are not a barrier to statistical analysis. I accept the small sample isn't necessarily a barrier to making a conclusion but it does mean you can be disproportionately affected by noise especially in domain like football where there are a lot of uncontrolled variables. Whilst I don't think it's worth ignoring (I'll get onto this a little later) it's worth keeping in mind that it might be a random fluctuation especially if it's just this season that is the outlier for us on a disciplinary front. It's also worth keeping in mind that one of the red cards counted for us shouldn't have been a red card (Sibley's rescinded red card). With all that said there is strong enough evidence to prompt further investigation to see what might be causing it and if there is anything that can be done to curb it. As an aside at a glance a couple things stood out about our numbers; our yellow cards are high but relatively in line for the rest of the league and that for the number of red cards we have it looks a bit odd that we've had no double yellows. Link to comment Share on other sites More sharing options...
brady1993 Posted July 28, 2020 Share Posted July 28, 2020 1 hour ago, Ghost of Clough said: One metric to measure discipline is the number of cards received. Obviously, we're one of the worst. However, another measurement is the number of fouls committed, putting us 9th overall. We committed just under 10% more than the best team in the league, yet nearly 30% less than the worst team. Maybe not actually that ill-disciplined? In fact, taking the first half of games into isolation, we're only 9 fouls worse than 2nd placed Bristol. As games go on, we're more likely to foul - in the final 15 we're only 9 fouls better than 3rd worst Cardiff. This seems to coincide with us pushing late on for a goal, getting hit on the counter, then taking tactical yellows as a result. Out of pure curiosity where are you sourcing your data, as I haven't seen that level of data provided in many places ? Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.