dkmmgcycbwc Posted February 15, 2008 Share Posted February 15, 2008 ... who reads Kickback. Publish these facts and then comment / explain / justify, please. Or are you all in the pockets of the SFA? You understate the case massively Hearts Record in all competitions over the same period is Pld: 354, W 157, D 80, L 117 With McDonald It is Pld: 26, W 4, D 12, L 10 This gives us a reasonable size sample Without McDonald Pld 328, W 153, D 68, L 107 This gives us the control Sample If we take the hypothesis to be that MacDonald's refereeing affects hearts ability to win games then the antihypothesis would be that the results of Hearts games are independent of whether MacDonald is refereeing or not Were the Antihypothesis true it would mean that the observed sample would be a random selection from the same population as the control. 26 games selected at random from the 328 game control sample would be expected to average around 12 wins. WIth a 90% probability that the number of wins would be between 8 and 12 The Probability of 26 games chosen at random from the 328 games played over the period containing less than 5 wins is 0.087% (1 in 1155). The antihypothesis can be discarded The hypothesis is proved. Link to comment Share on other sites More sharing options...
Commander Harris Posted February 15, 2008 Share Posted February 15, 2008 ... who reads Kickback.Publish these facts and then comment / explain / justify, please. Or are you all in the pockets of the SFA? You understate the case massively Hearts Record in all competitions over the same period is Pld: 354, W 157, D 80, L 117 With McDonald It is Pld: 26, W 4, D 12, L 10 This gives us a reasonable size sample Without McDonald Pld 328, W 153, D 68, L 107 This gives us the control Sample If we take the hypothesis to be that MacDonald's refereeing affects hearts ability to win games then the antihypothesis would be that the results of Hearts games are independent of whether MacDonald is refereeing or not Were the Antihypothesis true it would mean that the observed sample would be a random selection from the same population as the control. 26 games selected at random from the 328 game control sample would be expected to average around 12 wins. WIth a 90% probability that the number of wins would be between 8 and 12 The Probability of 26 games chosen at random from the 328 games played over the period containing less than 5 wins is 0.087% (1 in 1155). The antihypothesis can be discarded The hypothesis is proved. excellent post, I've been thinking of doing something similar for a while. It is truly unbelievable when you see our record with McDonald. proof beyond all reasonable doubt that he does not referee us fairly and that a referee can have a big influence on a game. Link to comment Share on other sites More sharing options...
Kemptons feet Posted February 15, 2008 Share Posted February 15, 2008 The result of my investigation is that the above is proven to be pointless. Link to comment Share on other sites More sharing options...
Deevers Posted February 15, 2008 Share Posted February 15, 2008 Scottish Press - Objective journalism!!!! Do me a favour...!!! Link to comment Share on other sites More sharing options...
Sherlock Posted February 15, 2008 Share Posted February 15, 2008 You should have sent it into the club first so they could have requested that McDonald was de-selected from refereeing our game. Imagine two clubs publically stating that they had no faith in our officials. Link to comment Share on other sites More sharing options...
Boab Posted February 15, 2008 Share Posted February 15, 2008 ... who reads Kickback.Publish these facts and then comment / explain / justify, please. Or are you all in the pockets of the SFA? You understate the case massively Hearts Record in all competitions over the same period is Pld: 354, W 157, D 80, L 117 With McDonald It is Pld: 26, W 4, D 12, L 10 This gives us a reasonable size sample Without McDonald Pld 328, W 153, D 68, L 107 This gives us the control Sample If we take the hypothesis to be that MacDonald's refereeing affects hearts ability to win games then the antihypothesis would be that the results of Hearts games are independent of whether MacDonald is refereeing or not Were the Antihypothesis true it would mean that the observed sample would be a random selection from the same population as the control. 26 games selected at random from the 328 game control sample would be expected to average around 12 wins. WIth a 90% probability that the number of wins would be between 8 and 12 The Probability of 26 games chosen at random from the 328 games played over the period containing less than 5 wins is 0.087% (1 in 1155). The antihypothesis can be discarded The hypothesis is proved. I've read that somewhere else,now where was it? Link to comment Share on other sites More sharing options...
Rawrrrrrrr Posted February 15, 2008 Share Posted February 15, 2008 What a pointless post What about the other variables? Location, opposition, team, time, weather, linesmen, light, wind etc etc Generally your stats prove nothing Maybe it was windier on the days we had mcdonald, maybe we were away, maybe we had suspensions or whatever, you get my point theres far too many variables for your post to prove anything Link to comment Share on other sites More sharing options...
Commander Harris Posted February 15, 2008 Share Posted February 15, 2008 What a pointless post What about the other variables? Location, opposition, team, time, weather, linesmen, light, wind etc etc Generally your stats prove nothing Maybe it was windier on the days we had mcdonald, maybe we were away, maybe we had suspensions or whatever, you get my point theres far too many variables for your post to prove anything if you read the post you can see that other variables dont really matter much as he is using the remaining games as the control - 328 games with different conditions - whether that be weather, suspensions or whatever. The fact that the chances of getting a similar set of results from that sample is less than one/tenth of a percent shows that in all likelyhood McDonald affects our results in a negative way. It is possible that a freak set of other factors came up on the days McDonald refereed us, but the chances of that happening are less than 1/10 of a %. Link to comment Share on other sites More sharing options...
Deevers Posted February 15, 2008 Share Posted February 15, 2008 What a pointless post What about the other variables? Location, opposition, team, time, weather, linesmen, light, wind etc etc Generally your stats prove nothing Maybe it was windier on the days we had McDonald, maybe we were away, maybe we had suspensions or whatever, you get my point there's far too many variables for your post to prove anything Take it you were not at the games that "Mr" McDonald officiated in involving us? If you had have been you would know wind, suspensions variables etc etc etc have nothing to do with his miserable, biased, thoroughly disgraceful performances. The stats speak for themselves. Link to comment Share on other sites More sharing options...
simplesimon Posted February 15, 2008 Share Posted February 15, 2008 Actually, as someone with a statistic degree, I'd have to say the sample is large enough to show he is affecting our results. Whether our players paranoia about him, particularly during the Levien years, is a bigger factor than his alleged cheating isn't proven. Link to comment Share on other sites More sharing options...
cylawny Posted February 15, 2008 Share Posted February 15, 2008 What a pointless post What about the other variables? Location, opposition, team, time, weather, linesmen, light, wind etc etc Generally your stats prove nothing [B]Maybe it was windier [/b]on the days we had mcdonald, maybe we were away, maybe we had suspensions or whatever, you get my point theres far too many variables for your post to prove anything Prancer you seem to a bit out of your depth here. p.s. Are you Graham Rix ? Link to comment Share on other sites More sharing options...
bishop1874 Posted February 15, 2008 Share Posted February 15, 2008 Maybe it was windier on the days we had mcdonald, yes thats what it is Link to comment Share on other sites More sharing options...
Dirk Diggler Posted February 15, 2008 Share Posted February 15, 2008 Are you Graham Rix ? Link to comment Share on other sites More sharing options...
Rawrrrrrrr Posted February 15, 2008 Share Posted February 15, 2008 Sorry but I still stick by the fact there are too many variables for those stats to count as evidence in what would be a potentially libelous article if not backed by evidence Whislt the stats are damning do they prove it enough to write a libelous article? No Additionally what about other teams? maybe he has the same in other games in which case it shows general incompetence as opposed to cheating The fact is whilst I agree to a degree, if I was a journalist I wouldnt print such and I doubt any editor would as they could find themselves deep in the doo doo Link to comment Share on other sites More sharing options...
Commander Harris Posted February 15, 2008 Share Posted February 15, 2008 Sorry but I still stick by the fact there are too many variables for those stats to count as evidence in what would be a potentially libelous article if not backed by evidence Whislt the stats are damning do they prove it enough to write a libelous article? No Additionally what about other teams? maybe he has the same in other games in which case it shows general incompetence as opposed to cheating The fact is whilst I agree to a degree, if I was a journalist I wouldnt print such and I doubt any editor would as they could find themselves deep in the doo doo a journalist could print the stats - they are facts and are indisputable. They couldn't say from them that McDonald was bent but they could safely say that having McDonald as our referee affects us negatively - for whatever reason. Nothing defamatory about that. As for your "variables" point, the method being employed here is a statistical one based on a control sample, this control sample is of a reasonable size,328 games, and would represent a variety of conditions incorporating those variables. For that reason it is valid to draw these conclusions. Link to comment Share on other sites More sharing options...
The Merse Posted February 15, 2008 Share Posted February 15, 2008 Excellent OP. And Commander Harris, I really don't see the point in trying to defend it. Someone says black, prancer says white. Link to comment Share on other sites More sharing options...
colinmaroon Posted February 15, 2008 Share Posted February 15, 2008 Sorry but I still stick by the fact there are too many variables for those stats to count as evidence in what would be a potentially libelous article if not backed by evidence Whislt the stats are damning do they prove it enough to write a libelous article? No Additionally what about other teams? maybe he has the same in other games in which case it shows general incompetence as opposed to cheating The fact is whilst I agree to a degree, if I was a journalist I wouldnt print such and I doubt any editor would as they could find themselves deep in the doo doo You will be able to knock me over with a feather if you did the same analysis of the same sized sample for Rankers and Smeltic, if they don't work out just about the opposite!!! But then that's probably down to the weather!!!!!!!!!!!!!!!!!!!!!!!!!!! PS The papers have never worried about saying something controversial to sell their sorry wares - they won't want to upset their neanderthal public, however!!! ................................ Link to comment Share on other sites More sharing options...
Dirk Diggler Posted February 15, 2008 Share Posted February 15, 2008 Excellent OP. And Commander Harris, I really don't see the point in trying to defend it. Someone says black, prancer says white. Topcat is due all the credit mate. It was he who posted it earlier. Link to comment Share on other sites More sharing options...
schillaci Posted February 15, 2008 Share Posted February 15, 2008 a journalist could print the stats - they are facts and are indisputable. They couldn't say from them that McDonald was bent but they could safely say that having McDonald as our referee affects us negatively - for whatever reason. Nothing defamatory about that. As for your "variables" point, the method being employed here is a statistical one based on a control sample, this control sample is of a reasonable size,328 games, and would represent a variety of conditions incorporating those variables. For that reason it is valid to draw these conclusions. Unfortunately, you are wrong. If I were referee-ing this for a scientific journal I would make the following criticisms: (1) This is a non-randomised sample ie the ref is selected by somebody and not randomly allocated. (Of course, that in itself may be the issue). (2) There is no blinding ie the players and MacDonald know who each other is (obviously). (3) The two groups need to be examined for any other contributory factors eg differences in proportions of home/away, playing against OF, other officials, league or cup matches etc. (4) An association between the variable and the outcome may have been demonstrated but this is not necessarily causative and the author's conclusions are not supported by the data presented. However, a carefully designed study with prospectively gathered data and appropriate multi-variate regressional analysis may demonstrate the stench of partisan officialdom. I don't need stats to know that tomorrow.... -Miko will not win a penalty. -We will have more players booked than Celtic. -We will have somebody booked for simulation. -In short he will do us again. Link to comment Share on other sites More sharing options...
The Merse Posted February 15, 2008 Share Posted February 15, 2008 Topcat is due all the credit mate. It was he who posted it earlier. Ah, makes sense. JKB's very own "statto". Link to comment Share on other sites More sharing options...
Commander Harris Posted February 15, 2008 Share Posted February 15, 2008 Unfortunately, you are wrong. If I were referee-ing this for a scientific journal I would make the following criticisms: (1) This is a non-randomised sample ie the ref is selected by somebody and not randomly allocated. (Of course, that in itself may be the issue). (2) There is no blinding ie the players and MacDonald know who each other is (obviously). (3) The two groups need to be examined for any other contributory factors eg differences in proportions of home/away, playing against OF, other officials, league or cup matches etc. (4) An association between the variable and the outcome my have been demonstrated but this is not necessarily causative and the authors conclusions are not supported by the data presented. However, a carefully designed study with prospectively gathered data and appropriate multi-variate regressional analysis may demonstrate the stench of partisan officialdom. I don't need stats to know that tomorrow.... -Miko will not win a penalty. -We will have more players booked than Celtic. -We will have somebody booked for simulation. -In short he will do us again. re point 4, I completely agree. All I have argued is that there is a relationship. Why this is the case is a different matter. I believe this encapsulates points 1 and 2. point 3 is perhaps worth exploring. I'll let topcat defend his own work as he is more than capable of doing so. Link to comment Share on other sites More sharing options...
AlanB Posted February 15, 2008 Share Posted February 15, 2008 -Miko will not win a penalty. -We will have more players booked than Celtic. -We will have somebody booked for simulation. -In short he will do us again. -We will finish the game with less than 11 men. -Sellick will get at least 1 penalty. -The sellick defence will be allowed to assault any hearts player at any time. Link to comment Share on other sites More sharing options...
nanananananana-angus Posted February 15, 2008 Share Posted February 15, 2008 Unfortunately, you are wrong. If I were referee-ing this for a scientific journal I would make the following criticisms: (1) This is a non-randomised sample ie the ref is selected by somebody and not randomly allocated. (Of course, that in itself may be the issue). (2) There is no blinding ie the players and MacDonald know who each other is (obviously). (3) The two groups need to be examined for any other contributory factors eg differences in proportions of home/away, playing against OF, other officials, league or cup matches etc. (4) An association between the variable and the outcome my have been demonstrated but this is not necessarily causative and the authors conclusions are not supported by the data presented. However, a carefully designed study with prospectively gathered data and appropriate multi-variate regressional analysis may demonstrate the stench of partisan officialdom. I don't need stats to know that tomorrow.... -Miko will not win a penalty. -We will have more players booked than Celtic. -We will have somebody booked for simulation. -In short he will do us again. You will more than likely be correct.Statisticians could argue about the variables all night but the things in bold will happen because of a constant,that constant being D.McDonald is a hobo ****. Link to comment Share on other sites More sharing options...
JamboAl Posted February 15, 2008 Share Posted February 15, 2008 What a pointless post What about the other variables? Location, opposition, team, time, weather, linesmen, light, wind etc etc Generally your stats prove nothing Maybe it was windier on the days we had mcdonald, maybe we were away, maybe we had suspensions or whatever, you get my point theres far too many variables for your post to prove anything Forget the variables: there is one constant! Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.