Vijay.Sharma Posted June 20, 2011 Posted June 20, 2011 Folks, Finally finished the analysis of the 20 ATG batsmen. And here are the quick results for you - Firstly, I'd like to thank ICF, Raghav_12, BossBhai, Akshayxyz, Sachinism, Sree_India, holysmoke, My two cents, and all others whose names I have missed (I know I have missed many names) for helping me with your inputs on this project. i have received inputs from members in other forums as well, and I believe the model is at a stage where it can be taken seriously as a partial way to compare batsmen across eras. There are 2 batsmen who have performed way above all the other ATGs and these two must be considered as Virtuoso Batsmen: Donald Bradman and Sachin Tendulkar One batsman's performance is way below the other ATGs: Wally Hammond Ten batsmen fall in the typical ATG range of 0.90 and 1.10 Seventeen batsmen fall in the range of 0.84 and 1.12 The following figure captures the Virtuosity Index The following figure shows the Parameter Indexes I have also attached the entire project "paper" with this message. Alternatively, you can access it here - Text Cricket's Virtuoso Batsmen Project version 2.0 eagerly awaiting your thoughts on the project and the various inferences (pls have a look at the attached doc or linked doc)... Edit - Abbreviations expanded: VI = Virtuosity Index CPI = Career Performance Index TOI = Top Opposition Index CDI = Condition Difficulty Index BCI = Bowler Challenge Index PCI = Peer Comparison Index CTI = Contribution to Team Index
Cricket Tragic Posted June 20, 2011 Posted June 20, 2011 That is really detailed stuff! I'll have a look through it and let you know what I think. :two_thumbs_up:
Tendya Posted June 20, 2011 Posted June 20, 2011 I read the entire analysis, it has some of the most complete cricket analysis done, You have considered every possible factor imaginable to decide an ATG in a statistical way. I would like to add just one point, There was a downward bias when it comes to the older batsmen, Hobbs, Bradman, Hammond etc. Other than that this is just perfect :hatsoff:
EnterTheVoid Posted June 21, 2011 Posted June 21, 2011 Nice one Nirvarnam. Now sell it to Cricinfo and use that money to follow the Indian team till 2050.
Sachinism Posted June 21, 2011 Posted June 21, 2011 What input did I give? I can't even make sense of the formulae :((
Raghav_12 Posted June 21, 2011 Posted June 21, 2011 Thanks for acknowledgement of my help, though I don't exactly remember how did I help :hmmm: . Will go through this stuff and will give my feedback, if any.
kabira Posted June 21, 2011 Posted June 21, 2011 I think ICF should give Vijay his own subdomain and he can host the project here.
mishra Posted June 21, 2011 Posted June 21, 2011 Vow. :clap: :clap: :clap:.. What a detailed analysis. I have no complains with the analysis. ..I know its not possible to include all the factors or make everyone happy.... My two cents.... I think PCI and CPI are related..... You may have wiegted it twice for each player. If possible can you intoduce something like general cricket quality index so that factors like fielding, home team pich, expectation, crowd support,professionalism can be mesured too... But if you dont do so... Its still best comparison i have ever seen :hatsoff: I know Wisden or ESPN or ICC will never do it
mishra Posted June 21, 2011 Posted June 21, 2011 I am just a fan. You are cricket historian,mathematecian,writter and expert... :hatsoff: again
zen Posted June 21, 2011 Posted June 21, 2011 So much analysis to find out that DBG is the greatest! .... From the graph, it appears as if Dravid is rated higher than Richards. In which world other than of stats can you have that :P
Roshanrocks Posted June 21, 2011 Posted June 21, 2011 why is my name not included in the list? :hitler:
mishra Posted June 21, 2011 Posted June 21, 2011 The existence of an unknown ninth planet was first proposed by Percival Lowell, who theorized that wobbles in the orbits of Uranus and Neptune were caused by the gravitational pull of an unknown planetary body ---- Morale: Mathematical apprpoach is allways correct, at times Results are surprising
The Outsider Posted June 21, 2011 Posted June 21, 2011 Inzamam is greater than Chappell and Barrington is greater than Richards - wonderful analysis. :nice:
dial_100 Posted June 21, 2011 Posted June 21, 2011 I think ICF should give Vijay his own subdomain and he can host the project here. I agree. If as a leading cricket forum, we could contribute towards establishing a solid process of determining cricket greats then it will help entire world. It will be nice if all could contribute in fine tuning the process. Constructive criticism will help rather than some sarcasm. Having said that I disagree with few outcomes but until I understand the whole process, I will hold on till I understand that whole document thoroughly. :two_thumbs_up: Vijay. Let us know which areas you dont feel confident that the measures could be 100% correct. What % of error margin do you consider in this whole exercise?
Ultimate_Game Posted June 21, 2011 Posted June 21, 2011 Again reinforces the fact that Bradman and Tendulkar are in a league of their own.
Vijay.Sharma Posted June 21, 2011 Author Posted June 21, 2011 Thanks a lot guys! RoRo, I am sorry I missed you and some other guys out in the OP. But I covered you guys in a way by saying "ICF"..lol! However, if you are one of those guys wondering what you helped with...lemme put it this way - they were not always explicitly mentioned but gave directions to the thought process. And some of those items were things that I didn't think of and so the name and the input clicked a photo in the mind which stayed their clearly...lol! Yeah, it would be great to have a sub-space on ICF to publish this model and everything else in it. A post on a site like this provides more visibility than posting it on my personal blog. And also it can help in creating a common workspace for whoever may be interested in joining in for the analysis or even if they just wanna provide some inputs. I also would like to start writing some articles and post them (be prepared for the pomposity in my expression). Dial_100, as a fan I myself do not agree with some of the results that came out. I actually mention this at the end that when I look at it from my fan's bias I find some results as "correct" and some as "wrong". But well the data shows what it shows. One of the items already on the cards is got to do with a continuous scale for Minnow Weight. Right now it is 0.5 : 1 :: Minnow : regular. Instead of using a cut-off date to say from this day the team became a Regular team, if we can make the weight a continuous thing such that before that date the team could be given a weight of 0.97 or 0.91 or 0.86...0.61...0.5. Basically 0.5 <= Minnow <= 0.99. Any thoughts on this? Mishragreat, main difference between CPI and PCI is that CPI includes all the subjective criteria (I refer to them as 'contextualization rules' in the doc) i.e. Minnow is treated differently than Regular, Diff innings will have diff weights, Home-Away treated differently, and RFF. And the career figures for the 20 ATGs are computed based on this. Whereas in PCI the normal career figures (what u see in statsguru) is compared with the figures of the particular ATG's peers (i.e figures of top 1 to 6 batsmen in all Tests played between the date of ATG's first Test and his last Test...of course ATG's figures will be subtracted from the Top 1 to 6 batsmen figures). Folks, please don't interpret the results as a ranking of these 20 players. It is not a ranking. It identifies clusters i.e. if the distance between some ATG(s) and other ATG(s) are significant. It is primarily a relativistic metric and involves comparison of performance within the group of 20 ATGs to arrive at their index values. Hence looking at it as a ranking system may not be right always. Because you'll find a Border > Sobers or Sehwag > Chappell does not mean they are ranked that way. Look at the difference between their index values. What you can do though, is pair-wise comparison by considering a gap of 0.15 as being significant. For example, while Inzy 0.919 and Gavaskar 1.083 fall in the 0.9 - 1.10 band, the difference between the two fellows is almost 17% (0.17+) so you can say Gavaskar is greater than Inzy for sure. But if you look at Miandad 0.883 although he is out of the typical ATG range of 0.90 to 1.10, if you compare him only with Inzy, they are similar...the difference is hardly 3-4%. Tendya, where do you think the analysis is biased against the pre-war ATGs? Or do you mean to say cricket itself is biased against them because they didn't play as many Regular teams, as many venues, as many innings, as many top rated bowlers. If the latter is the case then it is, by design and not bias. The main intent of doing this exercise is to provide context to these performances. If the context is reducing their performance indexes then it is a fair call and the model is working well. Mishragreat, the Home pitch thing is already taken care. but I am not sure how we can consider these elements: level of fielding, professionalism. I think these two factors definitely influence a cricketer's numbers but I have no idea how to incorporate them. Things like crowds, expectations, and other such things - I am not sure if they influence player performance at all. And even if they do, the influence maybe different for each ATG which makes it next to impossible to include as a common factor. For example, let's say for Inzy crowd shouting will throw him off his game but for Pawning it will help him. Or Miandad would love to shut up a huge opposition crowd when they are noisy whereas such a situation may affect Hammond...we just don't know about this.
Vijay.Sharma Posted June 21, 2011 Author Posted June 21, 2011 I agree. If as a leading cricket forum, we could contribute towards establishing a solid process of determining cricket greats then it will help entire world. It will be nice if all could contribute in fine tuning the process. Constructive criticism will help rather than some sarcasm. Having said that I disagree with few outcomes but until I understand the whole process, I will hold on till I understand that whole document thoroughly. :two_thumbs_up: Vijay. Let us know which areas you dont feel confident that the measures could be 100% correct. What % of error margin do you consider in this whole exercise? Error margin...I guess the +/- 10% band is a kind of error margin or tolerance band. All the indexes are relativistic so it wouldn't mean that each ATG's index value will have a +/- 10% error margin but that the margin is for all of them together. So maybe at the level of the individual the error margin won't be more than +/- 2 or 3%. By increasing the band of significance we can ensure that we don't prematurely say that some players perform at a different level than others.
mishra Posted June 28, 2011 Posted June 28, 2011 Mishragreat, the Home pitch thing is already taken care. but I am not sure how we can consider these elements: level of fielding, professionalism. I think these two factors definitely influence a cricketer's numbers but I have no idea how to incorporate them. Things like crowds, expectations, and other such things - I am not sure if they influence player performance at all. And even if they do, the influence maybe different for each ATG which makes it next to impossible to include as a common factor. For example, let's say for Inzy crowd shouting will throw him off his game but for Pawning it will help him. Or Miandad would love to shut up a huge opposition crowd when they are noisy whereas such a situation may affect Hammond...we just don't know about this. Its just an idea. Any addition to it is welcome. For example quality of fielding improved dramatically all over the world after South African came back to international cricket To measure quality may be you can set indicatiors. When was ceratin attribute introduced and became prevelant Format: Tests,odi and T20, day and night, floodlight, white ball, redball (when were they introduced) Batting: Defense and attack importance of pacing the inning.reverse sweep, heavy bats Bowling: spin bowling and fast bowling Fast bowling:Example for fast bowling will bodyline, Bouncers, inswing,outswing,reverse swing,yorkers,slow yorker, Sling action, Slow bouncers,Leg cutter,Off cutter Spin Bowling: Googly, Carrom ball,Doosra,Topspinner,Flipper,Off break,Leg break,Arm ball Umpiring decisions.
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now