Jump to content

Recommended Posts

Posted

What should be one's ODI stats, so as to be considered bradman-esque (in ODIs). i.e. Bradman (of 30s-40s in tests) kind of standout performance - in ODIs in modern era (preferably 1990s (or 80s) onwards. Few points that you might find useful that I have gathered from various sources. a) As per some statistical analysis on Don's era - it is widely accepted that he would have averaged about 75 (in tests) in modern era. In other words, 75 can be considered as Bradman-esque test average of modern era. B) If it helps, Sachin (or a combination of others) can be considered a benchmark for consistency across all formats in modern era. c) Any assumptions/(numbers) you are going to add/drop, and think should be considered by others too - please mention, I will update OP. Rules: a) This is a serious post with some serious purpose - that definitely is not to prove greatness of one player over other. It is more of an exercise to explore the limitation/strengths of statistical models (or other approaches) you people would want to use to do the prediction/extrapolation. So please contribute constructively. B) educated guesses, Calculators, google, Cricinfo, matlab ,outsourcing - are all allowed. c) Instead of debating over the purpose of this post, you better just leave a guess - or please just stay away. d) Please refrain from indulging into talks over conditions, bowlers faced, or other differences between eras. There are numerous other discussion threads for those purposes. e) Base your prediction purely on the basis of numbers - averages, 100s, 50s, etc.. and other such cricket stats - relative to peers. f) If it helps you can use stats/math model from other sports - (to study variance/distribution etc.) - be careful that there should not too much of difference in sports dynamics. e.g. It would not be wise to map some performance distribution curve from Tennis to Batting in Cricket.

Posted
:hysterical:
What part do you find funny? If people can predict Don's test average in modern tests, with high level of confidence ( I heard 75) - using some statistics/outlier science, then this should also be equally easy.
×
×
  • Create New...