Well, after about fifteen years of work, tearing my hair out, and driving myself insane at times, I finally settled on a statistical rating formula and thus I have completed a top 100 list of Test Batsmen that I am happy with.
I will now present my top 100 with a thread devoted to each block of ten players, and then as we get into the top 20, I will probably present 2 players at a time, which will make it similar to the CW Top 50 threads last year.
But, first of all, I would like to share how my ratings system works. At its core is the idea that each batsmen's average is adjusted according to the strengh of opposition he faced and the era he played in. Thus, you can expect the averages of such batsmen as Ricky Ponting and Virender Sehwag to drop, whilst the averages of players in the 1990s (an era full of great fast bowlers) to be adjusted upwards slightly.
I worked out the bowling averages of each team for each season of cricket from 1877 to the present day. I will not go into details of how I did so, but in my opinion, it is fairly accurate without being perfect. I did not build my own database. Instead, I relied on making excel spreadsheets and copying information from cricinfo's statsguru. Rating just one player took over 3 hours, so you can imagine how much time it would take to rate more than 100!
I have attached Stephen Fleming's ratings page so you can get some idea of the type of work I have done.
Anyway, here is how each player is judged:
Length of career
Career Average (adjusted)
Runs per innings (adjusted)
Away average (adjusted)
Top opposition average (adjusted)*
Strike-rate for innings of 50+
Number of centuries scored against top opposition
Number of double centures scored against top opposition
Number of significant innings against top oppotition**
Number of great innings***
Significant innings per match
Centuries per innings
Best series in terms of runs scored
Innings worth average****
25 Test peak composising of peak length (days played), adjusted average, adjusted runs per innings, centuries scored and significant innings
50 Test peak composing of peak length (days played), adjusted average, adjusted runs per innings, centuries scored and significant innings
* The top opposition has three levels, from the all time great teams, to very good teams, to middle of the road teams. Runs scored against the top tier teams have more weight. Minnows and poor teams are excluded.
** Posters may remember my top 100 test innings list. Brian Lara topped it with an innings worth around 25 points. A significant innings is an innings worth 3.50 and over in matches won, close draws, fighting draws and close losses. Thus, runs scored in bore draws or heavy losses don't count.
***A great innings is one worth 12.50 points or more.
**** The average points for all innings played, except in team scores under 150 with less than 5 wickets lost.
Finally, the qualification for entry is:
10 matches played AND a career of 1000 days or more between the player's first and last test match
The ratings of batsmen who played less than 50 test matches are adjusted downwards according to how many matches they played.
Thus, don't expect to find Martin Donnelly in this list or Vernon Philander in the upcoming bowler's list (not yet at least).
So, there we have it. Over the next two weeks or so, all will be revealed!
One final word: it is somewhat ironic, but the more work I did on this ratings system, the more I came to appreciate that statistics are by no means the only way to judge a player's worth. Rating systems like this should promote healthy discussion and debate without being the final word on a particular player's standing in the game's history.