Was Donald Bradman statistically the most dominant athlete?

The Game Room
1

See Don Bradman - Wikipedia

The section appears to be built around the work of statistician Charles Davis, who looked at the performance of athletes in terms of standard deviations above the mean. Bradman comes out on top based on batting average.



Athlete	        Sport	                Statistic	                Standard deviations
Bradman	        Cricket	                Batting average	4.4
Pelé	                Association football	Goals per game	3.7
Ty Cobb	        Baseball	                Batting average	3.6
Jack Nicklaus	Golf	                        Major titles       	3.5
Michael Jordan	Basketball	        Points per game	3.4

I can think of some possible holes-- maybe Babe Ruth should be measured on home runs per season, or cricket might be more variable.

2

Sorry about the bad formatting on the table.

3

No, cricket is not more variable, as you can see from the graph that I created to illustrate another thread. It’s the test match batting average. (Test matches being international cricket games that are scheduled to last 5 days.) Bradman’s average was just short of 100; the next three best are all between 60 and 61. It’s especially impressive because his career was broken by World War 2. (And that illustrates why you shouldn’t use measures like runs per season, because the length of seasons can vary, while games generally stay the same size.)

4

One trouble with this measure if I understand it is changes in the game. Think of Babe Ruth on a home runs per game (or at bat) basis. Home runs were quite rare in baseball until the Ruth era. Now they are much more common. If you look at the standard deviation across all baseball players with a certain minimum number of at bats of the statistic home runs per at bat I’d guess it will be high because of this time variation. If however you compared Ruth to his contemporaries you’d see he stood out very much more. For example in 1920, Ruth’s first season as a full time outfielder (he was an excellent pitcher before that), he hit 54 home runs. That was more than every other team in baseball except for the Philadelphia Phillies who played in a hitter friendly stadium.

5

I’d like to see where Gretzky would come in…he seems notably absent from this list. Goals or Assists (or total points?).

6

Here is a writeup of this “stat” wrt Ice Hockey - it puts Gretzky in the same realm as Bradman.

http://pappubahry.blogspot.com/2008/01/bradman-v-gretzky-v-orr.html

7

I think the fact that Bradman was judged by a career statistic rather than a season is significant. And I think a good case is made. Giles mentioned the list of record holders:

Donald Bradman - 99.94
Graeme Pollock - 60.97
George Headley - 60.83
Herbert Sutcliffe - 60.73
Eddie Paynter - 59.23
Ken Barrington - 58.67
Everton Weekes - 58.61
Wally Hammond - 58.45
Garfield Sobers - 57.78
Jack Hobbs - 56.94
Clyde Walcott - 56.68
Len Hutton - 56.67

Notice how there’s a slow rise from twelfth to second place - 56.67 to 60.97. And then first place leaps ahead by almost forty points.

8

I’d prefer an actual analysis rather than “eh, that looks about right” especially since it sounds like the author started the study with a bias in favor of the hockey players.

9

Well, here’s the top 20 for NHL Points Per Game (career):


1. 	Wayne Gretzky* 	1.921
2. 	Mario Lemieux* 	1.883
3. 	Mike Bossy* 	1.497
4. 	Sidney Crosby 	1.411
5. 	Bobby Orr* 	1.393
6. 	Marcel Dionne* 	1.314
7. 	Peter Stastny* 	1.268
8. 	Peter Forsberg 	1.250
9. 	Kent Nilsson 	1.240
10. 	Phil Esposito* 	1.240
11. 	Evgeni Malkin 	1.233
12. 	Jaromir Jagr 	1.220
13. 	Alex Ovechkin 	1.213
14. 	Guy Lafleur* 	1.202
15. 	Joe Sakic* 	1.191
16. 	Dale Hawerchuk* 	1.186
17. 	Pat LaFontaine* 	1.171
18. 	Steve Yzerman* 	1.159
19. 	Eric Lindros 	1.138
20. 	Bernie Federko* 	1.130

Similarly smooth until you hit the top 2. Now I’m not sure if the presence of one other similar player is enough to ruin the “most dominant” title. I’d say that both Lemiux and Gretzky were similarly dominant.

10

Thanks for the links, Jas09!

11

I still think Bradman gets the nod.

Here’s one way of looking at the numbers. Take the top ten in both sports and compare the range of the top ten to the range of the ten people below first place. For hockey, we’ll do it twice - once with Lemieux and once without him.

For cricket, you get a range of 43.00 points between 1 and 10 and a 4.29 point range between 2 and 11. So the drop-off factor for first place is 10.02.

In hockey, you get a range of 0.681 between 1 and 10 and a range of 0.650 between 2 and 11. A drop-off factor of only 1.048.

If you count Gretzky and Lemieux as a unit and calculate the drop off between 1 and 10 and 3 and 12, you have a range of 0.277 and a drop off factor of 2.458.

12

One thing to remember about that list is that Orr was a defenseman. The next defenseman on the list is Paul Coffey at 28th with 1.087. Of course that’s not how we should be rating defensemen, and therefore not a rating for best hockey player. So we can’t really talk about the best hockey player of all time (even excluding goalies which is a much different position) using scoring. Orr’s plus minus for his career was 597 (0.91 per game), Gretzky’s total was only 518 (0.37 per game). Larry Robinson was the leader with 730 (0.53 per game). I believe Robins is second in career plus-minus per game.

By this standard Orr is far and away the best non-goalie to play in the NHL, and there’s probably a good argument for his being the most-dominant player in any single sport. I can’t find right now a list of career plus-minus per game to make the exact comparison.

13

I don’t understand any of those stats but purely looking at the difference between best and next best do they approach they magnitude of Bradman’s deviation? It doesn’t look like it.

14

The point that often gets discounted in relation to Bradman iss the phenomenal outperformance over a 20 year international career, including the period through WWII when, based on his performance in 1st class cricket, his skills didn’t wane though his health did.

For the sake of the exercise I ran the following comparison.

I took (from ESPN Statsguru) all the world’s Test batsmen whose average was above 50 and who also played more than the 52 matches that Bradman played. There are only 28 of them, plus Bradman.

Then for each of them I identified the 52 matches of their career with their highest aggregate runs. From that I calculated their batting average from the best 52 matches of their career.

This of course hands the modern players a phenomenal advantage as they have played so many more games. Even allowing for that, of the 28, only 6 of them have a higher “52 match elective” batting average.
J.Kallis @ 122.7, S.Tendulkar @ 117.9, S.Waugh @ 117.6, R.Dravid @ 111.9, R.Ponting @ 110.5 and S.Chanderpaul @ 100.6.

But all of these only get their average above Bradman’s career average by excluding over 2/3rds of all the Test matches they played.

Bradman’s entire career was over 50% better than the best’s best.

DG Bradman (Aus)
Matches 52
Career Average 99.94
Best 52 Average 99.94
% Matches 100%

H Sutcliffe (Eng)
Matches 54
Career Average 60.7
Best 52 Average 61.6
% Matches 96%

KF Barrington (Eng)
Matches 82
Career Average 58.7
Best 52 Average 84.8
% Matches 63%

WR Hammond (Eng)
Matches 85
Career Average 58.5
Best 52 Average 83.6
% Matches 61%

GS Sobers (WI)
Matches 93
Career Average 57.8
Best 52 Average 87.9
% Matches 56%

JB Hobbs (Eng)
Matches 61
Career Average 56.9
Best 52 Average 62.9
% Matches 85%

L Hutton (Eng)
Matches 79
Career Average 56.7
Best 52 Average 78.6
% Matches 66%

JH Kallis (ICC/SA)
Matches 162
Career Average 56.1
Best 52 Average 122.7
% Matches 32%

KC Sangakkara (SL)
Matches 115
Career Average 55.8
Best 52 Average 96.5
% Matches 45%

SR Tendulkar (India)
Matches 196
Career Average 54.3
Best 52 Average 117.9
% Matches 27%

GS Chappell (Aus)
Matches 87
Career Average 53.9
Best 52 Average 83.4
% Matches 60%

MJ Clarke (Aus)
Matches 91
Career Average 52.97
Best 52 Average 80.2
% Matches 57%

BC Lara (ICC/WI)
Matches 131
Career Average 52.88
Best 52 Average 97.2
% Matches 40%

Javed Miandad (Pak)
Matches 124
Career Average 52.57
Best 52 Average 94.5
% Matches 42%

R Dravid (ICC/India)
Matches 164
Career Average 52.31
Best 52 Average 111.8591549
% Matches 32%

Mohammad Yousuf (Pak)
Matches 90
Career Average 52.29
Best 52 Average 79.0
% Matches 58%

HM Amla (SA)
Matches 70
Career Average 52.11
Best 52 Average 69.9
% Matches 74%

RT Ponting (Aus)
Matches 168
Career Average 51.85
Best 52 Average 110.4
% Matches 31%

S Chanderpaul (WI)
Matches 146
Career Average 51.67
Best 52 Average 100.5
% Matches 36%

A Flower (Zim)
Matches 63
Career Average 51.54
Best 52 Average 62.9
% Matches 83%

MEK Hussey (Aus)
Matches 79
Career Average 51.52
Best 52 Average 71.6
% Matches 66%

SM Gavaskar (India)
Matches 125
Career Average 51.12
Best 52 Average 96.4
% Matches 42%

SR Waugh (Aus)
Matches 168
Career Average 51.06
Best 52 Average 117.6
% Matches 31%

Younis Khan (Pak)
Matches 82
Career Average 50.74
Best 52 Average 72.68
% Matches 63%

ML Hayden (Aus)
Matches 103
Career Average 50.73
Best 52 Average 84.0
% Matches 50%

AR Border (Aus)
Matches 156
Career Average 50.56
Best 52 Average 98.2
% Matches 33%

AB de Villiers (SA)
Matches 85
Career Average 50.5
Best 52 Average 75.3
% Matches 61%

IVA Richards (WI)
Matches 121
Career Average 50.23
Best 52 Average 83.4
% Matches 43%

DCS Compton (Eng)
Matches 78
Career Average 50.06
Best 52 Average 66.7
% Matches 67%

Yeah, but.

We are comparing the world’s best as measured by their performance against the worlds best. Not some domestic inter-club competition.

15

As it happens, many of the world’s best ice hockey players play in the NHL, and did in Orr’s day. You may also be overstating how international Bradman’s career was. The Wikipedia article on Bradman has him playing against four different countries; the great majority were against England (37 of 52).

Incidentally, Orr was named most valuable player in the 1976 Canada Cup, the predecessor of the Hockey World Cup. The USSR, USA, Sweden, Finland, and Czechoslovakia were regular competitors in the event.

In any event, simply because a sport is played internationally does not make it the pinnacle of athletic achievement. Polo has surely been played in more countries for more years than gridiron, but I don’t think a reasonable polo fan would argue that the athleticism and talent of the best polo club in the world would therefore be greater than that of the best NFL team.

16

Sir Donald played against 4 teams in “Test” or international matches. Then as now those were the pinnacle of the game. Unlike now, the best level, first class was also very high and his performance ther were also spectacular. His 37 matches against England were incidentally against the best team in the world.

17

ahem Runs if you please.

Just to show Bradmans dominance. In test cricket, it’s a players performance over a series (a group of between 2 to 6 tests between two sides) is what is usually discussed.

In Bradmans career, his lowest series average was…56. In a series where he missed one match through illness and where the English employed a tactic of aiming for his head.

18

Walter Lindrum anyone? Might be in with a shout - was so good at billiards they had to change the rules as he killed it as a spectator sport, regularly racking up breaks in the 1000s and destroyed the competition so thoroughly that there was no effectively competition. Some of his records still stand nearly 80 years after he set them.

Weirdly, he was an Australian competing at the same time as Bradman.

19

Hockey is a glaring omission from Davis’s list. I’m willing to chalk it up to his nationality, since Australia has never had any major sort of hockey presence. However, it is a major international sport, and should be considered. Hockey’s omission means that a similar dominating performance on Gretzky’s part is ignored.

The following only looks at players with at least 100 games played in order to avoid outliers with only a handful of games played. Gretzky’s 1.92 PPG puts him 6.39 standard deviations above the mean value for skaters (non-goalies). He’s an incredible 9.91 standard deviations above the mean for career points for the same population. It’s worth bearing in mind that, despite retiring 5 years before his teammate Mark Messier (#2 in career points), he is 970 points ahead of Messier (only 83 players in league history have even obtained 970 points).

20

Funny you should say that. Australia actually has one of the world’s best Hockey teams. They’re Olympic and two time world champions. They consistently place in the top three teams in both competitions, and have done so from the 1960’s. On further reading, it appears you’re talking about a game that is derived from Hockey, played in fewer countries and by fewer people, called Ice Hockey. Please do try and be clearer.