Champions League – Benaud’s Law of Supremacy

When Richie Benaud passed away earlier this year, I was moved to re-read the publications of his which reside in my cricket library, one of which, Over but Not Out, includes this phrase:- A champion is generally regarded as a cut above the rest of the field. He was using the phrase in declaring Keith Miller as having been the greatest ever before going on to wonder if that definition could also be stretched to encompass Ken Mackay – apparently while “Nugget” was an undisputed champion, he acknowledged “Slasher” to be also a champion in his own way. This may possibly have been as a result of some guilt on the part of Benaud, after having interefered with his career when captain; Benaud acknowledges that Slasher may have been a better player had he not insisted on turning him from a batsman into an all-rounder “because it suited me to do so”.

On re-reading Benaud’s definition it occurred to me that I had a way to quantify it, using match impact. Match impact is described in more detail here, and the use of win probability as its basis is described here, while in this particular context we can use it to see who is the most dominant player at any given time and unearth Benaud’s champions.

There are two ways of measuring this dominance; one is the degree to which a particular player outshines his teammates, and the second is the degree to which a player outshines everyone in Test cricket at any point in history.

Ploughing a lone furrow

My database of impact profiles allows me to identify those players who, as Monty Python would have it, were like a stream of bat’s piss – that is, they shone like a shaft of gold when all around was dark (one of Shaw’s, I believe).

The current match impact is an average of the impact which the player had achieved in the previous five Tests and, as such, identifies the “hot hand”, so to speak. The form of every player of course exhibits peaks and troughs from time to time so that the hot hand may not necessarily be the best player overall, but as regards those with the highest measured dominance, I would expect that they would most likely be great players.

So why not just use the ICC player ratings?

Good question. The ICC has rated every player going back to the beginning of Test matches (as well as ODIs and T20s) so it’s possible to look at their top 100 batsmen or bowlers at any point in time. There are two problems with using the ICC ratings for this exercise, one pratical and the other more philosophical.

The practical problem is that we don’t have access to the ICC database so we’d have to look at the top 100 batting ratings after each of the 2200+ Tests individually, then repeat the exercise to rate the bowlers. The philosophical problem is that we could still only rate the top batsman and bowler on their batting or bowling alone; though we could then combine the two to see which had the greatest degree of dominance, fielding would of course not be taken into account. Additionally, I don’t know how the ICC rating algorithms operate so I can’t comment on whether or not a margin of 20 in batting is equivalent to the same margin for bowling.

In comparison, match impact produces a single rating for each player which takes into account batting, bowling and fielding, thereby enabling us to compare all players on a level playing field, as it were, which implies a different philosophy than that applied by the ICC ratings.

Benaud’s Second Law

Why second law? Well, I think that Benaud was intending his definition to be applied to champions as compared to all other players, whereas with the first part of this study i want to look at those who have ploughed the lone furrow by being head and shoulders above their peers. If we look at players who, going into a Test match were most on top of their game, we get this list:-

DoD represents degree of dominance. The above table is by and large not too much of a surprise – Lara and Hadlee I would have expected, Flower too for that matter. There are two omissions from the above list, both of which are linked to an artifact of the method. The first is that I have excluded all players from early Tests, as there are many new players with no rating, making it easier for a regular player to dominate; this means we lose Spofforth (21.65%) and Ulyett (22.04%). The second omission is specifically Bradman in 1946/47 – that side boasted the likes of Sid Barnes, Arthur Morris, Lindsey Hassett, Keith Miller, Ian Johnson and Ray Lindwall, which is a spectular line-up to be dominant over, however at that time these players hadn’t yet played in five Tests so had not generated a match impact rating, which as I mentioned in earlier features is the duration over which impact is averaged.

Another way to look at those who flew higher than their peers is to aggregate all of the degrees of dominance for each player – those who rate highest would be identified as being clearly superior:-

Although I don’t show here DoD on a per-Test basis, clearly Barnes would reign supreme in that case.

Benaud’s First Law

I would consider all of the above players to be champions, but that doesn’t represent Benaud’s intended definition – to be a champion under Benaud’s First Law requires that the player dominates all players, not just those on his own team.

In order to show dominance over a reasonable duration, I broke Test cricket history into blocks of 50 Tests and totalled each player’s match impact in each block. The most dominant players are shown below in chronological order, with a threshold of a 30% DoD margin over the next best:-

In some ways this is not a fair measure – the use of 50-Test blocks mean that for example Fred Spofforth doesn’t show; Spofforth was dominant in the first 25 Tests, but Charlie Turner’s dominance in the next 25 Tests means that, over the first 50 Tests taken together Spofforth appears not to have been as dominant. However if 25 Tests was used then dominance in later eras would not be highlighted as clearly. Also, the use of 1-50, 51-100 as a delineation is necessarily arbitrary, as players can be dominant through 50 Tests encompassing Tests 76-125 for example, which measure would be dispersed using this method.

Nonetheless, this has thrown up a number of players who were clearly dominant for a period, such as Botham in 1981 and Flintoff in 2005. However, towering over even these is Garry Sobers who, whilst also showing as the most dominant most often, shows for his most dominant period as having been more than twice as dominant as any other player ever.

Garry Sobers – an absolute colossus of match impact.