October 24, 2014

The bias of umpires

Understanding historical trends in decision-making might help you deal with the iffy calls you receive. Or maybe not

Bowler-captains hold the advantage when it comes to getting decisions in their favour © AFP

The cricket season not long past was my first back in the arms of an old lover after a three-year break. The old pleasures I'd sought were still present, if less frequently found: the sublime thonk of a cover drive; the giddy relief of the unspilt dolly; the faux-nonchalant smile when a bowling change brought a key wicket. And so too were the old annoyances, primus inter pares being poor umpiring decisions (those of which I was the victim, at any rate) - decisions made with the emotions, or in attempting to right an earlier wrong, or with undue haste, or with too much relish in the raised finger (do these Koertzens practice this signal?).

While a longstanding and bone-deep respect for the admittedly nebulous spirit of cricket kept even the slightest flicker of dissent from my face, it did make me question a couple of assumptions.

Firstly, I had always thought that captains received the rub of the green with an umpire's decisions, particularly in club cricket, where the latter's marks and advancement depend on the captain's approbation. I hadn't ever really been one to ingratiate myself with the ice-cream men, despite seeing the pragmatic value of doing so, yet I had presumed the "fine-wine effect" of three years away - umpires getting sozzled and forgetting I used to be a fractious so-and-so - might afford me the benefit of the doubt with the iffy calls. Not the case.

Which brings us to the second point: how often are the marginal decisions now going in favour of the bowler as a result of Hawk-Eye technology (an established truism of the professional game)?

Luckily these are exactly the sort of matters explored by Douglas Miller, a man with the wherewithal and inclination to investigate them, which he did in March 2011 in a study published by the Association of Cricket Statisticians and Historians. Miller painstakingly scrutinised all County Championship matches played between World War I and 2010, discovering, among other things, that historically captains had indeed benefited from umpiring leniency. From the start of the 2000s, however, this stopped being the case; or rather, it did of batsmen-captains, who are now without advantage, while it remains true of bowler-captains that they are statistically favoured - probably because the former have to take their foul mood off stage, whereas a spurned bowler-captain can chunter away continuously, chafing at the umpire's resolve.

While the benefit of the doubt has shifted away from batsmen (and batting captains), the umpires chosen from the domestic pool for international honours has moved from the "not-outers" to those slightly more trigger-happy fellows

Another trend highlighted by Miller was the increasing prevalence of lbw decisions as a percentage of all wickets, with spikes caused by changes to the lbw laws in 1936 and 1972, then later under the influence of Hawk-Eye, a trickle down from the international game that has trickled further still into recreational cricket.

Nothing in the least surprising there, you're doubtless thinking, yet this modulation of umpires' subconscious bias has affected how they are perceived by those chosen to select domestic umpires for higher honours. Miller shows that while the benefit of the doubt has shifted away from batsmen (and batting captains), the umpires chosen from the domestic pool for international honours has moved from the "not-outers" to those slightly more trigger-happy fellows - the evidence of technology having borne out their instincts, it would seem.

Miller analyses all umpires to have stood in over 50 Championship games (a total of 164) and - with a degree of statistical licence, given the impossibility of tracing back all decisions to their individual adjudicator - crunches the numbers to find out how much they deviate from the norm over the duration of their career. This index places each umpire in a scale, from the biggest "outer", Harold Elliott, to the biggest "not-outer", Kevin Lyons, with Dickie Bird directly above him, 46% less inclined to give a batsman lbw than his era's Mr Average.

What's interesting here is that of the top 30 umpires in the list (the "outers"), only six were given Test match duties. Of the bottom 30, however, 21 were accorded such honours. The conclusion is that those hitherto perceived as "outers" - the likes of the perennially overlooked Ray Julien, say - may well have been ahead of their time. It's not so much that they gave the benefit of the doubt to batsmen; rather, they reduced what they thought of as just cause for doubt (big stride forward, for example). It could also have shown that most county captains were batsmen, and their annual meeting became a platform for their collective disgruntlement with a perceived lack of protection. Ahem.

(Incidentally, this wasn't the only bias - or bias toward a bias - that Miller examined. A companion study showed that prior to the appointment of neutral umpires in Tests in 1993-94, away batsmen in Asia were far more likely to be fired off lbw than elsewhere: 55% in India, 65% in Pakistan, and 182% in Sri Lanka.)

Miller's study also demonstrated that the deviation above and below the norm was getting ever narrower, perhaps because ECB mentoring and feedback is producing greater decision-making consistency across the umpires panel. Maybe there's no more clamouring to bowl at certain umpires' ends.

A final myth that Miller dispels - perhaps one that is quite surprising - is the notion that it's the ex-bowlers who are the biggest "outers" and the former batsmen who are most reluctant to flex the finger. On the contrary, the index reveals plenty of "poachers turned gamekeepers", which probably says something quite profound about human nature. All of which led me to wonder what I will do if when I take up the white coat. Of course, there will be absolutely no bias whatsoever, but I'll make sure I practise the raising of the executioner's digit, just to convey the required tone.

Scott Oliver tweets here