Thursday, March 03, 2011

Two LBW Decisions III: An obvious error with DRS?

Update: I've added to my argument as to why Watson's LBW should have been harder to overturn than Bell's. See the third paragraph after the second screenshot

These are the hawkeye screenshots of the two decisions. The first is of Shane Watson's LBW decision. The second is of Ian Bell's LBW decision. In both instances, the original decision was Not Out. In Watson's case, he was given out upon review, while in Bell's case he was given Not Out. This presents a fascinating case in the story of DRS and Hawkeye. It also helps to unravel the agencies of the various entitites involved in the process. I remain skeptical of the capacity of the DRS method to produce good decisions, although I thank Jonathan for his thoughtful comments on my two previous posts about this episode. I think he's generally right that the ICC DRS code allows for far greater discretion that it is usually given credit for (on this blog as well).

The questions I've had all along about the DRS remain, even though many of them have been answered in part or at least refined over the last 12 months or so - Why DRS? Why involve players? How does the DRS determine what is marginal and what is obviously wrong (or right)? Where are judgments made in the DRS? What is the point of the public spectacle - the effort to lay bare the decision making process in public? How much of this is actually possible? How is evidence produced from simulations (video replays, hawkeye, hotspot, super slo-mo) used?

A comparison between the Watson and Bell LBWs, made 10 days apart in the same tournament, in which the same technology and the same broadcaster was at work suggested on the surface, that one of the two decisions must be wrong. In Watson's case, Umpire Kettleborough changed his mind, even though, if everything went well, the Third Umpire would have had no option but to point out that the evidence from the technology was inconclusive, while in Bell's case, Umpire Bowden didn't.

A strict reading of the Hawkeye simulation, would suggest that Bell's decision was easier to reverse than Watson's. This is because of the following three reasons:
1. The impact was shown to be middle-and-off in both cases, in Bell's case it was lower down, and the ball was heading towards middle (towards the stumps), while in Watson's case it was higher and heading towards the off-stump (away from the stumps).
2. The distance of the pitching point and the point of impact was greater in Bell's case than in Watson's case. This matters according to Hawkeye's criteria for evaluating the quality of the prediction (the greater the distance, the better).
3. Bell is shorter than Watson, and was also trying to play across his pad, hence it is arguable that his stride must have been just that little bit shorter than Watson's (Watson was playing straight).

Hence, the chance of the Bell delivery missing the stumps is arguably lower than that of the Watson delivery missing the stumps. It would be instructive to see what the computed margin of error actually was in each case.

So why did the decision that should have been less likely to be reversed, get reversed, while the other one didn't? Here are three possible reasons:

1. Umpire Tucker and Umpire Saheba, the two third Umpires interpreted the TV Replay differently, and applied different levels of emphasis on the 2.5 m aspect of the situation.
2. The match situation was different in each case. In Bell's case the game was in the balance, while in Watson's case Australia were decisively ahead in the game and this was unlikely to change anything.
3. Umpire Kettleborough had a specific doubt, which he cleared with Umpire Saheba, whereas Umpire Bowden was more generally skeptical of the decision.

In any event, what we have here is a classic marginal situation. In one case a marginal decision was overturned, in the other, it was not. The use of the DRS was entirely futile. Should it have gotten involved in at all? And once it did, did it cause an obvious error to be made in at least one of the cases, because it couldn't distinguish usefully between the marginal decision and the obvious error?

No comments:

Post a Comment