Throughout 2016, some 47 Test matches were played around the world. Of them, only seven resulted in draws and three of those were so severely weather affected no result was even possible.
Of the other two draws, the match between West Indies and India at Kingston was badly affected by rain (the equivalent of a day and a bit was lost), and so too the Lord’s match between England and Sri Lanka (even more play lost to rain).
So that leaves two draws which weren’t weather affected, and they both involved England: the essentially dull match in Cape Town when both teams hit more than 600 on a spectacularly flat wicket in the first innings, and the one in Rajkot where a bolder declaration from Alastair Cook might have put India under severe pressure. As it was, they still lost their sixth wicket with more than half an hour to go on what was generally considered an unusually flat wicket, even by sub-continent standards.
There were certainly several instances of teams succumbing to pressure late in the day when a draw seemed almost assured. England in both Dhaka and Chennai last year, and Pakistan in Melbourne are examples that quickly spring to mind.
This lack of draws in 2016 is a fairly remarkable development, even allowing for the more expansive tactics that modern cricket has adopted. Between 2000 and 2015, just under a quarter of Tests (24.4% to be precise) finished as draws. In 2016, it was less than 15%.
The first Test of 2017 has ended in such a thumping win for South Africa that it would have been wrapped up well inside three days had Faf du Plessis chosen to enforce the follow-on. Sri Lanka effectively lost the match on the second day. First, they allowed South Africa to convert a platform of 297-6 to an all-out score of 392. Then they collapsed from 56-1 to 110 all out, losing nine wickets in less than 20 overs.
There was no excuse for this. All our ball-tracking data suggests fairly modest amounts of swing and seam movement throughout the match, and particularly over the course of the opening two days. There was also largely predictable bounce. What there appeared to be, visually at least (this is an aspect that cannot yet be calculated scientifically) was a reasonable amount of pace in the wicket.
Batsmen appeared to be hurried at times, particular when facing the quicker bowlers such as Kagiso Rabada and Lahiru Kumara. But there was none of the really extravagant, unplayable movement that many of the pre-match pundits had forecasted.
Perhaps the best concerted spell of bowling in the whole match came when Sri Lanka began their first innings. A very high percentage of balls were edged or missed by Dimuth Karunaratne and Kaushal Silva, but somehow they got through the first 15 overs. Then, Silva was unhinged by a short-pitched ball from Rabada (Sri Lanka were particularly bad at leaving the short stuff) and could only guide it onto his stumps. But the ball was getting softer and Sri Lanka got through more overs. Less was happening… until the batsmen started losing their heads.
In the blink of an eye, Karunaratne, Kusal Mendis, Angelo Mathews, Dhananjaya de Silva and Dinesh Chandimal were sent packing. And the daft thing is, it wasn’t a glut of particularly good balls that did for them. Mendis went slogging the spinner and Karunaratne also brought about his own downfall, slapping a short, wide one to point. Mathews was at least playing defensively when Rabada had him caught at the slips – but the ball wasn’t threatening the stumps. De Silva was skipping down the track when trapped lbw, and Chandimal’s wild, flat-footed nicked drive was one of the worst of the lot.
That was essentially game over, the specialist batsmen all gone – and the tail unable to salvage anything from the wreckage. Whatever happened from that point – and there were four sessions of cricket to follow – the only viable result was a heavy Sri Lankan defeat.
Let’s look at weighted wicket probability (WWP) for the five key Sri Lankan dismissals – the second to sixth wickets in the first innings, the ones that turned hope into despair. Now you may (or may not) have read in previous blogs how WWP works. It essentially uses the tracked characteristics of each delivery to assess its wicket-taking danger – a unique and (we think) rather clever analytical tool. The final number in the grid below is the percentage of times we would expect each delivery to take a wicket.
And here we are:
- Mendis (slogging spinner) – 0.01 = 1%
- Karunaratne (cut to point) – 0.013 =1.3%
- Mathews (defensive prod edged) – 0.023 = 2.3%
- De Silva (lbw down the track) – 0.016 = 1.6%
- Chandimal (flat-footed drive) – 0.013 = 1.3%
So, rather appropriately, the highest of these five deliveries on WWP is the one which saw off Mathews – and he was the only one of these five batsmen not to be playing an attacking shot for his dismissal.
When we rated James Anderson’s swing-bowling masterclass at Headingley (also against Sri Lanka), we found his average WWP was 2.13%. The average of these five dismissals in question is just 1.5%, a significant amount less.
To go back to where this blog started off, Sri Lanka should not be castigated for failing to draw the Test as such. It’s possible this match was never meant to be drawn. But they certainly should have been able to bat for longer than 43 overs in the first innings.
Don’t be surprised if 2017 brings us another famine on the draw front.