Also on cricketcountry.com
By Michael Jones
Almost as soon as Pakistan sealed a one-wicket victory over India in the Asia Cup 2014, a list started circling on cricket sites of supposedly amazing similarities between that match and the only previous One-Day International (ODI) between the two teams to end in victory by such a margin — at Sharjah in 1986. This was the match when Javed Miandad famously hit Chetan Sharma’s last ball for six to seal the win and with it the Austral-Asia Cup.
Even ESPNCricinfo posted the list on their Facebook page, with the comment “This will blow your mind”. On the face of it, the parallels do seem to make an incredible coincidence — but are they really as astonishing as they seem? Or are our minds too easily blown by lists compiled in such a way as to make them appear more unlikely than they really are?
Let’s examine some of the similarities listed: how improbable are they? Firstly, both matches were played at a neutral venue. This is hardly remarkable: a significant proportion of ODIs are staged at neutral grounds, which becomes even higher when one of the teams involved does not play home matches due to security concerns (it’s even less of a coincidence that one match was held in the UAE, the other in Bangladesh, more than 2000 miles apart; but of course it would spoil the effect to mention that). In both matches, Pakistan won the toss and chose to field; neither of these events is notable — each has a 50 per cent chance of happening in any given match. If they hadn’t chosen to field, they couldn’t have won both matches by margins of one wicket and no one would be making the comparisons at all.
India scored an identical number of runs in the two matches — the only fact on the list which really is a significant coincidence; given the number of possible different totals in an ODI innings, the chances of the totals in two particular matches being the same are quite low. It seems possible that whoever drew up the list started by noticing the one genuine coincidence and added whatever else they could find in an attempt to make it look more impressive. Again, a closely related statistic which was different [the number of wickets lost] is ignored.
Krishnamachari Srikkanth and Rohit Sharma, who both opened in the matches in question respectively, hit the same number of sixes, as did Miandad and Shahid Afridi, who both made the winning hit. The identical numbers of sixes are just a matter of picking out the numbers which match and ignoring the ones which don’t; both pairs hit different numbers of fours. Zulqarnain and Saeed Ajmal both made golden ducks, which isn’t too surprising: neither of them was in the team for his batting, or they wouldn’t have been at No 10. The fact that Tauseef Ahmed and Junaid Khan both took a single to get the other batsman on strike is hardly even worth mentioning: it’s exactly what a No 11 would be expected to do in a tight finish with a better batsman at the other end.
The pattern throughout the list is that those statistics which were the same in both matches are emphasised, while those which were different are not mentioned. This is exactly how the improbability of a set of ‘coincidences’ can be exaggerated: by careful choice of examples. Imagine throwing two dice — if they are thrown only once, they will probably land on different numbers; it might even be considered a coincidence if the numbers happen to be the same. If they are thrown a thousand times, then most of those times they will land on different numbers, but sometimes the same one. By listing all the times they land on the same number and ignoring all the times they don’t, one can produce a list of many pairs of events which had identical outcomes, giving the impression that between them these constitute a remarkable coincidence — when in fact they can be expected from any long enough sequence of throws.
This, in effect, is what happens when someone compares two cricket matches with the aim of finding as many similarities as possible. A single match generates hundreds of statistics: runs scored, balls faced, boundaries hit, overs bowled, wickets taken, methods of dismissal, extras conceded, outcomes off a particular ball and so on. If the same list of statistics is generated for two matches, most of them will be different, but a few will almost certainly be identical. Discard all the different ones, highlight all the identical ones, and the result is a list of similarities between the two matches which can then be presented as an ‘incredible coincidence’ — which is clearly what the author of the Sharjah/Dhaka list has done.
To illustrate how such a list can be produced between any two matches at all, consider another result which has happened twice in India’s ODI history: a tie against England. The two matches in question took place in the same calendar year — at Bangalore in the group stage of the ICC World Cup 2011; and Lord’s in the fourth match of a bilateral series a few months later. By the same process of picking out the statistics which match and ignoring the ones that don’t, another list can be compiled of seemingly ‘astonishing’ coincidences. The following events were common to both matches:
In the same way, a list of ‘amazing’ coincidences can be made to link any two matches chosen at random; perhaps the biggest coincidence of all would be two matches which showed no similarity whatsoever.
(Michael Jones’s writing focuses on cricket history and statistics, with occasional forays into the contemporary game)
Play Fantasy Cricket & Win
Cash Daily! Click here