Tuesday 3 February 2015

Martin Guptill and the form myth

Every season there seems to be a cause célèbre among NZ cricket fans. In 2013 the call was that Brendon McCullum wasn't scoring enough runs, and needed to be dropped. In 2013-14 it was that Peter Fulton wasn't scoring enough runs and needed to be dropped. This season the overwhelming majority of cricket talk in New Zealand has been about one man: Martin Guptill. Apparently he isn't scoring enough runs and needs to be dropped.

In either calls to Radio Sport or comments on the Vietchy On Sport facebook page there have been at least 21 players suggested as being a better option as an opener than Martin Guptill. People have suggested different ways that he might get injured in order to get him replaced in the squad.

But the opinion that Guptill's significantly out if form is not just confined to the uninformed public (I consider anyone that suggest Michael Pollard, Peter Ingram or Kyle Mills as replacements for Guptill uninformed). There have been a number of the country's sports journalists join in. In a quite well written and balanced piece, Andrew Alderson noted that Guptill "struggled for form." Charlie Bristow talked of Mike Hesson needing "to handle Martin Guptill's stuttering form." Mark Geenty commented that the top order was carrying "significance and concern." Guy Heveldt said that Guptill is "under immense pressure to find some form before the World Cup begins." Daniel Richardson said that Guptill is "out of touch", "has done little to inspire confidence" and that his "form is a concern."

And it's easy to see why they felt like that. Guptill was dismissed for a duck in the first over 3 times against Sri Lanka. In two of those innings he was dismissed first ball. In another one he was lucky to survive a DRS malfunction and a hard-to-tell-if-it-bounced-or-not catch appeal. He got 66* in Auckland, but his other innings saw him score a total of 58 runs in 6 innings. Those 58 runs came off 94 balls, giving him an average of 9.67 and a strike rate of 61.7. It's hard not to be alarmed by those numbers.

However once we look a little deeper an interesting pattern emerges. Here is Guptill's numbers by bowler (I started compiling this before the Pakistan series, I'll include those innings at the end):

BowlerBallsRunsWickets
KMDN Kulasekara48344
NLTC Perera25321
SMSM Senanayake17110
RAS Lakmal1660
KTGD Prasad36200
HMRKB Herath220
AD Mathews28191

The issue that Guptill had was not so much lack of form as it was he struggled against Kulasekara.

Here's a better comparison:

BowlerBallsRunsWicketsS/rAverage
KMDN Kulasekara4834470.88.5
rest of SL12490272.645

If we exclude Kulasekara, he averaged 45 at a strike rate of 72.6. Now that's still not dominating, but it's certainly not a record to be particularly concerned by. In Guptill's career, no bowler has dismissed him more than 3 times, except for Kulasekara, who has dismissed him 6 times. The remarkable thing about that statistic is that Kulasekara has only bowled to Guptill 8 times. All 6 dismissals have been before Guptill has reached 15. In the one time that he made it to the 10th over against Kulasekara (the Auckland innings) he only faced 14 balls from him in the first 10 overs, and he scored off two inside edges that missed the stumps and had one ball fall just short of a fielder. The other time that he made it past 15 he only faced a total of 14 balls from Kulasekara before he was dismissed by Mathews. The spell that Kulasekara held over Guptill is one of a bowler having an edge over a batsman, rather than a batsman being out of form.

One bowler having the wood over a batsman like this is not actually as unusual as one would expect. Ewen Chatfield dismissed Allan Border 4 times in one series, Sarfraz Nawaz dismissed Gordon Greenidge every match in one series, as did Dilhara Fernando to Paul Collingwood, and went on to dismiss him in 9 out of the 14 matches that they played against each other. Faoud Bacchus was dismissed 3 times in 5 days by Jeff Thompson. Stephen Fleming was dismissed early by Chaminda Vaas a number of times also. Stephen Boock dismissed Martin Crowe 10 times in 15 matches.

It can be easy to say that a batsman is out of form, when it can actually be either a technical issue or a psychological issue.

I then looked through to see if Martin Guptill had a similar issue with other right arm inswing bowlers. At international level, most right arm bowlers tend to swing the ball out, so there is not a huge number of them to look at. However, two that sprung to mind are James Anderson and Chris Martin. Obviously Guptill hadn't been able to face Chris Martin in ODI's as they played in the same team, but there were some games that they played against each other in domestic cricket. In the 3 List A games that I could find, Chris Martin had never dismissed Guptill. He had been dismissed a couple of times by Hamish Bennett, and also by Shane Bond, who both swing the ball in, but both were considerably faster than Kulasekara, to the point that they are not really similar bowlers at all.

The next step was to look at his matches against Anderson. When playing for Derbyshire, Guptill never played against Anderson, but for New Zealand Guptill played 5 matches against Anderson. Anderson dismissed him twice in those 5 matches, but he's scored a total of 78 runs off 85 balls that he faced, for an average of 39 and a strike rate of 91. That's not anything to be particularly alarmed by, but once we look a little closer a different picture emerges.

In their first meeting Guptill either edged or was beaten by 8 of the first 10 deliveries he faced from Anderson. In the next innings it seems that the ball didn't swing. Guptill faced 10 deliveries from Anderson before pulling a hamstring and having to retire hurt. In the next innings he mostly watched Anderson from the other end, only facing 8 deliveries from him in the first 10 overs. The next time he faced Anderson he hit New Zealand's highest ever score, 189 (155), however, his first 10 deliveries from Anderson saw him edge one and play and miss two. Anderson went for 31 off 31 eventually, but most of those runs came at the death, once the ball had stopped swinging. In the next game Anderson got Guptill out with the 8th ball, but he'd already found the edge twice, once dropping just in front of a fielder.

I spoke to a former coach of Guptill's at domestic level to ask him if he felt that theory had any merit. He said that Guptill had always had difficulty early on against swing. He said when he was coaching it was left arm inswing that was the issue. Guptill hasn't been dismissed by left arm swing bowlers much recently, but it could be that the old issue is re-emerging in a slightly different form.

He has actually struggled against the new ball in general. He has been dismissed for less than 15 in 41 of his 96 innings, and his median score is only 21.5. This means that in half of his innings he has been dismissed for 21 or less. In some ways he is probably more suited to batting at 3 or 4, but that's not likely in the current NZ lineup. Once he's through the early period, he's a much harder batsman to bowl to.

In fact, his average score of innings over 15 is 65.57. This means that once he gets to 15, he can be expected to add another 50.

This made me wonder if he was going through a particularly bad run due to Kulasekara, or if it was actually just an expected combination of results that would have to come up at some point.

One of the ways to test this is to randomly select 1000 groups of 6 innings from his innings, and see how many of them are have a lower average, strike rate or batting index than the 6 innings against Kulasekara in this series. If less than 10% are below that, then we can say that he is probably out of form. More than 10% under the Kulasekara innings' and then we don't actually have enough evidence to say that it's likely to just be coincidence.

Here's the results of my randomization:

We can see here that while these innings are not great, they are within the range of what we would expect a normal 6 innings block to be.


The red line here is the middle 80%. This would have a batting index between 11 and 57. Anything outside this would be considered exceptional.

Guptill's batting index for the 6 innings that he faced Kulasekara was 14.3, so there were about 15% of the randomly selected series that had a lower index than the actual one.

What this means in context is that we can't say from this data that Martin Guptill was particularly out of form against Sri Lanka. While it's clear that he had struggles with Kulasekara, it's not clear that he was out of form generally.

There is some significant question if form actually exists. Generally a batsman's average is a better indicator of their next score than their recent average or their last innings. (Although a fuller exploration of that will have to wait for another post).

Finally we can look at a fuller breakdown, using some information that came in after I started writing this article

This is adding in the two matches against Pakistan, and looking at his average against Kulasekara and against everyone else.

BowlerBallsRunsWicketsS/rAverage
KMDN Kulasekara4834470.88.5
not Kulasekara260205478.851.25

Overall, I believe there is no evidence to say that Martin Guptill is out of form, nor is there any to say that he was out of form against Sri Lanka. His low scores there could be the result of his struggles against Kulasekara or a product of his general pattern of inconsistency throughout his career. But either way he can be expected to continue on scoring runs whenever he gets a start. One of the things that he will need to work on is getting a start against the swinging ball more often, particularly the one that is swinging in to him.

6 comments:

  1. I've had some comments saying "isn't this proof that Guptill shouldn't be picked because he regularly scores under 20 and had an issue against swing bowling?"

    To answer that: his median score of 22.5 is actually higher than the overall median of all preening batsmen in the past 2 years. Most opening batsmen have roughly half their innings under 20. That's the risk when you play against the new ball.

    Also, swing bowling is very hard to face. That's why quite a few teams use swing bowlers that bowl at 125kph early on, rather than seam bowlers who bowl at 140. It's why Kyle Mills, Nuwan Kulasekara and Zaheer Khan have all had such good careers. Also, most right arm swing bowlers don't create the same angles as Kulasekara because their default delivery goes away from the batsman. Having an issue with one bowler at the start of the innings is certainly not a reason to drop Guptill.

    ReplyDelete
  2. I've had some comments saying "isn't this proof that Guptill shouldn't be picked because he regularly scores under 20 and had an issue against swing bowling?"

    To answer that: his median score of 22.5 is actually higher than the overall median of all preening batsmen in the past 2 years. Most opening batsmen have roughly half their innings under 20. That's the risk when you play against the new ball.

    Also, swing bowling is very hard to face. That's why quite a few teams use swing bowlers that bowl at 125kph early on, rather than seam bowlers who bowl at 140. It's why Kyle Mills, Nuwan Kulasekara and Zaheer Khan have all had such good careers. Also, most right arm swing bowlers don't create the same angles as Kulasekara because their default delivery goes away from the batsman. Having an issue with one bowler at the start of the innings is certainly not a reason to drop Guptill.

    ReplyDelete
  3. Great stuff. Can we trust the NZ selectors and management appreciate this?

    It seems odd to ask for more information after reading such a detailed post, but I did wonder how Kulasekara had dismissed Guptill. I assume, but don't know, if those dismissals support the argument about Guptill's difficulty with the (in) swinging ball.

    Chris (Declaration Game)

    ReplyDelete
    Replies
    1. He got him a number of different ways. He was caught off the outside edge from the one that didn't swing after 2 that did. He was lbw from one that swung in. He hit a bad ball straight to cover, and he edged a wide one that swung in then seamed away to Sangakkara. All but one was a very good piece of bowling, and the two that weren't first ball were situations where he was set up by the previous deliveries.

      Delete
  4. I enjoy the article and the deeper analysis.
    Completely off topic but I was wondering whether you could find who the best (professional) sports team in the world is (was) by win percentage - google doesn't really have the answer and it might bring up some conclusions.
    I'm betting the All Blacks would be up there with a win percentage of around 84%.

    ReplyDelete
    Replies
    1. I honestly have no idea. I also don't know how I'd find out. Dynamo Berlin won the East German football league 10 years in a row. They might be an option.

      Delete