Tuesday 31 July 2018

Fakhar Zaman's amazing start

18 matches. That's all it took.

1 year and 45 days to go from relative obscurity to being one of the most talked about cricketers in the world. Fakhar Zaman has certainly burst onto the world stage.

I wanted to look into the context around the record, and also to see how his start compares to others.

I then will look at how well we can predict someone's eventual career record based on their start.

The first question is, if we add in context, how does the innings stack up.

Over the past 3 posts, I've looked at using a measure of Added Value to see how much better a player performed than the average player for their era, against the same opponents in the same locations.

That basically shows how much that player is contributing to the team's cause.

How does Fakhar Zaman stack up against others after 18 matches? Is he still on top of the pile, or is the fact that a number of his matches were against Zimbabwe going to count against him?

The answer is that he's still very close to the top.

He's just behind Viv Richards, but ahead of everyone else.

Of these 10, he's 3rd by average, 4th by strike rate and top by batting index.

The difference in that field shows just how much more high scoring the modern era is, when compared to the early days of ODI cricket.

The next question is: how unusual is it to score 1000 runs in 18 consecutive matches. Is that something unique, or is the only unique thing that he did it at the start of his career?

Here is the list of the most runs scored in a batsman's best 18 match sequence:
I had to learn how to turn my computer screen sideways to be able to get a long enough screen shot to include Zaman, but he is there, and there have only been 58 batsmen ever manage to score 1000 runs in 18 consecutive matches, so making the list when you've only played 18 matches is very impressive.

Important to note that this only includes matches played for their country. I did not include matches for the World XI, Asia XI, Africa XI etc.

At the top of the list is Virat Kohli, who has just recently overtaken Sachin Tendulkar in terms of total runs scored, but is still a way behind the little master in terms of value added (although the metric of best 18 consecutive matches is probably not one that they really care about too much.

The next question was how well the initial 18 innings predicted later success.

I chose an arbitrary set of matches (101-150, and compared a players initial results with how well they did in that later set. There was more of a relationship that I was expecting, but there was still an awful lot of variation.

The model that can be constructed off that suggests that Fakhar Zaman is likely to score between 1595 runs in that period and 3270. That's basically the difference between below average and the second best 50 match stretch ever. That would suggest that the runs are not a great predictor of future success.

The next thing to check is batting average. Is the batting average a good predictor?

Again, the evidence suggests that it isn't. Of the 4 players who averaged over 50 in their first 18 innings, only 1 averaged over 50 later in their career. If we pull the benchmark down a bit, of the 16 who averaged over 40 at the start, 12 averaged over 40 in innings 101-150.

Using the model from this data, we would expect him to average somewhere between 29.4 and 67.4 That's the range from Hailton Mazakadza to AB de Villiers.

Next I thought I'd try seeing how added value worked as a predictor.

Again this was at best a moderately good predictor. There is a more clear relationship, but there is still an awful lot of noise.

This measure suggests, however, that we can be confident that he will go on to be an above average batsman. The confidence interval for the model here is 110 to 1138. To put that in context, in that period Michael Clarke had an added value of 67.6, Younis Khan had 87.6 and Stephen Fleming had an added value of 49.4.  If he ends up with a career as good as those 3, he will certainly be an asset for Pakistan.

 Finally, I decided to try using a data science technique to make a prediction based on all the information that I have.

This tended to slightly underestimate the results for batsmen and overestimate the results for bowlers, so once I took that into account, I could make a prediction range for how Zaman's career would turn out.

The range of values that came out were 567 to 930. That puts him in some very good company. Only 16 batsmen had an added value above 567 in that period. It suggests that we can be confident, based on his start, that, given he is still playing after 150 matches, he will be very successful. The "given" condition here, however, is very important. Only 20 out of the 94 players who made a good start with the bat to their careers went on to play 150 matches.

In conclusion, Fakhar Zaman has made a very good start to his career, probably the second best of all time. However, that doesn't mean yet that we can start being too bullish about his career. The start to his career suggests that he is likely to be successful, but by no means is he certain to be.

1 comment: