September 6, 2011

Time for enhanced fielding stats

Peter Della Penna
Michael Clarke took a smart catch to get rid off Kumar Sangakkara , Sri Lanka v Australia, 1st Test, Galle, 2nd day, September 1, 2011
Just how much of a difference does fielding make in cricket?  © AFP
Enlarge

At the start of Australia’s commanding performance in the field on day two in Galle, Ricky Ponting took a spectacular catch at short cover to dismiss Tillakaratne Dilshan and gave Trent Copeland his first wicket in Test cricket. It was only a matter of time before Tony Greig uttered a classic cricket commentary cliché to describe the sequence during the slew of replays that were shown of the catch, afterwards. “He doesn’t drop many,” Greig declared.

I don’t know about the rest of the cricket viewing public, but I am quite curious to know for a fact just how many Ponting drops, especially compared to Kamran Akmal. The Argus review recently brought this to light and while the main elements of the review deal with how to fix other aspects of the Australian set-up, two paragraphs that dealt with fielding were what caught my attention most.

“For catching and fielding specifically, the panel recommends introducing explicit measurement of catching and fielding efficiency for all first-class and international players and teams,” reads the bottom of page 20 in the review. “These should also feed in to player rankings/performance incentives. One simple measure would be catches taken as a percentage of chances created. Chances could be weighted by difficulty if required. The same could be done for run outs. Measures of this nature have been standard practice in baseball and other sports for decades and should become standard in Australian Cricket.”

Just Australian cricket? Why not all cricket at all levels around the world? Cricket is a sport that prides itself on statistics more than most, but in terms of fielding stats, as well as what could be termed as negative stats, the sport is way behind. As the Argus report notes, fielding stats in baseball are standard. Things like fielding %, put-outs, assists – especially outfield assists – errors, baserunners thrown out or picked off by a catcher, are all marked in the fielding stats column in baseball.

By comparison, catches are tracked for fielders and total dismissals – catches and stumpings – for wicketkeepers in cricket, but the only tangible negative statistic for a fielder in cricket are the byes conceded by a wicketkeeper. The equivalent stat in baseball for a catcher would be the number of passed balls conceded. That’s extremely weak.

At the moment, reputations are made for fielders in cricket through anecdotes. Excluding wicketkeepers, Rahul Dravid has the most catches in Test history yet his fielding prowess is not talked about in the same way as Jonty Rhodes, Mark Waugh, Bob Simpson or Colin Bland. There is no historical record of runs saved between cover and backward point, just stories.

The lack of elaborate fielding stats is partly caused by the absence of specific scorecard designations assigned to fielders in cricket. When one looks at a detailed baseball scorecard, things like F-9 and 6-4-3 DP would be instantly recognizable to signify a fly out to the right fielder and a double play involving the shortstop, second baseman and first baseman. A cricket scorecard shows the name of a player who took a catch but nothing to note where it was taken. Only those with a dedicated knowledge of a team would know that “c Hayden b Warne” most likely involved a catch taken at slip instead of on the midwicket boundary.

It’s hard to properly designate a scorecard number to each player relative to their fielding position in cricket due to the fact that the field can change at any moment depending on the captain’s strategy at any stage of a game. However, even simple things could be applied such as designating a catch taken in a so-called catching position like the slips or gully, a catch taken in the circle in front of square, a catch taken on the leg side or the off side, a catch taken on the boundary, etc.

Similarly, drops should also be designated relative to where the chance was missed. Baseball scorecards immediately assign where a miscue occurred, i.e. E-5 is an error made by the third baseman. Drops or missed chances are not documented in cricket. Once again, stories of batsmen making the most of a second chance – such as Brian Lara being dropped by the Durham wicketkeeper on 18 and going on to score 501 not out for Warwickshire in County cricket – are only passed down through anecdotes and not in official scorecard documentation.

Curiously, cricket also chooses to reward batsmen such as Lara for a fielder’s misdeeds instead of punishing the fielder and sparing the bowler from negative stats. If a baseball hitter reaches base as a result of an error by a fielder, the hitter does not get credit for a hit. Simultaneously, the pitcher is not punished statistically if a run scores as a result of that hitter reaching base. This is reflected in his earned run average (ERA).

Unfortunately, cricket bowlers see their stats go in the wrong direction if a fielder has robbed them of a wicket. One thing that could be changed is to create a similar type of ERA statistic for bowlers and teams. One line would track total runs conceded during an innings. However, if a batsman is dropped off a bowler, any future runs conceded by that bowler to the corresponding batsman should be listed separately to note total runs vs. earned runs. Also, all runs scored after a missed chance when a team is nine wickets down would count as unearned. The unearned or bonus runs scored by a batsman after the drop would also be tracked in the team total.

For example, dropped or missed chances had a major impact on the just concluded Test match between Zimbabwe and Pakistan. According to the ESPNcricinfo commentary, Tino Mawoyo was dropped on 27 and went on to make 163 not out. Brian Vitori was dropped on 0 and made 14. So in effect, Zimbabwe scored 150 unearned runs in a first innings total of 412.

Pakistan benefitted even more from sloppy fielding by Zimbabwe. Mohammad Hafeez (dropped on 11, made 119), Misbah-ul-Haq (2, made 66), Younis Khan (30, made 88), Umar Akmal (13, made 15) and Sohail Khan (3, made 11) seized 240 unearned runs in a first innings total of 466. Take away the unearned runs from each side and Zimbabwe should have entered the second innings with a 36-run lead instead of a 54-run deficit. This shows the impact that fielding had on the match.

Baseball isn’t the only American sport that is way ahead of cricket in terms of defensive or negative stats. Ice hockey has the valuable plus/minus statistic which essentially tracks goals scored for a team compared to goals allowed while a particular player is on the ice. Last season, New Jersey Devils forward Ilya Kovalchuk tallied 60 points and led the Devils with 31 goals. However, his minus-26 rating was the worst on the team. Despite his offensive skills, the opposition had a far better chance of scoring a goal while he was on the ice than the Devils did.

Basketball tracks turnovers as well as assist to turnover ratio, a strong indication of a point guard’s risk/reward ability and the ball-handling skills of other players on the court. Steals and blocks are tracked and an NBA All-Defensive Team is named at the end of each season. Likewise, Major League Baseball awards Gold Gloves at the end of each season to recognize the best fielders at every position.

While the ICC now holds an annual awards ceremony which hands out gongs for Cricketer of the Year, Test Player of the Year, Emerging Player of the Year and Twenty20 International Performance of the Year, there is no such trophy to recognize the world’s best fielder. Even if there was, there would be a minimal amount of hard evidence to judge it on. Catches win matches, but apparently they won’t win you any awards in cricket.

Just because detailed fielding stats were not kept before in cricket doesn’t mean they should not be instituted for the future. Sabermetric stats such as OPS are modern obsessions in baseball thanks to people like Bill James and present a different way of viewing a player’s value. In the NFL, quarterback sacks were not an official statistic until 1982 but are now a major benchmark in judging a defensive player’s impact. It’s time cricket followed suit and started to officially track more detailed fielding stats.

Peter Della Penna is a journalist based in New Jersey

RSS Feeds: Peter Della Penna

© ESPN Sports Media Ltd.

Posted by Chris Cowles on (August 28, 2012, 17:49 GMT)

It looks a bit complicated to me. You can't plot every variable. How would you distinguish a c&b that was a quick reaction on a follow through from being handed a dolly for example? So, Bland, Solkar, Waugh and Rhodes were great fielders without having any record to show for it. Let it be known by anecdote. Stats can be absurd - Bill Johnson's batting average on the 1953 tour being the most obvious example - and some have no meaning, maiden overs for instance, which have no bearing on averages. I saw a football match which turned on two moments of brilliance a couple of minutes apart: a great save by Gordon Banks and a brilliant goal by Bobby Charlton. Stats show a 2-0 scoreline instead of 1-1. Only the anecdote can tell the true story. Before getting more figures into the game, instead please tidy up the written English.

Posted by Arun MAK on (October 30, 2011, 3:01 GMT)

Cricket and Baseball has been mixed & analyzed & compared so much. Just bcos both are similar doesn't make them same. Federer may not be as good in Squash too, tho both games look similar and the rackets too. Where is the author trying to lead the readers? This page is about cricket in U.S. Adopting cricket to stats levels of baseball should not be the idea, rather just adopt cricket for what it is. From the days of marking notches on a piece of wood to mark runs to this present day stats cricket has come a long way. Getting too deep into stats in each and every aspect would only kill the joy and spirit of cricket. Let's play it 'The Gentlemen's way', as it has to be.

Posted by Arun MAK on (October 30, 2011, 2:44 GMT)

Nice article. there was a comment about how many drops, in comparison to akmal. well, i wonder how the author missed India's kiran more in the list. Gooch's epic 333 + 123 contained drops that would be nearly in double figures and most by Mr. More. who then went on to be even a selector, and reportedly has a coaching camp too. (Kamran! are you a trainee from that camp?)

Posted by Catches Win Matches on (September 14, 2011, 12:02 GMT)

In a professional sport, where not only pride but millions of dollars are at stake, how can you have detailed statistics for 2/3 of the game but leave 1/3 of the game virtually stat free.

Despite many years of coaches and players saying that fielding is so important, until people started monitoring fielding performance, it was only the high calibre fielders who really cared. Now with software programs an analyst is able to give comprehensive data on each player's fielding performance. Each and every event can be logged so there is no hiding from the truth.

If a player achieves good performance, recognise it; if they achieve poor performance, find out what, where, when & why so that it can hopefully be prevented. Accountability is the key word here. Does the player contribute positively or negatively to the teams fielding performance? Human nature means we only remember the really good or the really bad; the other 90% gets lost in the depths of opinion, nostalgia & hearsay.

Posted by Rikki Gerhardt on (September 9, 2011, 21:58 GMT)

I've thought about a system where the scorer approximates the location where the ball first touched the ground, and where it was fielded, for every delivery. And the fielders involved. Might take a lot of paper. But emulating ball-tracking systems will be worth it.

There is a fielding award (not sure if there is an actual trophy) in county cricket, relying on modified traditional fielding statistics. Not sure how much it is adjusted by approximating catching chances.

One thing about errors in baseball though... it has been found that ERA is less descriptive of a pitcher's true skill than plain RA. That definitely doesn't mean that errors are not worth noting.

I still wouldn't adjust batting and bowling averages for errors though, since any level of fielding mishap affects batting (or possibly the other way around). To some degree it is arbitrary.

Posted by Chris on (September 8, 2011, 19:52 GMT)

Couldn't possibly work. Not all drops are created equal.

What about a flying dive where second slip only just gets the slightest of hands on it? Was that a chance? A half-chance? Perhaps they should be credited with only half of their subsequent runs.

What about one coming back at the bowler at a rate of knots, which they just manage to knock down, but would technically be a dropped chance, even though they'd have needed Matrix-style bullet-time to actually have a chance (see also: silly point "drops").

Posted by Sir_Freddie_Flintoff on (September 8, 2011, 13:01 GMT)

This is a very nice suggestion. It's pretty amazing that after 130 years of test cricket we still don't have any fielding stats other than catches taken.

However, the baseball stats that you've mentioned, like Fielding Percentage, are highly flawed. For example, according to these traditional metrics, Derek Jeter is a superb defensive Shortstop. But advanced metrics show, he has been a below average fielder all his life. He has poor range, which means he doesn't get to enough batted balls. So, Sabermetrics stats like UZR or Dewan's +/- are far better metrics, although they are still work in progress.

Having said that, we need to start somewhere with fielding stats in cricket. But I don't expect advanced stats like UZR to come up any time soon.

Posted by youngkeepersdad on (September 7, 2011, 23:12 GMT)

Cricket, like all games, involves an element of luck. That luck can go against a bowler, when the captain insists on the wrong field or butterfingers drop everything, or for him when a captain's tactical masterstroke or superb fielder pulls off a blinder. Even a great stop which keeps a vulnerable batsman on strike for an extra ball can tip the balance. It is impossible to document all such detail, however there is certainly no reason why scorecards can't outline where a catch was taken, the missed chances off a batsman or bowler and even particularly close lbw calls for starters.

Posted by Mustafa Pasha on (September 7, 2011, 21:28 GMT)

Great article!

As a kid growing up in Pakistan in the 90s I had the pleasure of watching two all time great bowlers in their pomp. I am of course referring to Wasim and Waqar. Even then Pakistani fielding left a lot to be desired and I always wondered how many wickets the two Ws would have ended up with if they had the catching quality of say Australia or South Africa to back them up. If cricket had implemented your suggestions two decades ago, I would have had a number to put to that wondering.

I am also glad that you mentioned Kamran Akmal. At the very least we should have a statistic for wicket keepers, recording how many chances they have fluffed. That way I could be secure in the knowledge that all the curses, screams and things flung at the TV after yet another Akmal 'Special' would not come back to bite me in some sort of horrible cosmic joke.

Looking forward to your next piece!

Posted by Sean on (September 7, 2011, 0:51 GMT)

As someone who follows both baseball and cricket, this is a subject I have given some thought to. However, it should be noted that measures such as errors and assists in baseball are quickly becoming outdated in favor of sabermetric defensive stats (UZR, +/- system, Defensive WAR, and a slew of others). Errors are a one dimensional way of assessing defense. Likewise,an assist is a stacking number that is based on opportunity and variables. While it would be interesting to track these in cricket, they couldn't give you a true measure of a fielder's ability. You would have to consider how many runs a fielder saves with his speed and range. And how would you weigh chances by difficulty? Its a subjective criteria.

Comments have now been closed for this article