Apps To Fill The Void In Fielding Metrics > Cricket News, cricinfo, mobilecric, cricbuzz, livescore and more

Cricket news - The apps working to fill the void in fielding metrics

While CricViz sticks to the basic quantifying of fielding data, ProFielder is more extensive and aligned to the requirements of a fielding coach

Here's a question. In one match, a player scores 153 from 121 balls as his team wins by 106 runs. In another, he puts down a simple catch. The opposing batsman uses the reprieve to move from 14 to 84 in a tight game that the opposition wins by just 14 runs. Which of the two contributions - the century or the dropped catch - was more telling?

The sharper-eyed will recognise that the player in question was Jason Roy, who was rightly lauded for his matchwinning innings against Bangladesh early on in the World Cup. But how much focus has gone into his dropped catch that let Mohammad Hafeez off the hook during Pakistan's win at Trent Bridge?

Cricket has exhaustive statistics that help us quantify performances by a team, a batsman, a bowler or a captain. And yet, fielding data has until recently been limited to catches taken or stumpings completed. There have been no records of dropped catches, let alone missed run-outs or fumbles - at least publicly. While teams employ data analysts to compile and analyse this information for their own strategic purposes, the viewing public have thus far been left in the dark.

Recognising this void in cricket data are two apps that have emerged into the public realm during this World Cup: CricViz, which brings the "next level of cricket analysis" to fans, and ProFielder, a creation of renowned fielding coach Chris Taylor. Both have their own unique ways of assigning numbers to fielding performances, a territory few have ventured into.

"Fielding analysis is an area of the game that is heavily underrated and underused," Taylor tells Cricbuzz. "The ProFielder tool has incredible details that you don't usually see on TV. Everything from the moment a ball is hit by the batsman is recorded - where the fielder is positioned and how he stopped the ball, with a tumble, a dive or a roll. A throw is also monitored. How accurate the throw was, which end it was to, and by the end of the game we have an overall throw quality and number of shies a team made. There is an algorithm that runs in the background to calculate the throw rate and throw quality."

CricViz, on the other hand, takes into account each incident that could affect the final team score. A run saved, a dropped chance, a misfield or a missed run-out would all go into their database. Each of these events is then assigned a plus or minus value based on the impact it had.

Instead of categorising all such events under one umbrella, CricViz also identifies the probability of the fielding event to be a success or failure - 90% being the easiest and 5% being the hardest - and assigns the run value with a few basic guidelines in place. The fielder leaderboard you see on TV in the World Cup broadcast is generated by CricViz. The runs saved or leaked is assigned a positive or negative value based on impact.

"The percentage represents the likelihood of the event being executed successfully," says Freddie Wilde, an analyst at CricViz. "We use basic guidelines as well, for example we know that around 70% of slip catches are successful - so that is the starting figure for slip catches, then we go higher or lower than that depending on the difficulty of the event." In Roy's case, CricViz assigned a 90% chance of success for the catch he shelled.

But how does one gauge the cost of a dropped catch? It depends on how you measure it.

In the past, one might have said that Roy's dropped chance at long-off cost England 70 runs - the difference in Hafeez's score when he was put down and when he departed. But that is an oversimplification, given that other Pakistani batsmen would have scored runs from the deliveries Hafeez faced after the drop. When it came to quantifying the cost of Roy's drop, CricViz evaluated the runs Hafeez was expected to score based on when he was dropped in the innings. This forecasting model added 18 runs to what Hafeez had scored when he was dropped.

But - and herein lies the difficulty in creating fielding metrics - other approaches could end up with a different value to 18 depending on where the emphasis is placed. This is also true when it comes to assessing the overall fielding performance of a team in a match.

For example, ProFielder's system said that England's overall fielding performance against Pakistan cost them 21 runs - the difference between winning and losing.

Yet CricViz disagreed, saying that as a whole, England in fact saved runs in the field, with their positive impact events outweighing the fielding blunders, which included Roy's drop as well as captain Eoin Morgan allowing four runs through his legs. This is because CricViz uses two metrics to evaluate fielding events - Impact and Ability.

"Impact measures the impact that fielding has on the final score, and Ability measures the quality of fielding," explains Wilde. "Impact considers when events happen in the innings - earlier is more valuable; Ability disregards when things happen and therefore is a more effective way of gauging the quality of fielding. A good way of distinguishing between Ability and Impact is if a catch was dropped off a no-ball, Ability would still punish the fielder because it is looking at evaluating the ability of the fielder, but Impact would not because the catch has not impacted the scorecard."

While England's fielding performance against Pakistan elicited different evaluations from the two different tools - and thus provided some insight into the subjectivity of fielding metrics - everyone agreed on the quality of the home side's showing during the opening match against South Africa. Taylor and several others, including England analyst Nathan Leamon and CricViz, rated it as one of the best fielding performances by a team in international cricket.

One of the highlights of that game was Ben Stokes' sensational one-handed catch to dismiss Andile Phehlukwayo in the deep. The fielder himself admitted he was out of position initially which made the catch more difficult than it could have been. While it was a remarkable catch in the end, would he be able to pull it off again? Shouldn't he be informed by the fielding coach that he needs to stay in the right position?

The two fielding data teams record events like this in different ways, and offer different insights. Stokes' catch had just 5% probability of success according to CricViz, who deemed it to be worth +11.4 runs.

ProFielder, meanwhile, notes how difficult the catch was and recognises if the fielder was out of position only if it's a regular occurrence. It also identifies "hotspots" in the field or areas that are targeted regularly at a certain stage of the innings or against a particular bowler. The deep mid-wicket fence is a major hotspot when a leg-spinner - in this case Adil Rashid - is bowling to a lower-order left-hander like Phehlukwayo. The app can also pick out some of the reasons behind a poor fielding display - such as fielders placed in the wrong positions, or responding negatively to the pressure they are put under during the fielding event.

"ProFielder can give insights into where a particular fielder is more effective - where they make a mistake more often or which area gives him a higher chance of making an impact in the field. Fielding at mid-wicket needs a different set of skills compared to fielding at short fine leg for instance," Taylor explains.

"In terms of pressure, if a ball is tucked down to fine leg and the batsman are only looking for a single, the fielder can take a bit of time to return the ball. It's not the same as a quick single when the ball is hit in front of backward point. You need to be proactive then and attack the ball. There is an amount of pressure that comes in. The 'pressure gauge' (in the ProFielder app) in a way shows if a player is a poor fielder when under pressure by batsmen or situations. There are algorithms again to judge this."

While CricViz sticks to the basic quantifying of fielding data, ProFielder is more extensive and aligned to the requirements of a fielding coach. Rather than viewing one as better than the other, it would be better to see them as complementing each other to provide a more holistic analysis.

Neither tool is just for outfielders. Both CricViz and ProFielder recognise wicket-keeping on par with outfielding, and highlighted Shai Hope as one of the best wicket-keepers in the early part of the World Cup. The keeper converted four out of four chances in West Indies' first match against Pakistan and ProFielder recorded an overall wicket-keeping index of 98 for his performance.

While being simple enough to understand, CricViz and ProFielder give the sort of detailed information that is often discussed in a more casual - and often completely subjective - manner on commentary. The data gathering in itself, though, remains subjective - two viewers might judge what constitutes a 'chance' differently, for example.

CricViz has a data gathering team in South Africa, who have a series of guidelines by which to measure performances in the field. ProFielder too has people tracking everything from fielding positions to rolls, tumbles and dives. What constitutes a tumble or percentage of success of a chance is highly dependent on the person recording the event. Also, a catch taken when the opposition needs three to win or when a team needs 50 runs to win is categorised similarly at the moment. But both parties acknowledge these loopholes.

"This part of the analysis is obviously subjective but we like to think we have a framework in place to remain consistent and fair," Wilde says. "Impact considers the time in the innings when things occur but it doesn't consider the match situation. That is something we are looking at doing as a third measure which considers how fielding influences win probability. But it's not something we do right now."

Live Home Series
Player Schedule
Points Table Cricket Rankings
Cricket Clothing Cricket Equipment
Cricinfo Apps Download