Cricket Moneyball

Although I haven’t been paying close attention to this year’s IPL, simply because it’s a big year ahead and I need a bit of a rest before the World Cup and the Ashes, I am incredibly interested in how T20 and cricket in general might evolve in the coming years. I’m especially interested because I recently read Michael Lewis’s Moneyball and found it very interesting indeed.

For those who aren’t familiar with the book, it describes how Oakland A’s general manager Billy Beane, who was once a highly drafted college baseball player who bombed badly as a pro, managed to take a team with a small budget to the World Series finals. He did it by finding market efficiencies – in other words, identifying undervalued players who could perform a specific role incredibly well, even though they weren’t brilliant all-rounders or the archetypal athletes other teams coveted.

In doing so Beane created a highly competitive team despite spending comparatively little money on wages. And he soon transformed the way franchises thought about baseball. Instead of looking at how tall a player was, or how many replica jerseys he might sell, Beane was only interested in statistics and the ‘value’ a specific player could bring.

The parallels with cricket aren’t exact but I’ve always thought they’re relevant. Just look at the recent success of Harry Gurney. The Notts left-armer has rarely excelled in first class or list A cricket, but his left-arm angle and good cricket brain has made him an effective performer in T20, even though his reputation before this year’s Big Bash was hardly stellar. Beane would’ve gobbled up someone like Gurney as soon as he hit the market. He wouldn’t care that he wasn’t as slick and stylish as a Mitchell Starc. He’d bring him in to perform a specific role and he’d do it well.

I’ve often wondered how long it would take for T20 cricket to catch up with baseball in this regard. Although T20 has always been portrayed by many (including this site!) as merely a bit of hit and giggle, I do think it’s becoming more interesting now that the statisticians and analysts are getting involved. It’s intriguing how somewhat random players like Wayne Madsen can win lucrative T20 contracts in the PSL. Basically someone is doing their homework and trying to find efficiencies. And I think this makes T20 a much more serious and stimulating business – even though aesthetically it will never be the equal of first class cricket in my opinion.

Consequently my eyes lit up when I saw the following interview with data analyst Dan Weston on the Betway Insider blog. Weston is a former gambler who previously used his statistical expertise to win money. These days he applies his knowledge to identify players for subcontinental franchises instead. He’s therefore ahead of the game when it comes to cricket statistics – which makes him something of a revolutionary in a way that Peter Moores never was.

Weston is a former gambler who uses millions of balls’ worth of cricket data to calculate and predict the success of cricketers in all formats around the world. Like Billy Beane he eschews traditional scouting methods, which tend to rely on feel / gut instinct, and instead focuses solely on stats. Players therefore become numbers rather than personalities. It might sound a bit cold but at least his decisions are evidence-based.

He believes the key is to know which statistics are vital and which ones to ignore:

Cricket is full of inane data like: ‘This is the slowest century by an English batsman on a Tuesday.’ It’s completely worthless. Cricket teams make mistakes. You name it, they’ll make that mistake, be it selection, recruitment, in-game tactics”

Weston’s solution was to set up a business that supplies data to teams, players and agents that aims to eliminate these mistakes. His model is a complex one. He takes a players’ individual data (average, strike rate, economy) and makes adjustments to it based on recency, opposition quality, and the conditions in which each tournament is played in.

Because I follow American football I know they use this kind of statistical data regularly on the other side of the Atlantic. A site called Football Outsiders, for example, uses a metric called DVOA (which stands for defensive adjusted value over average) to determine who the best teams and best players are. Rather than using simple metrics like ‘metres made’ or ‘touchdowns scored’ they work out rankings for players which take into account specific circumstances and the toughness of the opposition.

If they did one for cricket, for example, they might find that Alastair Cook was top of the run scoring charts but merely average in the DVOA tables. This is because (as I’ve demonstrated ad nauseam over the years) he made so few hundreds against top bowling attacks in tough conditions and instead excelled at making huge scores when conditions were benign. Rather than arguing about how good certain players are on social media, DVOA actually decides these arguments for us.

Weston uses a similar formula to calculate expected stats for players at upcoming events. This enables him to recommend players to teams around the world, including for the IPL and T20 Blast.

Cricket is a conditions-driven sport. So a T20 Blast match at Canterbury will be a pace-orientated affair, whereas in Dhaka it’s going to be spinner-friendly and low-scoring.

If a batsman performs well at Canterbury, does that really apply to a match in Dhaka? Probably not. There’s limited relevance. So I analyse how historically similar players have made the transition from one league to another.

I might be asked to find a pace bowler for the T20 Blast, where an Australian will be quite highly-rated, as opposed to the IPL, where they haven’t thrived as much as their reputation would suggest because of the quality of the league.

‘Reputation’ is a key concept here. Weston believes that reputations can fool a team’s management into making poor selection and recruitment decisions. Picking on reputation is lazy. Instead he prefers what cricket fans might call horses for courses selections – although it’s a bit more sophisticated than that.

Lots of high-profile players are signed based on reputation rather than current ability. Take Brendon McCullum: he’s got a poor record against spin bowling, he doesn’t keep wicket anymore, yet subcontinental teams are signing him as a marquee player. It makes no sense whatsoever.

Interestingly, Weston believes that a statistics-driven approach will eventually lead to a resurgence in the popularity of specialists in cricket. This is obviously music to my ears. Although Ed Smith might not be so keen as it somewhat undermines his whole selection philosophy.

While T20 teams often try to fit as many batting and bowling options into their team as possible, Weston begs to differ. Why? Because cramming a plethora of all-rounders into a team can negate their effectiveness – especially when it comes to the bowlers.

You don’t want to stick an all-rounder at No. 9 because he’s just not going to bat in T20. The average No. 8 faces about seven balls per match, and the average No. 9 faces about four balls per match. If those guys are required to face more than the average, your top order batsmen haven’t done their job properly.

For No. 9, 10 and 11 you just want an out-and-out specialist bowler who would perhaps then be capable of playing a five-ball cameo. If you pick too many all-rounders you end up compromising where they bowl, because often they turn out not to be very good death bowlers.

Weston’s approach is certainly working. And he has the numbers to prove it. Indeed, teams around the world are slowly beginning to take notice. For example, last year he wrote an analysis of 10 English players who would perform well in the subcontinent. His top five all got signed by franchises … including Wayne Madsen!

As a result I expect we’ll see a lot more analysts and statisticians involved in T20 cricket in the coming years. Yes this might make T20 cricket even more like baseball, but at least it will give spectators and cricket followers something else to think about.

The sight of six after six disappearing into orbit soon has lost its lustre as far as I’m concerned. I prefer test cricket for this very reason. It’s less repetitive and the ebb and flow of the game slowly seduces you. However, if I knew there was something more strategic going on behind the scenes then I might be inclined to take T20 more seriously. Even if one doesn’t particularly enjoy watching the format on the field, one can still appreciate the science behind assembling the perfect roster off it.

James Morgan

23 comments

  • Sorry James, your usual excellent writing, but I just can’t get interested in T20 stuff anymore!

    • For years it has been said that test cricket is dying. I remember reading a Geoff Boycott book in the early 90s where he said that in the future there will only be limited overs cricket. And now players such as Jos Buttler say there will only be T20 in the future.

      But a cricket world consisting of only various T20 leagues would be tedious wouldn’t it? How would it work?

      • Even as someone who likes its a T20 only future would get to the point of overkill pretty quickly I’d probably pick a team in a couple of comps, cut out watching matches as a neutral and watch other sports the rest of the year.

        There are still plenty of challenges for the format to overcome for that to happen as most are not making a profit and would be unviable if an expanded or second IPL were to overlap with them.

  • An enjoyable read and an interesting concept as far as selection goes, but perusing stats is not my idea of a fun afternoon.

  • When T20 first became a big thing back in the mid-2000s, and I was a baseball-stat-loving PhD student with spare time on my hands and a superfast computer, I calculated a series of new T20 stats and player appraisal methods. I’ve yet to see any evidence that any of these “professional” statisticians have managed to come up with anything anywhere near as useful or sophisticated as I came up with in a few weeks coding.

    Take this nonsense for example:

    “You don’t want to stick an all-rounder at No. 9 because he’s just not going to bat in T20. The average No. 8 faces about seven balls per match, and the average No. 9 faces about four balls per match. If those guys are required to face more than the average, your top order batsmen haven’t done their job properly.
    For No. 9, 10 and 11 you just want an out-and-out specialist bowler who would perhaps then be capable of playing a five-ball cameo.”

    This is almost shockingly stupid. It confuses averages with distributions, and what “should happen” with what “might happen”.

    If one of my analysts came out with something this daft I’d be appalled. Whether or not the top order batsmen “should” have done a better job is irrelevant. The fact of the matter is that every X number of games, you’re going to be in a situation where you’re 7 down with 5-6 overs still to go and a competitive target still a long way off. The ability of your number 9 at this point can make a huge difference in your probability of winning the game.

    What’s more valuable, a bowler with an adjusted econ of 8.0, and the ability to score 30 off 15 if needed to save a game, or a bowler with an adjusted econ of 7.8, and no ability with the bat whatsoever?

    (Its a trick question, you don’t have enough information about how his bowling fits into your plans). Looking at individual’s adjusted econ is daft, you need to look at the impact on the overall team adjusted econ between him and his likely replacement (Effectively bowling VARP).

    • I’m not a statistician but I expect the numbers of a team winning from being seven down with 6 overs to go must be pretty close to zero even If you had the best number nine in world Cricket walking in.

      • I don’t think so – I’ve seen plenty of games where a top order collapse has been followed by a late recovery to post a par score, that has then been defended.

        I remember being at the Oval when Middlesex were something like 80-7 in their first innings of a London Derby… recovering to get to something like 130 and winning by ~5 runs or so.

  • If Mr Weston thinks Canterbury (where I watch 90% of my cricket) is pace orientated…then I suspect I won’t be investing my hard earned pennies according to his advice!

  • Interesting read from Richard Gibson in the DM about the unfolding economics of HHH. Revelations include one county pro being asked to take a £20k pay cut and players being afraid to speak out becuase of the threat to their retirement pot.

  • There was a Pakistan expert commenting pre and post games in the recent T20s Pakistan franchise league who was very good. He was against hiring all big hitters and filling the teams with all rounders. The wickets were variable and the bowling good. This put the batsmen on their mettle. He admired quality batsmen who could score in all conditions and said every team needed one batsmen who could act as glue if wickets started to fall, to rotate the strike, pierce the field and get partnerships going. If you are a batsman who targets sixes only a good bowler could tie you down. The rise of the bowler is making T20 more of a contest so captaincy skills and batting skills come more into play. Seeing wickets fly is also entertaining- ECB please note. If the six hitters do manage then to hit a six against tight bowling it is a thrill.

  • I’ve never understood the desire to draw any comparison between baseball and cricket. I know of no one who is interested in both and throughout the world, apart from some peripheral interest in Australia, a country with more general cultural links to the US than most, with their common roots in frontier lifestyles, no other country I can think of that bothers with both.
    The attitude to stats is completely different, with baseball’s total obsession compared to cricket’s almost oddball fascination with the more obscure, for entertainment purposes above anything else. When we buy our annual ‘Playfair’ booklets to get us through the season we are more fascinated in the biographical details of players than their stats. Similarly with Wisden, it is the oddball features that are the most read and commented on. Only when major landmarks are about to be reached do cricket fans seem to pay much attention. I know of no cricket fan who defines any player by their stats, with the possible exception of Bradman. It is the experience of watching the greats live that we remember, certain shots, balls or great pieces of fielding, not whether they make hundreds or take 10 wickets is almost immaterial. How many of us out there would say after watching the likes of Warne, that we think of him in terms of stats above all else.
    The only common factor is marketing, which has become a universal science in most team sports.

    • I don’t entirely agree with you. I am a traditional cricket fan, preferring test cricket by far and I also like baseball! As an MCC member too and I know many other members who enjoy baseball. The strategic and tactical approaches do bear some remarkable similarities actually and of course the collection of statistics around both games is both legion and common.

      Spending a day watching players and chatting for hours about their stats and those that have gone before is all part of the joy for many of us that don’t just go to get drunk and sing. I suspect we are a dying breed but we not dead yet!

      • I didn’t say there weren’t any who liked both, merely I didn’t know any. The key here is that, as I mentioned, apart from the Aussies, no other country in the world seems to care less with the possible exception of the Japanese, who have no interest in cricket. I agree there are strategic and tactical similarities, but these are more to do with the similarities in the games’ structures. Look at the respective cigarette cards. In baseball they are totally stats oriented, in cricket they are more biographical with references to major achievements.
        All of us cricket fans like chatting about the relative merits of players, but stats are hardly the most important part of this, unlike with baseball fans. Indeed they are often a major bone of contention. When we compare cricketers merits accross the generations stats are pretty meaningless, otherwise there would be no debate, as stats are written in stone. It’s no coincidence that ‘Top Trumps’ cricket has never taken off amongst kids, despite its success elsewhere.

        • You might be surprised to learn that baseball and cricket are actually about on a par in terms of international enthusiasm. The most recent cricket and baseball world cups had more or less identical numbers of countries competing, prize money, live spectators and tv spectators. Whereas cricket is only really a major sport in a handful of large ex-British colonies, baseball is the number 2 sport in the whole of North and Latin America and East Asia, and is also significantly more popular and prominent across Europe (with several professional leagues).

          • We’re getting off the point here, which was the comparison between the two. I am not disputing baseball’s popularity, especially in South America, as its next door to the home of the sport and has many from that continent playing it professionally in the US, with many of the games televised so the youngsters can see their heroes play. On the other hand cricket as an old colonial export makes it pretty obscure globally. It’s not that I hate baseball, it’s just I don’t understand the desire to compare it with cricket. The only crossover I see is with fielding, where the flat fast throw has undoubtedly been a recent import and for the betterment of the game.
            Baseball is a more accessible sport to play and understand than the more arcane cricket. It’s natural to hit and throw, but not to bowl and bat, where natural instincts have to be curbed, throwing being a foul and cross batted swipes tending to get you out.
            As I’ve said before, the fact that you can spend a full day watching part of a game with no end result, maybe not even seeing your favourites on the field atall, gives cricket a unique mentality amongst its supporters. Baseball, like other sports, is over in one sitting. It doesn’t make it better or worse, just different.

    • Huh. Plenty of people like both cricket and baseball, maybe you just don’t know many people.

      The two games are surprisingly similar. The easy-paced rhythm, the laid-back stadium experience, the inherent nostalgia, the almost mystical quality, the sheer number of people who just love the game and would happily sit and watch any team play any time anywhere, the fans snoozing in the bleachers. the remembrance of endless childhood summers, ignoring the dismissive sneers about being “slow” and “old-fashioned”, the sounds of leather on willow/ash/maple, a cold beer, the feel of the sun on your cheek, the long afternoons in the outfield, the arcane and often unwritten rules, the jobbing pros, the young hopefuls, the lifetime habit of checking in every day to see the scores, even if you can’t watch the game.

      Visiting a minor league or spring training baseball game is very much like visiting a county championship match. Watch an amateur game in this country, you’ll see its very similar to club cricket.

      • Who are these plenty of people? If there was significant interest in baseball over here we would have major league coverage in much the same way as we get the NFL. TV shows virtually nothing, not even when the World Series is underway.
        The similarities in structure undoubtedly exist as you have pitchers and batters and fielders, trying to get each other out and restrict runs. However, being American the winning is promoted as key. The thing about professional cricket is you can quite often play for a number of days and have no result. Can you imagine your average American sports fan putting up with that for the beauty of the game.
        The baseball nostalgia is for achievement, success if you like. The main teams all come from cities, where the pace of life has always been traditionally hectic, with little longing for hazy lazy days.
        Even the highly romantic and rather beautiful ‘Field of Dreams’ (if you build it he will come) is very much a one off, with no imitators before or since, requiring a certain intellectual sensibility of nostalgia. Your average working class fan did not appear too interested, as was evidenced by its poor reception at the box office. A labour of love it certainly was, but a commercial success it wasn’t, despite having a major box office draw in Costner, largely because there was no contest involved.
        I know this is only a theory and there to be shot down, but I believe the passive Englishness we like to celebrate in cricket and the overt American patriotism celebrated in Baseball are entirely different animals but they do reflect the differences in national character. I don’t think it’s possible for the two to operate in the same way.

        • We do get major league coverage – BT Sport 2 shows virtually nothing else for 6 months of the year, and there are a huge number of subscribers to MLB tv.

          Having been involved in both, the UK baseball and softball community is many times larger than the American Football community, with leagues up and down the country.

          The MLB are playing 2 games at the Olympic stadium this summer. Despite ticket prices north of £200, all 160,000 tickets sold out in under 3 minutes and are now re-selling on viago for 4 figures.. The NFL games at Wembley don’t even sell out anymore.

          “being American the winning is promoted as key.” Having played in the states, no its not.

          “The baseball nostalgia is for achievement”. having lived in the states, no its not.

          “The main teams all come from cities” Whereas the Oval, Edgbaston and Trent Bridge are surrounded by fields?

          I’m not going to argue every detail of baseball with you, there isn’t much point, as clearly you really don’t know very much about the sport.

          • Not sure why we need to discuss baseball here. Actually I quite like it and used to watch it in the States.But it’s not cricket by a mile, and who watched BY Sport 2? I’ve never heard of MOB tv. Would be interested in the viewing figures.

  • The Billy Beane story and the subsequent success of Oakland followed by the Mets was obviously designed to squeeze money out of the high salary bills. Statistical analysis clearly has merit but when compared to cricketing situations the critical mass is so different. One baseball season has 30 plus teams playing 162 games with a potential 20 more through the playoffs. This is a much larger database than any cricket format all built into a six month period.

    One statistic I like to use for test batters is a measure of how many hundreds they score per innings played. I find it a very interesting set of data.

    • One baseball season has 30 plus teams playing 162 games

      ^^^ This is the key. Baseball moneyball is a true statistical exercise where the odd game your whole team has a shocker gets lost in the big numbers and evened out over the rest of the season.
      If in a T20 season of 15(? – don’t know how many the IPL has any more) you have two or three collapses it makes much more of a difference to your season’s performance as a whole. Which is why teams like the insurance of all-rounders in the lower order. Similarly the ODI World Cup, where, especially in the knock-out stages, a single bad game loses the tournament.
      In baseball, even in the World Series, you can lose nearly half the games and still walk away with the prize.

  • Hi James, I love your site and check most days for new content. Sorry to be “that guy” but it’s market inefficiencies. If the market was efficient, there would be little opportunity to exploit it

  • Isn’t DVOA a team statistic though?

    We’d need Wins Above Replacement for bowlers and batsmen shouldn’t we?

    I’d also love a Depodesta approach to working out the runs and wickets needed per game to win a tournament?

FOLLOW US ON TWITTER

copywriter copywriting