'Moneyball' shows cricket's cliches are crooked

It is indeed true that the amount of analysis that goes on in cricket is much less that in baseball, but that doesn't mean it doesn't lend itself to more and more analysis. Look at what Andy Flower has achieved and tell me that he isn't using statistical analysis and getting his bowlers in particular to bowl with very strict plans to players. It is making people like Bresnan look like world beaters, but in reality they are simply maximising their ability and bowling to a very good plan. This doesn't mean he is that much better than say Darren Gough of 10 years ago, it is that he is being better advised about which plans work and which don't. As for saying that cricket is more or less complex than baseball, I think complexity is a very narrow word to use. There are potentially more variables to measure when analysing cricket from a macro sense, game situation, pitch, weather, etc. However, measuring success is i think one of the hardest measurements in cricket. Monty Panesar blocking out to draw the game against oz, is probably the single most statistically valuable innings ever, but does that mean you sign him as a batter? Where I think it can be extremely valuable is for measuring perfomance under pressure, but then we already have the world ratings for bowlers and batsmen. It would be extremely interesting to try to quantify the value of an entire bowling attack, because where England for example don't have a Mcgrath or a Warne, they undoubtedly have a world class attack, without having a stand out player. Using analysis to give very clear plans to the bowlers is patently very good mentally for the players particularly when the plans work. Beyond everything else, what is the absiolute truism is that no team stays on top forever in cricket, but planning how to unravel a team is extremely difficult. Wasnt' the decline of the WI and Aus simply caused by retirements and old age?

Phil, in order to measure DIPS (DIBS?) you need some way of measuring fielding. It is relatively easy in baseball because batsmen and pitchers have little control over balls in play. I have looked at (again, not published) runs scored against and for per ball independent of boundaries (on the basis that a boundary is unstoppable), but it wasn't satisfying. Not least because boundaries are stoppable, but also because year-to-year measures are heavily dependent on which team was being played, grounds. There really is a paucity of comparable data in cricket, especially as the ball-by-ball statistics sabermetricians have produced are almost non-existent. An aside, the thing that did come out of the study was that boundaries per ball massively increased around the year 2000. I haven't had a chance to see if that was an artifact of bad data (unrecorded boundaries) or an actual effect, but it equated to 2-3 runs per hundred balls (which equates to 203 runs on a batsman's average too)

Great discussion above. As a cricket stats tragic, I adored Moneyball. Together with Freakanomics which I read at the same time, the book has had a deep impact on me in many aspects of my life as it has resulted in me consistently questioning many accepted wisdoms. Anyway... Bill O'Reilly had his own version of sabremetics when it came to One Day cricket. He was the first to espouse that one day teams should just pack themselves with 11 batsmen (a few of who could hopefully bowl) and start throwing the bat from ball 1. Soon after his death, Sri Lanka won the World Cup playing a variation of that approach.

You know what would be a fun study - some fielding independent studies of bowlers measuring their impact in a manner similar to DIPS in baseball - wish I had the time. It is even more relevant now considering the review system and more consistency with LBW's . To give you an idea, here is a good defintion of DIPS "In baseball, defense-independent pitching statistics (DIPS) measure a pitcher's effectiveness based only on plays that do not involve fielders: home runs allowed, strikeouts, hit batters, walks, and, more recently, fly ball percentage, ground ball percentage, and (to much a lesser extent) line drive percentage. Those plays are under only the pitcher's control in the sense that fielders (not including the catcher) have no effect on their outcome." Does anyone know if this has been done with cricket ? It really helps with baseball and working out if a pitcher has been lucky over the course of a season - a few good comments above have touched on this , more so with getting tail enders out . Thoughts (I really rushed this post - hope you see the general point of it..)

"the real lesson of moneyball, i think, is to encourage questioning and critical thinking about game." This is right on the mark.

Lolly - if you want heavier stats websites for baseball check out fangraphs, and baseball think factory You'll be impressed

You are looking more at the end result not what drives it in tlking wickets and runs. For baseball the two results are not giving up runs and scoring runs what James saw was that OPS was the best statistic to determine how likely you were to score runs. What you need to look at is what behaviours drive scoring runs and stopping your opponent from doing so. Will it be as easily identifiable as in baseball – no but is it doable – I think yes.

but wouldn't that subjectivity recorded stat be better than the fat zero we've got now? You could do runs of shots through that players "area"

Nice article B. I wrote the following a few days ago for my blog (www.therestijustsquandered.com) after reading about Moneyball and was going to post it on roar but it appears my thunder has well and truely been stolen. I guess that movie has put stats on the brain for everyone. Well, no point posting now, but here it is if anyone's interested. It covers a few point being debated right now... Whilst I readily admit that cricket needs more statistics about as much as Todd Carney needs another boozing scandal, it is worth floating the idea that the current cricket numbers might not be telling us what we really want to know. As it stands in the wonderful world of bowling stats, we can easily see a players's economy rate (avg runs per over), strike rate (wicket per balls bowled) and of course the measuring stick used to compare all those who roll the arm over - the bowling average (runs per wicket taken). But whilst these three stats provide us with a few numbers to throw around in pub discussions (most likely with Todd Carney) they tell us very little in terms of a bowler's impact on a match or the calibre of player they are getting out. Have they removed the other teams most dangerous player with a test average of 60? Have they broken a longer standing partnership than Bert & Ernie? Have they sent the opposing teams captain's middle stump cartwheeling faster than a pre pubescent russian gymnast? These are questions that the current abundance of cricket stats give us no insight into. To take a simple & current example - the Australian first innings in Johannesburg, Ismar Tahir took the tidy figures of 3-55. But his 3 wickets were an out of form Brad Haddin, Number 9 Peter Siddle and a genuine bunny in Nathan Lyon. Whilst Philander's figues of 1-47 and Morkel's 1-62 looked significantly weaker performances, the former broke an 174 run opening stand which was the beginning of the end for Australia and the latter cheaply removed the Australian captain and first test centurian, Michael Clarke. It could easily be argued that in this particular instance the contribution of Lobster Morne & The Philanderer were at the very least 'on par with Ismar.' I'm sure if one was to dig a little deeper they could find countless examples of flattering statistics of those cleaning up the tail and others where a bowler's contribution is being seriously undervalued. To combat this problem, I propose something of an effectiveness rating which takes into account the average of the batsmen removed and the strength of the partnership broken. Maiden overs could also be factored in as they build pressure and have an undeniable psychological bearing on the batsmen, even in the test arena. Dropped catches could also be incorporated in the rating as a bowler that endures matches, or even entire series, without luck and reward is generally forgotten about a year later when the selectors are only examining the numbers on the paper. Like wise with batting, no one likes a fair-weather player who only scores a tonne after strolling out with his team 2/300 on a flat second day deck. Runs made in crisis situations are immensely more valuable and although I don't have the statistics to back it up (cause they don't yet exist) I dare say blokes like Ian Healy and Michael Hussey would perform better than their averages allow if these elements were factored in. Hawkeye, strike rates, run rates, over rates and wagon wheels that seemed to be a lot tastier back in primary school - it can easily be argued that cricket is already a bit stats heavy. But compared to the endless streaming of baseball numbers or Quarterback's pass 'rating' which are standard fare for those American stat obsessives, there could be room for additions or even tweaking in the cricket arena. Whilst some might laugh at the Yanks' tendency to obsess over stats, Brad Pitt's latest offering Moneyball shows the power of a statistics spreadsheet and an example of a practical application of them. For the record I haven't seen the movie, but anything with the guy who ate all the pies from Superbad in it is usually good for a giggle. I am usually very dismissive of people who call for rule changes or more statistics, but in this case I think it would be for the betterment of comparing cricketers. Exactly how it could be implemented I have neither the time nor the statistical capabilities to calculate. I'll happily play the Dale Kerrigan role of 'Ideas Man' on this one and like a nut that's just outside off on a pretty good length, I'm leaving it through to the keeper. The one thing I am sure of however, is the new 'effectiveness rating' (go ahead and call it something snappier if you so desire) would give Ritchie & the boys something new to talk about this summer and help us separate the men from the boys in the cricketing arena.

What if that top order wicket was Hughes? Hardly a rare scalp.

Yes, loads of space for more stats analysis. It's one of the things I love about baseball.

great article and thread. one thing i remember from moneyball (the book) and background reading are discussions about how baseball lends itself to statistical analysis, for a number of reasons. 1 is that the game really only has a certain of possible for each situation (i.e. # outs, strikes etc, then hit to which fielder, throws to which base) so that the game can be be described by numbers. 2 is, as above, the HUGE sample size. where the innovation of Bill James (father of sabremetrics, which lead to moneyball) came in is that he questioned, essentially, the key performance indicators of baseball, which were batting averages, and pitching ERA's, among other things. using stats, he found value in players in less obvious ways. on-base %, and many others. With cricket, the KPI's are runs for batters and wickets for bowlers. and really, there is no substitute for this. So i think, the opportunity for statistical analysis to unearth new valuein this way is limited. the sample sizes and maningful data are much too small anyway stats analysis about who was a better bowler etc, posthumously, is different and for the tragic fans only However, what Bill James himself says about sabremetrics was really about was questioning long-standing logic/myths of the game. this goes back to the original post. I think cricket has HUGE scope for questioning the accepted logics of the game. i.e. thorough analysis of batters shot placements, dismissals etc. this is probably already occurring and increasingly so. the real lesson of moneyball, i think, is to encourage questioning and critical thinking about game.

Catches dropped, Ian, certainly (for all fielders in fact, it would be like tackles missed/made), but runs saved can really only be judged on the boundary. In the infield, it becomes completely subjective how many runs may have eventuated had the fielder not been there...

Great post Coorey...and others, good discussion all round...except oikee of course. Baseball game situations are way more complicated than some mention on here. Being 0,1 or 2 outs with runners on base or not on base, means far more permutations. Relief pitching, matching like handed pitchers with batters, matching pitchers who have an edge over certain batters. When to bunt, sac fly, steal base, Even a hitless at bat can be a good outcome, i.e gaining a walk, or just facing 8-10 pitches that tires the pitcher. With cricket i would like to see more analysis on TV on how batsmen lose their wicket. i.e where they've been caught, no. of LBWs , bowled, outs v spinners, v fast bowlers. When they were out e.g opener with 10 outs and eight in the first 2 overs after a break. Bowlers record in opening overs of a spell compared to last few overs of a spell. More weight on scoring runs in 1st innings or when it's a "live" game i.e if game is petering out to a tame draw. There is so much more analysis that can occur in cricket that would show us a lot more about trends and be able to compare players. (i'm talking about Test cricket as T20 is just fluff).

I think both you (kick to kick) and Phil Coorey are suffering from language and definition difficulties. I 100% agree that hitting a home run off a top class pitcher throwing at 100mph is probably the most difficult skill in sport BUT that is difficulty not complexity. (Although those that have seen the West Wing might appreciate this, "Ask any sportsmen what the hardest thing to do in sports is and they'll all say hit a baseball, but a coach once told me the most difficult thing to do is walk into the dressing room at half-time of the Superbowl [insert relevant big game here] and change the strategy that got you there because it's not working anymore.") Simply put there are a greater degree of variables to track and quantify in cricket making the numerical analysis of the game a more complicated task than that of baseball. It is honestly not up for debate at all (Phil in his post below can "agree to disagree" but that doesn't make him correct) - it is simple objective fact. Cricket is the more complicated game. Not more difficult to play. Not better. Just more complex.

I'd generally agree with Brendon, other than: "Losing a game now so you can win a game in the future is pointless. The end result is the same – you’ve gained nothing."...unless the game you've won by creating player depth is the final. and "If the younger player can contribute more than the older player immediately then of course he should be in the team. But if the younger player is left out of the team for another two years, he won’t disappear or suddenly become old himself. If a cricketer gets to first class level he’s not going to suddenly stop playing because he was overlooked for a team."...except that, these days, that is almost certainly what will happen and to think otherwise is simply another attachment to old wisdom. Two years is a very long time in a sportsmans career, especially when players are considered past it once they approach their middle thirties. So if the younger player is left out, he'll head to England for county cricket or India for the IPL. Maybe he'll improve and return to be considered again another day, maybe not. I'm not saying it is right or ideal, but the management of sport is a very different thing these days than it was. It was a huge thing for D'Ollie to uproot his family and move to England in pursuit of opportunity. These days it is done at the drop of a hat with as little thought.

Cricket needs much, much better stats, especially defensive stats. Its amazing something as obvious as "catches dropped" isnt statted for wicket keepers. Catches dropped and runs saved are obvious. I'd very much like a "recognised batsmen dismissed" stat for bowlers as well.

Oikee, Stop being a code warrior. A better question would be "Could Rich Thompson have played cricket for Australia" ? http://web.theabl.com.au/news/article.jsp?ymd=20110722&content_id=22128446&vkey=news_l595&fext=.jsp&sid=l595

An interesting article re the Moneyball movie and concept - http://mobile.slate.com/articles/sports/sports_nut/2011/09/more_moneyball_same_problems.html This suggests that the movie and its story are just that - a good story - in that the key players for the Oakland team were pitchers (who, like bowlers, win matches) who had been picked up the old fashioned way. At the same time it acknowledges that more advanced statistical analysis than was done previously has a place.