Monday, 27 December 2010

Nice idea from Italy

I got a lovely e-mail from Marco in Rome the other day. Christmas Eve it was, and Marco revealed that he is a mad-keen Derbyshire fan, though he has never lived in the area. It illustrates again the pulling power of the county and how many people follow their fortunes.

Marco suggested that one method that Derbyshire could try to improve their fortunes was the Moneyball method, used with great success in baseball. Now I have to admit that baseball has never been a sport I have followed, though I went to a game in Memphis a few years back and enjoyed the family experience of the game.

Anyway, Marco highlighted that the Oakland Athletic baseball team, with a playing budget of 'only' $41 million, became competitive with bigger clubs who spent three times as much. As explained in Wikipedia:

"The central premise of Moneyball is that the collected wisdom of baseball insiders (including players, managers, coaches, scouts, and the front office) over the past century is subjective and often flawed. Statistics such as stolen bases, runs batted in, and batting average, typically used to gauge players, are relics of a 19th century view of the game and the statistics that were available at the time. The book argues that the Oakland A's' front office took advantage of more empirical gauges of player performance to field a team that could compete successfully against richer competitors in Major League Baseball."

In cricket terms, it essentially means that you pay less attention (though naturally some) to batting and bowling averages and more to the other things that people bring to the side, like fielding, team spirit and attitude. It was used successfully in the first year of IPL by the Rajasthan Royals in selecting a team. Players could win matches with brilliant fielding and a quick thirty as much as those who  scored more runs but were less interested in the team ethic.

It is, of course, what Eddie Barlow brought to Derbyshire cricket in the 1970s and Dean Jones in a later period. You get people playing as a unit and the unit is an unstoppable force. When everyone realises that eleven performances are essential to win matches (unless you have a couple of outstanding individuals) success is generally a consequence. It has been done in other sports too - Ipswich under Alf Ramsey and Nottingham Forest under Brian Clough had few outstanding players, but they worked for one another and it brought dividends. Leicestershire won the T20 with an average team that never knew when they were beaten.

If John Morris and Luke Sutton can fashion a strong team spirit for 2011 then Derbyshire can also be competitive. Whether they formalise an interest in the Moneyball system I can't say, but the principles are sound ones and I expect to see our teams work under a captain that I trust to lead from the front.

Thanks for your mail Marco and for your kind comments about the blog - and do keep in touch!

On a different tack, I can only assume that another correspondent the other night isn't a regular reader when he said that I never criticise Derbyshire. I do, but I try to make it constructive and don't do it for the sake of it, like some. When we play badly I say so, when I think things could be done better (eg membership rates, querying why we built a stand when we can't fill the ground) I say so. At the end of the day though, as I've said before ad nauseam, I am a fan. I could probably support a team that won more trophies quite easily, maybe Nottinghamshire as my folks live inside that border, but that's not going to happen...

And that next trophy, whenever it happens, will be all the sweeter because of that.

7 comments:

  1. Cheers dude, I appreciate the fact that you mentioned me on your blog but I have to partially disagree about the team spirit & attitude part. While I do agree that a great team spirit is needed to help all the players to perform at their best, you still need talent to compete. The point about "moneyballing" is to find data as objective as possible to spot possible weaknesses on the open market, which, as we know, in cricket is based mainly on three things: averages, totals and subjective analysis ("I've seen something in that player"). My point is that the best thing to do would be to put some context to numbers (a quick one, a run in Taunton is not the same as a run at Edgbaston, but how are they related?) or to collect data unknown to other clubs (such as for fielding, knowing exactly how many runs in the field is worth a player would gave anybody a huge advantage). Basically, to search for the best stats possible to judge a player. Of course, this won't mean that the team should stop the "subjective" analysis; this is just meant to integrate it. Eyes alone can be deceitful: you might see a fielder leaping right and left and holding onto every catch, and not notice that he's got a very small range, while somebody who's so fast that everything he does looks extremely easy goes completely unnoticed. Of course it would not be something fast and easy to do, but the earliest the better, and given Derbyshire current position (financial and in the rankings), I maintain my opinion, why not give it a shot?
    Cheers

    ReplyDelete
  2. Can't see a reason not to Marco!

    ReplyDelete
  3. It's a nice idea but I don't think it is as simple as you are making out. Simplified, Beane wanted players with a high on base percentage and high percentage of walks from their at bats. The principles behind Beane's theories were aided by the massive amount of statistical data widely available in Baseball, a simple sport - just 8 outfield positions with a win or lose philosophy.

    The data available in Cricket would not only be expensive but also be very subjective based on how the viewer interpreted how many runs a fielder had saved, the influence of a Captain upon the field, the state of the game, the conditions, etc. The data collector would have to find a way to level all of the influences out in order to compare players like-for-like. It's what I try to do as a Sports Economist.

    The 'Moneyball' principle allowed Beane and the Athletics to select players in the MLB Rookie Draft that no other team wanted as highly as thus give the A's leverage in low balling the player into a contract. Unfortunately the pool of talent ready to play first class cricket is far smaller!

    Sorry to put a downed on things! It's a nice idea though! The book is a good read too!

    Gareth

    ReplyDelete
  4. Hi Gareth, I never said or dreamed that it would be easy to do so, nor that it would be possible to reach immediately the same depth as in baseball (a decade or more, I reckon), or for what matters that DCCC will go on to fight for every trophy in three years time if it will follow the Bill James/Billy Beane route :)

    But still, you have to start somewhere, and I find astonishing that we have no objective gauges about fielding, apart from catches and runouts! There is no doubt that at the moment we can only dream about the accuracy of baseball's UZR (Ultimate Zone Rating) or Pitch/FX and such, both because of cricket's variable situations and lack of data, but it doesn't justify the current desert, and besides that there are still lot of areas where an improvement could be done even now: an example of a stat that would be easily obtained, the difference in percentage between the average of a bowler and the average of the averages of the batsmen he dismissed: If I take 2-60, and the two wickets were two guys averaging 50, it is a better performance better than a 2-30 against two tailenders who are averaging in single figures. Then, it is a perfect stat? Of course not, it has so many flaws that I don't think I have enough space here to list them all :) But, does it tells us more than the pure average? Yes it does, no doubt about it.

    Now I don't want to bother anyone so I will not put down a few other examples, but then I'm just a random cricket fan, it is too absurd for me to hope that somebody whose job is cricket in the side I support could do something in this direction? For sure it will not harm us, as for the cost... you just need an extended scorecard and a couple of more scorers to start, how much could it be?

    Anyway it's good to see that there is interest about this subject!

    Marco

    ReplyDelete
  5. Hi Marco,

    Sorry if it came across like I was having a go at you. I wasn't and it wasn't suppose to sound like I was.

    I like your idea about the boilers and averages, which analysis could be completed on now quite easily with the statistics available. It could be long winded but at least it is achievable. At least the results might help you select a balanced bowling attack or identify underrated bowlers. However, the same effect could be quickly achieved by simply looking at whom the bowlers has dismissed.

    I'm trying to think of ways to remove the different game scenarios from the data to allow comparisons of players i.e. how could you compare the SRs of a batsman shutting up shop to save a game to one whose been bowled fodder in order to set a contrived ending to allow a result. There must be a way though.

    I like your thinking and suggestions though. They are very interesting!

    I'm not sure all the moaners on 606 will allow money to be spent on some statistician rather than another signing though! ;-)

    Cheers

    Gareth

    ReplyDelete
  6. Well, something that would be good to compare batsmen would be to dissect their averages in categories: how much a batsman is averaging (and at which SR) against bowlers who are taking wikets at 20 or less, 21-25, 26-30, 31-35, 36-40, 41-45 and more than 46. If somebody pumped his average or SR because Tom Poynton was bowling, it would be shown this way. Then, about the game saving ability, I think you could look at average balls faced in certain situations, however this would be a bit more complex, I have to think some more time about it. At the moment I'm working (veeeery lazily) to an amateur system to adjust First Class and List-A data by ground and country.

    Ah, just a quick note: given that in County Cricket a win is worth 16 points and a draw just 3, the ability to save a game has become a lot less important (and marketable) than the ability to set up results. Personally I think that this is a bit unfair, for as long as these are the rules, to sign an overseas batsman has become nearly a waste (of course only from a First Class perspective). The current point system is way, way too unbalanced.

    Cheers

    Marco

    ReplyDelete
  7. Ah, I forgot:

    A way to put a comparable value over everything that's happening on a cricket field, is it to understand how much in percentage that action has steered the game to a probable record.

    Example: a team is 200-5, chasing 300. Our possibility to win is, say, 50%. If I made a stunning runout, our possibility to win might become of 70%: the worth of my action could be valued as 3,2 championship points (20% of 16, the points for a win. Then we should see if it was only my merit or captain's as well, but for the sake of the principle let's keep it simple now). If we're trailing by 300 in our second innings at the beginning of day 4 and I score 60 off 250 balls to the close of play, the value of my innings could be explained as the difference in probability of a draw (worth 3 points) at the moment I went at the crease and the moment I got out (so it would be negative, if I went for a duck). You add all these data througout the season for all the player, and at the end, you know how many championship points was due to anyone of them. But then, It's a bit early for such a system. I hope I explained myself well enough :)

    ReplyDelete

Please remember to add your name. Avoid personal comment at all times. Thanks!