Skip to main content

What's Better; Heuristics or Statistics

My favourite cricketing site (Sideon view aside) is Edmund Baylis' "Red Ball Data".  Sadly he posts less than he used to but at the start of the 2021 season he predicted the finishing positions for the county championship.  I've always had a suspicion that a lot of statistical work has a negative value, not adding much to the accuracy of predictions at the cost of a lot of additional complexity

So Red Ball data providing free to air predictions gives us an interesting mini experiment.

What I've done below is, in the left hand column, set out the final county championship tables for divisions 1&2.  The middle column is Red Ball data predictions (Red Ball had a few teams in equal positions which I've just ordered any old way) and the right hand column is the teams, first sorted into division 1&2 and then by 2021 finishing position.  


ActualRed Ball Data2021 Order
    
1SurreyLancashireWarwickshire
2LancashireEssexLancashire
3HampshireSomersetHampshire
4EssexKentYorkshire
5KentSurreySomerset
6NothhamtonshireHampshireEssex
7SomersetYorkshireGloucestershire
8WarwickshireNorthhamtonshireNorthhamptonshire
9YorkshireWarwickshireSurrey
10GloucestershireGloucestershireKent
11NottinghamshireNottinghamshireNottinghamshire
12MiddlesexDurhamDurham
13GlamorganGlamorganGlamorgan
14WorcestershireMiddlesexMiddlesex
15DerbyshireWorcestershireWorcestershire
16DurhamSussexLeicestershire
17SussexLeicestershireDerbyshire
18LeicestershireDerbyshireSussex
    
    
Out by #    0    32    46
   
The figure to look at is at the foot of each table. The actual championship table perfectly predicts the final county championship table and is "out" by 0 places. The Red Ball Data predictions are out by 32 places, or 1.7 places a team & the rule of thumb, 2021 finishing position table, is out by 46 places 2.5 places a team. A pretty easy victory for Red Ball Data and statistics. 

The outperformance came largely from three teams. Warwickshire, last years winners came 8th this season, Red Ball data predicted 9th and would have been spot on if it hadn't been for this . Surrey this year's champions had a poor year in 2021 and Red Ball Data had them up at 5th for 2022 and Red Ball Data correctly predicted Kent would finish in mid - table, whereas their 2021 form saw them predicted to be last in the 2022 championship division 1. 

Hopefully Red Ball Data will publish his predictions for the 2023 championship and we can see if the outperformance against heuristics continues.

Comments

Popular posts from this blog

County Championship Salary Cap

This is post about salaries in county cricket. The first class counties are subject to a cap and a collar on amounts paid in wages to cricketers.  They must pay above a collar, currently £0.75m, and below a cap, currently £2m. There is an agreement for both the collar and the cap to increase over the next funding round to 2024. In 2024 the collar will be £1.5m and the cap £2.5m What is less clear is what payments count towards the cap and collar.  I assume employers' national insurance (a 13% tax on wages) isn't included.  Similarly I assume payments to coaching staff don't count towards the cap as if they did, Somerset, Lancashire and Yorkshire would all be over the current £2m cap.  I've gone through the accounts of the first class counties to see what, if any, disclosure, they include on players' wages.  What gets disclosed varies enormously, quite a lot for some counties, nothing for others.  Additionally there is a possibility the information include

Mo Bobat and County Cricket

Cricinfo has this  interview with ECB "Performance Director" Mo Bobat.  Bobat makes an interesting claim about county cricket, "Take something like county batting average. We know that a county batting average does not significantly predict an international batting average, so a lot of the conventional things that are looked at as being indicators of success - they don't really stand true in a predictive sense."  And later in the article there is a graph, showing county averages plotted against test averages for 13 English test batsmen.  This is reproduced below. better than random? raw data suggests no meaningful link between championship and test averages 20 25 30 35 40 45 50 55 60 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 Test County Championship Sam Curran England players' batting averages

English County Cricket Finance: 2018 Bentley Forbes Rankings

I have gone through the most recent financial statements for the English first class counties,  made an estimate of the financial strength of each and given them a Bentley Forbes Consulting ( TM ) financial sustainability ranking.  The overall table looks like this. County      Profit Assets Ranking Position Essex   4   4   4   1 Surrey   1   7   4   1 Nottinghamshire   5   5   5   3 Somerset   2   8   5   3 Derbyshire   8   3   5   5 Leicestshire    6   6  6   6 Sussex  15   1  8   7 Middlesex  14   2  8   7 Kent     9   9  9   9 Worcestshire    3  15  9 10 Gloucestshire   7  12  9.5 11 Northamptonshire   11  13  12 12 Glamorgan   16  10  13 13 Durham     12  14  13 13 Yorkshire    10  17  13 15 Warwickshire   17  11  14 16 Lancashire   13  16  14 17        The approach is to rank the counties for profitability and balance sheet strength and combine the two measures in a sustainability ranking. The balance sheet strength is itself a combination of thre