By Ged
October 31 2005

Nine punters, three computer-generated benchmarks, sixteen first-class matches, seven dwarfs, six geese-a-laying but only ONE winner and ONE small cricket-related prize. As the domestic cricket season ends, so does the Middlesex 2005 prediction game. Ged demonstrates that the...

...computerised benchmarks achieved better predictions than humans, but only a person can win a prize, so read on and find out who was hot and who was not.

The following nine punters played for glory in the Middlesex 2005 prediction game:

¨      Ged Ladd

¨      Barmy Kev

¨      Janet Davis

¨      Warren Ladd

¨      JC67

¨      David Slater

¨      Patrick Kingsley

¨      Tom Skelton

¨      James Duffy

 

There were also three computed entries, by way of benchmarks:

¨      1^st Class Averages (with “norms” for players who don’t have 1^st class averages)

¨      Middlesex 2004 averages (with “norms” for players lacking Middlesex 2004 averages)

¨      Poll of Polls (average of the 9 real entries).

 

Ged has observed throughout the season that picking previous averages seemed to be a better guide than the “wisdom” of the competing humans!  However, only real people can win a prize, so there was plenty to play for even though the computer did best!  If you recall, Ged has promised a small, cricket-related prize for the best human entry.

 

Ged has now recomputed the standings table at the end of the season, eliminating those Middlesex players with too few matches or innings to allow their averages to count towards the results.  In this interesting season for Middlesex, that left very few bowlers who were in the squad at the start of the season still in the standings at the end!

 

The computer has done far better than all the real people.  The Middlesex 2004 average has been far and away the best predictor throughout the season and was still best at the end of the season.  At the half year stage JC67 was standing in poll position for the prize and had exceeded the 1^st Class Average predictor.  But by the end of the season, 1^st Class Average was also doing better than all humans (see table below). 

 

Name	Bat Score	Bowl Score	Total Score	Ranking
1st class average	74	58	132	2
Middlesex 2004	52	52	104	1
Gedd Ladd	88	81	169	4
Barmy Kev	108	68	176	5
Janet Davis	136	109	245	12
Warren Ladd	108	129	237	10
JC67	101	78	179	8
David Slater	139	104	243	11
Patrick Kingsley	104	81	185	9
Tom Skelton	89	89	178	7
James Duffy	73	81	154	3
Poll of polls	90	87	177	6

 

James Duffy was the best placed human at the 12 match stage and indeed retained his lead over the rest of the humans in the end.  Congratulations then to the winner, James Duffy.  The promised small, cricket related prize will be winging its way towards you.  Do make sure you let everyone know what it is once it arrives.

 

The wooden spoon slot was fiercely contested.  At the half-season stage Janet was in that slot, while at the 12 match stage David Slater seized the wooden spoon slot from Janet.  Warren Ladd (no relation) has also threatened for the wooden spoon throughout.  But in a gripping finale, Janet has (just) managed to retain the wooden spoon slot, followed closely by David and Warren who should both be mightily relieved, although still somewhat chastened.

 

Ged himself unfeasibly came a decent second (amongst the humans), beating off a feeble challenge from Barmy Kev, the perennial “could do better than this” merchant of school reports and competition results.  James, Ged and Barmy Kev are the only three who can console themselves that they did better than one of the computer-generated benchmarks – the poll of polls (the average return by the nine real players).

 

If anyone wants to see the spreadsheet model that includes all the predictions and the formulae for computing the winners, you are very welcome to do so.  Please e-mail Ged at gedladd@hotmail.co.uk and he’ll e-mail you the spreadsheet.

 

Ged has enjoyed running the game and hopes people have enjoyed playing it.  Next year it would be nice to get some of the more pessimistic regulars from the board participating, so we can genuinely test whether the pessimists or the optimists are the more realistic followers of the team.  The spreadsheet model can accommodate as many players as want to play, so the more the merrier for 2006.  Suggestions for changes, improvements etc. will all be welcome, again to gedladd@hotmail.co.uk