Fully Scientific Selection

[Manee]

Hi,

Was wondering what reservations the CC posters on CW have about a domestic or international team employing a more scientific and less intuition based selection policy (i.e. the sort of stuff that is regarded as common sense on these forums) for their team, rigorously.

For example, we would not have seen the debut of Cummins as quick as we did and he picked up six wickets. We would have seen the debut of Vinay Kumar in Test cricket earlier and he did poorly and perhaps is not Test quality, though I am aware of the dangers of judging after just one match.

I cannot think of better examples atm, but the question I am asking is, is the consensus of selection in CW via a statistical method the best way to select teams or does this system throw up some major errors and anomalies and what are these?

 

[Flem274*]

Very Simplistic Scientific Selection of New Zealand from the start of our summer

Peter Ingram
Brendon McCullum
Mathew Sinclair
Ross Taylor
Kane Williamson
James Franklin
Kruger van Wyk
Daniel Vettori
Tim Southee
Mark Gillespie
Michael Mason

Make of that what you will...

edit: Cribb's CPL database selections would be more scientific. Would be itstl what his sides are off pure CPL.

Basically, I'm all for scientific selection but there is a point - Franklin's Law - where ya gotta stop. There is also room for the Cummins/Bracewell selection, but if it flops the selectors should be rightly slammed.

You also have to include an interpretation with your raw data i.e. think about how likely Player X is to transfer his assets to international cricket and what you are looking for in your side.

 

[Spikey]

in the future all selection panels will have been made redundant by the PEWS database

 

[Spikey]

this will lead to England having to forfeit a must win Ashes test due to the database being down

 

[Manee]

You also have to include an interpretation with your raw data i.e. think about how likely Player X is to transfer his assets to international cricket and what you are looking for in your side.

And is this something that can be codified at all? Is a fast-medium or RF bowler more likely to convert skills or is it strictly case-by-case?

 

[benchmark00]

Would be the worst thing in the world itbt.

 

[Flem274*]

And is this something that can be codified at all? Is a fast-medium or RF bowler more likely to convert skills or is it strictly case-by-case?

Definitely case by case. I'm all for Medium Rights.

 

[Flem274*]

Would be the worst thing in the world itbt.

You're the sort of **** who would pick James Marshall aren't you?

 

[Prince EWS]

edit: Cribb's CPL database selections would be more scientific. Would be itstl what his sides are off pure CPL.

Yeah I've made a lot of posts about this. In the match Watling kept wicket against Zimbabwe, CPL would've picked the exact same New Zealand XI that New Zealand actually did, bar van Wyk in for Watling. van Wyk has since gone on to show that hardly would've been the worst selection ever as well.

Not even I believe teams should select completely off something like this though. Fully scientific selection as stipulated in the thread title would be beyond ridiculous for several reasons.

I would actually post the "What CPL thinks each Test side should be" lineups up, but it'd compromise the CWPL draft going on at the moment.

 

[Ruckus]

Selection has to be based both on stats and observation, not one or the other. E.g. there are heaps of cases of players who have awesome initial records in domestic competion, but later fade into mediocrity. If you were using just blind stats to pick them, it might seem the right decision when their record looks impressive but it would be missing the crucial observations that should tell you their record flatters them and they aren't actually good enough for the highest level.

 

[Manee]

Definitely case by case. I'm all for Medium Rights.

What about the other way around. I'd expect a very quick but wayward bowler to dominate Indian domestic cricket (e.g. Mohammad Sami) but struggle in Test cricket.

 

[Arjun]

An Indian T20 team that's based largely on T20 specialists picked from the Syed Mushtaq Ali Trophy.

 

[Manee]

An Indian T20 team that's based largely on T20 specialists picked from the Syed Mushtaq Ali Trophy.

eH?

 

[Flem274*]

What about the other way around. I'd expect a very quick but wayward bowler to dominate Indian domestic cricket (e.g. Mohammad Sami) but struggle in Test cricket.

Well that's where interpretation comes in.

And I don't necessarily think a quick bowler would get away with being wayward, especially in a seamers graveyard like India. Pretty much every test nation has had quick but wayward bowlers in their domestic comps who can't take any wickets because any FC batsman worth their salt can dispatch them.

 

[silentstriker]

The term "scientific" doesn't really make sense in this context.

And no, it wouldnt work because we simply have too many unknowns.

 

[Manee]

The term "scientific" doesn't really make sense in this context.

Idiocy on my part, am happy for any moderator to change the title to "Fully Statistical Selection".

 

[bagapath]

Fully Scientific XI

Einstein
Newton (c)
Watt
Franklin
Galileo
Pasteur (wk)
Darwin
Edison
Marconi
Pascal
Faraday

 

[unam]

Making selection fully statistical wouldn't work because there are so many things that selectors would considered before selecting a player.
- playing conditions in different countries differ hence a player who does well under one condition wouldn't necessarily do well in other conditions.
- other things would be to see if the player would fit in the team environment, so stuff like players attitude and his dedication towards the team and the how long would the player be able to play, his fitness level.
I feel selectors should use statistical as well as subjective approach

 

[honestbharani]

Napolean Einstein would lead the time, right?

 

[weldone]

The thing is that the way human brain thinks of a situation (common sense, or whatever we call it) follow an algorithm. Whenever a selection panel discussion happens, it follows an algorithm. It takes various inputs - age, test record, FC record, form in the last year or so, record in a certain situation or against a specific opposition - process through some if but only if for and other loops - and come up with the squad. However, it is more tricky than that. There are a few things which are in selectors' minds but possibly are very very difficult to be put in the form of an algorithm. Wicketkeeping skills, fielding skills, captaincy, political considerations (Ganguly-Chappell relationship for example), balance of the side (probably this one is less tough to be made a part of the algorithm than previous ones) etc...

So, for the time-being the answer is no. But I think if bagapath's XI tried hard, they could come up with a good algorithm