The flaw in batting averages

[pasag]

Came across this and was wondering what our resident statisticians thought of it:

The flaw in batting averages - Sanchit Maini & Sumit Narayanan

 

[Lillian Thomson]

There is no flaw at all in batting averages. It's a pure fact. If you start introducing other factors it's no longer a measure of what actually happened.

 

[Isura]

There is no flaw at all in batting averages. It's a pure fact. If you start introducing other factors it's no longer a measure of what actually happened.

On what basis? Being a fact doesn't imply that it isn't a flawed measure. How about if we only look at total wickets for bowlers. It is a fact, but it is still a flawed way to look at bowling performance.

 

[Top_Cat]

There is no flaw at all in batting averages. It's a pure fact. If you start introducing other factors it's no longer a measure of what actually happened.

Any mean, being a calculation with associated standard deviation, by definition, is not a fact. not a measure at all of what actually happened; number of runs scored is and that's basically it.

I agree with the writer of the article, averages are an incomplete measure of batting skill and results. Not sure what would be or indeed if you can come up with a single measure.

 

[slugger]

under the system now.. it has bevan at the top of the pile.. but strangely when this guy does it based on his formular bevan slips too 4th and players such as tendulkar, ponting and richards move above him. and id be suprised to fine anyone on cw that though bevan should be higher than those 3. and also being 4th on the list for bevan still demonstrates how good of a player he was.

 

[kanenzl]

this article doesn't take into account that batsmen coming in later in the innings more often than not have to take a higher amount of risks than someone coming in earlier in the innings. therefore if they are not out after being in for 10 overs accelerating the scoring rate they rightfully shouldnt have this included in their number of innings.

of course there are plenty of valid reasons for the other side of this debate too

 

[KiWiNiNjA]

Averages are just statistics. They are by no means an absolute measure of skill.

There is only a problem when people measure a players skill on average alone, not taking in variables such as the situation, opposition etc

There is no need to change anything, people just need to realise that averages aren't everything, they need to be put in context, and they can sometimes be misleading.

 

[bagapath]

well, richards, tendulkar and ponting are certainly the top 3 one-day batsmen of all time with bevan at no 4 for sure. if the new statistical method comes up with this result it could be right. but this alone cant make one method better than the other. stats are just that, numbers. greatness is much much more than that. and trying to define it in your own terms is great fun....

 

[Shaggy Alfresco]

Batting averages are flawed because they don't take into account dropped catches

 

[Neil Pickup]

It's an interesting analysis, and makes sense for what it's trying to do.

 

[Laurrz]

Urgh!!! @ people who over analyse with stats...

 

[nightprowler10]

It's an interesting analysis, and makes sense for what it's trying to do.

Agreed.

 

[Burgey]

Batting averages are flawed because they don't take into account dropped catches

Nor do series results. Owing to Pietersen's dropping by Warne in the 5th test at the Oval in 05, England didn't deserve to win the Ashes.

 

[Top_Cat]

Urgh!!! @ people who over analyse with stats...

Same, especially when they're either misused or too much emphasis is placed on data which is either wrong, misleading or incomplete. Misuse of stats is par for the course most of the time. 13.4487% of people know that.

 

[bond21]

pfft bean counters.

 

[Lillian Thomson]

On what basis? Being a fact doesn't imply that it isn't a flawed measure. How about if we only look at total wickets for bowlers. It is a fact, but it is still a flawed way to look at bowling performance.

A few deluted souls like to believe that they can prove than X is better than Y by producing a Venn Diagram which is fair enough if it makes them happy but as far as official records of stats are concerned nothing should be included other than the facts.

Any mean, being a calculation with associated standard deviation, by definition, is not a fact. not a measure at all of what actually happened; number of runs scored is and that's basically it.

What utter tosh. A batting average is the number of runs scored divided by the number times dismissed, both are irrefutable facts and a 100% accurate record of what actually happened.

 

[Richard]

Averages are just statistics. They are by no means an absolute measure of skill.

There is only a problem when people measure a players skill on average alone, not taking in variables such as the situation, opposition etc

There is no need to change anything, people just need to realise that averages aren't everything, they need to be put in context, and they can sometimes be misleading.

Pretty much. Averages aren't supposed to be an "average score per innings". It's rare for a batsman to make the same score twice in a row, never mind more than that.

It's simply a fairer way of expressing total runs. The object of batting is to score runs without getting out, so therefore any attempt to treat not-outs as different to what they are currently will garner no interest from me.

 

[Goughy]

It depends what you want to suggest.

If you want to show runs per dismissal then current averages do exactly that.

If you want them to be a closer appoximation to relative ability then they can be played with in a number of different ways. Firstly take runouts out. They are a dismissal but dont relate to batting ability.

The one I liked was using the average of completed scores over a not out as the value. ie 87* would be worth the average of all competed innings over 87.

TBH, averages do their job but there is context and situations that must be taken into account.

Its a flawed system but a decent one, though I dont mind people trying to create their own more thorough method.

 

[GGG]

Yeah there are a lot of factors that sway someones averages, take Brendon McCullum for example, if he was Australian he wouldn't make the team till he was 25+ same with Taylor and a host of other New Zealanders, West Indians and Proberbly a few English. So these players quite often struggled for the first few years making there average lower than if they were from a stonger cricketing nation like Australia, India or South Africa where they would have to ply their trade in a domestic comp for years. Say Martin Crowe didn't play till he was 22, his average from 22 - he retired is 50, I know he had 3 years of international experience but there are plenty of players in the Australian domestic comp that are better than many international players.

 

[Top_Cat]

What utter tosh. A batting average is the number of runs scored divided by the number times dismissed, both are irrefutable facts and a 100% accurate record of what actually happened.

It's funny how I've seen you admonish people for speaking out of their bounds of expertise. You should considering following your own advice.

One more time, with feeling; number of runs is the data, average describes the data.