Adjusting the batting average based on actual quality

[srbhkshk]

Here's a few more retired bowlers of the modern era with careers of importance, outside the very top bowlers list. The regression model extrapolation holds up better than I expected:

Ian Botham:
Regression Adjusted Average = 27.03 Delta = -1.374

Anil Kumble:
Regression Adjusted Average = 27.86 Delta = -1.789

Jacques Kallis:
Regression Adjusted Average = 33.91 Delta = 1.260

Did you ever put out these numbers for the active batting and bowling greats? Or does the system not support them atm?

 

[shortpitched713]

I put no stock in these kind of ratings exercises at all, but your issues here look like "shortpitched713 problems" rather than actual problems with the system. You think Philander and Kaneria are better than they are so when they don't end up as high as you want it's because of "nationalism"

Those are just two examples. I tend to just think, underresearched players at that ranking level + flagging voter interest created a scenario where the big 3 fan base countries ended up being way overrepresented at the lower end of the poll result.

 

[shortpitched713]

Did you ever put out these numbers for the active batting and bowling greats? Or does the system not support them atm?

It would require more work. Which I'm not really able to do at the moment, or even for the foreseeable future.

My methodology was laid out pretty clearly, for anyone who wants to continue that work.

 

[howitzer]

“Let’s just take a random guess and adjust someone’s averages to what I think they should be” - topic of this thread

Yeah. In the spirit of this I'm adjusting Woakes' batting average to 99.95.

 

[Red_Ink_Squid]

Yeah. In the spirit of this I'm adjusting Woakes' batting average to 99.95.

You'r underrating him imo.

 

[trundler]

Lyon got in disgracefully late. Very near Ashwin, if not better than him.

Herath not making it at all was disgraceful, much better bowler than Jadeja.

 

[ma1978]

Herath not making it at all was disgraceful, much better bowler than Jadeja.

if he was much better, how come his performances are so much worse

 

[trundler]

mid he was much better, how come his performances are so much worse

Incorrect

 

[Nintendo]

Kim Jong-un saves North Korean innings with top score of 807* — The Betoota Advocate

30 September, 2015. 15:03 ERROL PARKER | Editor-at-large | Contact THE SUPREME LEADER of the Democratic People’s Republic of Korea (DPRK) has paved the way for a North Korean victory on day two of their test match against Cuba in Pyongyang yesterday. The 32-year-old scored a quick-fire 807 not...

www.betootaadvocate.com

Kim jong un made his infamous 807* against Cuba and that game got test status somehow. Massive minnow bully.

 

[TheJediBrah]

Herath not making it at all was disgraceful

Guess they couldn't fit him in. Narrow doors

 

[ankitj]

Kim Jong-un saves North Korean innings with top score of 807* — The Betoota Advocate

30 September, 2015. 15:03 ERROL PARKER | Editor-at-large | Contact THE SUPREME LEADER of the Democratic People’s Republic of Korea (DPRK) has paved the way for a North Korean victory on day two of their test match against Cuba in Pyongyang yesterday. The 32-year-old scored a quick-fire 807 not...

www.betootaadvocate.com

Kim jong un made his infamous 807* against Cuba and that game got test status somehow. Massive minnow bully.

Parody? Can't be sure hence asking.

 

[RossTaylorsBox]

Extremely parody

 

[TheJediBrah]

Can't be sure

how

 

[ankitj]

how

It's obviously fake news but North Koreans could be believing this in earnest so it may not be parody. Check this:

North Korean Soccer Coach Talks to 'Dear Leader' Via Invisible Phone

North Korea returned to the World Cup for first time in 44 years to play Brazil. Coach Kim Jong-Hun says he talks to the country's reclusive dicactor via an invisible cellphone Kim Jong Il invented. Hundreds of Chinese were rented to sit in the stands and root for North Korea.

abcnews.go.com

 

[RossTaylorsBox]

I don't think making up cricket scores is part of the propaganda regime over there.

 

[honestbharani]

Parody? Can't be sure hence asking.

3 Idiots thing.

 

[Coronis]

I don't think making up cricket scores is part of the propaganda regime over there.

Beliefs like these are exactly why they’re succeeding

 

[shortpitched713]

Have gone ahead and done this for the bowlers too. The methodology here was to set ICC Ranking average as numerator, and multiply by MVP model score (which usefully adds in a nice away record consideration as well) and place the era adjusted bowling average as the denominator. The correlation is not as strong as for the batsmen ( just shy of .7 ), but it's still not bad. Inclusion of spinners only reduced that correlation by a negligible .01, so I feel comfortable including them, although they end up with very different residuals. I have to also account for the fact that all things are not equal ( i.e. spinners are used more in stock roles regardless of quality). So while these bowlers are sorted in the order of the original regression value, I made an adjustment based on the original residuals to slightly increase spinner averages (and very, very slightly decrease seamer ones). The final results look a bit more "realistic" for both bowler categories.

Bowler	Regression Adjusted Average	Actual Average	Delta
Glenn Mcgrath	21.42793​	21.64​	-0.21207​
Muttiah Muralitharan	23.27961​	22.72​	0.559614​
Dale Steyn	21.77292​	22.95​	-1.17708​
Curtly Ambrose	21.83116​	20.99​	0.841163​
Richard Hadlee	22.14589​	22.29​	-0.14411​
Malcolm Marshall	22.36784​	20.94​	1.427839​
Allan Donald	22.42278​	22.25​	0.172784​
Joel Garner	22.79065​	20.97​	1.820645​
Shaun Pollock	22.99114​	23.11​	-0.11886​
Imran Khan	23.56562​	22.81​	0.755615​
Wasim Akram	23.566​	23.62​	-0.054​
Shane Warne	25.25136​	25.41​	-0.15864​
Vernon Philander	23.60542​	22.32​	1.285421​
Dennis Lillee	23.74988​	23.92​	-0.17012​
Waqar Younis	23.8438​	23.56​	0.283803​
			0.3408​

The average delta ended up being slightly positive ( meaning more of these players came from the somewhat more bowling friendly eras of the overall modern era).

Also, it turns out you can't square the circle of batting friendly conditions for SA bats, and bowling friendly ones for SA bowlers, unless you're a bowler like Philander who has a significantly worse record away (therein being hurt by the MVP model consideration). So Steyn ends up coming out as the only bowler with a negative delta magnitude over 1 .

			ICC/MVP	Career Impact Model		Multiplied
Murali			7432292.962​	67459.11158​		50137588.02​
McGrath			8173927.159​	55142.4876​		45073067.7​
Hadlee			6640173.385​	59726.78749​		39659622.46​
Steyn			7390804.515​	51991.62017​		38425990.11​
Marshall			6235483.868​	57977.13897​		36151551.48​
Ambrose			7267354.325​	40830.00099​		29672608.43​
Donald			6139761.699​	41925.2685​		25741115.77​
Imran			4487644.032​	45693.82073​		20505760.19​

I've included the raw values from the ICC/MVP model that I used to calculate the initial regression for this. Now comparing it to the raw values of the more impact oriented model I had made in the below thread, and then multiplied together to get a new value for each bowler:

Lillee/Trueman/Akram/Donald - how would you rank them?

Yeah many of those on the list were worldclass earlier than listed It's just their best averages over a 6 year period. It doesn't necessarily represent their best 6 years. And almost none of these guys would have had a peak we would define as exactly 6 years.

www.cricketweb.net

Think it's interesting to see how the models sort of differ in what they emphasize, but in the end have the same top few bowlers. Might make another regression using this, but cba right now, so leaving this as a placeholder.

 

[kevinw]

Isn't this what points rankings are meant to take into account. A guy who played mostly on roads and averages 50 is probably not as good as a guy who played mostly in England or NZ and averages 45....but rankings take into account those vagaries.

 

[Coronis]

Isn't this what points rankings are meant to take into account. A guy who played mostly on roads and averages 50 is probably not as good as a guy who played mostly in England or NZ and averages 45....but rankings take into account those vagaries.

In theory yes. In practice they do far from a perfect job.