Cribbage's Standardised Test Averages (UPDATED November 2018 - posts 753-755)

[Prince EWS]

Fazal > Waqar

Haha, this was caused by the bug it seems.

 

[Prince EWS]

• I never thought I'd see a statistical analysis built around averages and longevity that ranked Warne ahead of Murali, at least not without just straight up excluding some games. Given the algorithm does address the two main things Warne point to -- Murali playing on more helpful pitches and disproportionately playing weaker opposition -- I suppose it makes some sense, but I'm still pretty surprised we've got Warne at #2 now.
  
• Cowie h4x is bigger than ever.
  
• We now have Nissar and Amar h4x to join him, I guess owing to their dead home wickets.

So were these (although Warne is still higher than I'd have expected -- 5th, still ahead of Murali).

 

[Prince EWS]

I've fixed it in the doc now, will edit my posts above later.

https://docs.google.com/spreadsheet...RJyxqnl8WU_sZvK5SbDe109OA/edit#gid=2016770811

 

[aussie tragic]

Sorry, can't get past the Lillee rating. He has 355 wkts @ 23.92 in 70 tests vs Flintoff 219 wkts @ 33.35 in 78 tests but they're basically ranked the same

 

[aussie tragic]

Actually Lillee, Thomson and Walker all treated harshly here. Could it be that because they were one of the best pace attacks Australia ever had, the resulting Australia pace friendly rating has unfairly impacted on their individual rating?

 

[Prince EWS]

Actually Lillee, Thomson and Walker all treated harshly here. Could it be that because they were one of the best pace attacks Australia ever had, the resulting Australia pace friendly rating has unfairly impacted on their individual rating?

Nah, it doesn't rate a country's pitches pace-friendly just because the bowlers are good. It compares the home bowlers do with how the quicks from the same team do away, and then it also compares how the quicks of visiting teams do with how they did against the same team's batting lineup at home. So your quicks can average 12 at home and still not have your pitches rated as friendly if they also only average 13 away, or touring bowlers get rekt. As an example, right now the Australian pitches are rated worse than the Bangladesh pitches for fast bowling (dead last) even though they've had Starc, Haze, Cummins, Patto etc bowling for them on them. You can see how multiple good bowlers didn't cause the great West Indians to suddenly have bad standardised averages too.

I don't even know that "home pitch flatness" is the reason for those players not being rated particularly highly; I haven't checked that out. They all lost a couple of years due to WSC which will hurt them in the longevity stakes, and some of Walker's and Lillee's best bowling was in that period too. Plus, regardless of pitch analysis away cricket is deemed harder than home cricket, and Lillee played a disproportionate amount of stuff at home (roughly twice the amount of his away games I believe). Walker and Thomson not having particularly spectacular base averages didn't help them either. There are lots of factors playing against all of them I think.

 

[Prince EWS]

Let's say Jan 1954. Whenever you can.

Spin-friendliness:
New Zealand - 1.51
Pakistan - 1.28
England - 1.11
India - 1.06
South Africa - 0.90
West Indies - 0.82
Australia - 0.73

Pace-friendliness:
New Zealand - 1.44
England - 1.16
Australia - 1.09
South Africa - 1.07
Pakistan - 1.06
India - 0.77
West Indies - 0.74

Must've been some low scoring ones in NZ, haha.

 

[Prince EWS]

Chanderpaul's batting ranking is a bit of a put-off for me personally.

On this, I suspect that if I looked into the methodology the article below dives into for dealing with not outs, Chanderpaul would be more accurately rated (and possibly Barrington as well).

Inbox: How to account for not-outs more accurately when assessing batsmen - ESPN

That might be my next little project on this. I'll give it a few months though as I think spamming constant formula changes in a short period takes a bit of away from it all. You're not going to get invested in reading a rating list if you know it'll probably change drastically tomorrow despite no matches having occurred.

 

[aussie tragic]

Thanks for detailed response. Still seems strange that every Aussie pace bowler from the 70s and 80s have such increased averages compared to other nations bowlers from the same period.

Lillee, McDermott, Thomson, Hughes, Walker, Alderman, Hogg and Lawson all have significantly higher averages ( I couldn't find one that didn't)

 

[Prince EWS]

Thanks for detailed response. Still seems strange that every Aussie pace bowler from the 70s and 80s have such increased averages compared to other nations bowlers from the same period.

Lillee, McDermott, Thomson, Hughes, Walker, Alderman, Hogg and Lawson all have significantly higher averages ( I couldn't find one that didn't)

Yeah, there could be a number of explanations for bowlers of the same type in the same team to all have an increase. Off the top of my head:
* a bowler-friendly era in general
* home pitches that suited their bowling style, relative to other pitches at the time
* their team playing a disproportionate amount of home games
* their team playing a disproportionate amount of games against weak opposition
* their team playing a disproportionate amount of their away games in a country with pitches that suit their bowling style

I'm definitely not saying those things were all true in the 70s/80s for Australian quicks (some of them obviously don't apply here), but there are a number of reasons you could find that sort of trend without it being any sort of flaw. For example, Indian and Sri Lanka spinners this decade have all received an increase due to their home pitches, and that's perfectly fine -- that's the sort of thing this is designed for, and it's more or less the same trend as you've identified there (even if the explanation isn't quite so obvious). So I don't think it's strange or even necessarily some sort of flaw that that's happened. I'll go through it some time in the next couple of days and let you know exactly why I think it has done that (it's probably largely the pitches, with a special case for Lillee given he played so much of his career at home).

 

[aussie tragic]

Yeah, there could be a number of explanations for bowlers of the same type in the same team to all have an increase. Off the top of my head:
* their team playing a disproportionate amount of games against weak opposition

I had to lol at this one...obviously you never watched Australia play cricket in the 80's

 

[trundler]

Australia were the weak opposition in the 80s.

 

[Prince EWS]

I had to lol at this one...obviously you never watched Australia play cricket in the 80's

No no no, I was just listing all the possibilities I could think that could make all players from one team (in a general sense) in one area suffer a hit. Obviously some of them don't apply to this particular case at all.

 

[weldone]

PEWS, what kind of formula do you use to arrive at the final value? Personally I use the following for ODI.

(Avg of current rating - constant1) * (longevity - constant2)^constant3

In your case obviously avg of current rating can be replaced by standardised averages. Have you ever tried anything like this? Or is your formula drastically different?

 

[Prince EWS]

BatValue = (((StAvg / 36.1) ^ 2) * (BatLongevity ^ 0.5)) ^ (1 / 2)
BwlValue = (((31.77 / StBwlAvg) ^ 2) * (BowlLongevity ^ 0.5)) ^ (1 / 2)

 

[Prince EWS]

So basically I do it very similarly to you, except I use a division instead of a subtraction when comparing the standardised averages to the predetermined baseline.

 

[andmark]

I love that with the standardised averages, Don Bradman has could be flimisily argued to have been a better allrounder than Jacques Kallis.

 

[Prince EWS]

I love that with the standardised averages, Don Bradman has could be flimisily argued to have been a better allrounder than Jacques Kallis.

Allrounders are interesting. If you take the square root of the multiples of bat value and bowl value, to find a sort of "rounded allrounder value", this is the top 25:

1. GS Sobers (WI) - 2.46
2. JH Kallis (SA) - 2.33
3. Imran Khan (Pak) - 2.20
4. TL Goddard (SA) - 2.08
5. WR Hammond (Eng) - 2.03
6. KR Miller (Aus) - 2.00
7. SM Pollock (SA) - 1.91
8. N Kapil Dev (Ind) - 1.89
9. Sir RJ Hadlee (NZ) - 1.88
10. FMM Worrell (WI) - 1.84
11. MA Noble (Aus) - 1.83
12. Mushtaq Mohammad (Pak) - 1.80
13. W Rhodes (Eng) - 1.79
14. R Benaud (Aus) - 1.77
15. GA Faulkner (SA) - 1.76
16. AK Davidson (Aus) - 1.75
17. Shakib Al Hasan (Ban) - 1.73
18. G Giffen (Aus) - 1.73
19. WW Armstrong (Aus) - 1.71
20. W Bates (Eng) - 1.71
21. JR Reid (NZ) - 1.70
22. IT Botham (Eng) - 1.69
23. Asif Iqbal (Pak) - 1.69
24. KD Walters (Aus) - 1.67
25. SR Waugh (Aus) - 1.66

Seems to be slightly bat-skewed, which doesn't surprise me much when you look at the values at the tops of the batting and bowling lists (ie. the best batsmen are slightly higher than the best bowlers, even ignoring Bradman). This seems to be because great batsmen tend to play on a bit longer than great bowlers and therefore have more longevity.

 

[andmark]

Allrounders are interesting. If you take the square of the multiples of bat value and bowl value, to find a sort of "rounded allrounder value", this is the top 25:

1. GS Sobers (WI) - 2.46
2. JH Kallis (SA) - 2.33
3. Imran Khan (Pak) - 2.20
4. TL Goddard (SA) - 2.08
5. WR Hammond (Eng) - 2.03
6. KR Miller (Aus) - 2.00
7. SM Pollock (SA) - 1.91
8. N Kapil Dev (Ind) - 1.89
9. Sir RJ Hadlee (NZ) - 1.88
10. FMM Worrell (WI) - 1.84
11. MA Noble (Aus) - 1.83
12. Mushtaq Mohammad (Pak) - 1.80
13. W Rhodes (Eng) - 1.79
14. R Benaud (Aus) - 1.77
15. GA Faulkner (SA) - 1.76
16. AK Davidson (Aus) - 1.75
17. Shakib Al Hasan (Ban) - 1.73
18. G Giffen (Aus) - 1.73
19. WW Armstrong (Aus) - 1.71
20. W Bates (Eng) - 1.71
21. JR Reid (NZ) - 1.70
22. IT Botham (Eng) - 1.69
23. Asif Iqbal (Pak) - 1.69
24. KD Walters (Aus) - 1.67
25. SR Waugh (Aus) - 1.66

Seems to be slightly bat-skewed, which doesn't surprise me much when you look at the values at the tops of the batting and bowling lists (ie. the best batsmen are slightly higher than the best bowlers, even ignoring Bradman). This seems to be because great batsmen tend to play on a bit longer than great bowlers and therefore have more longevity.

I wouldn't have guessed Goddard would've been so high. Wally Hammond is a bit of a weird one in there as well.

 

[trundler]

Is there a statistical method which doesn't overvalue batting all rounders?