DoG's Test Innings and Bowling Performances. Updates thread.

[Days of Grace]

Starting a **** storm... but @Days of Grace can we please get S.Waugh, Borders and Kallis highest rating?

Of the innings I have checked so far:

SR Waugh	200	West Indies	648	Kingston	1995​	14.78​
AR Border	196	England	485	Lord's	1985​	14.52​
JH Kallis	162	England	664	Durban	2004​	14.12​

 

[Jarquis]

Do you mean by percentage of overs bowled? Should I include all bowlers in the innings? If I did that would probably make it worse for Laxman, since Australia used 9 bowlers in total.

You could argue that the Australian attack wasn't entirely suited for the conditions. They played three quicks and one spinners, whereas two specialist spinners would have been better. Was MacGill on that tour?

Yeah, I think by % of overs bowled would be a better reflection

 

[Starfighter]

If you wanted to be extremely tedious, you could actually go into the ball-by-ball record and find out the percentages of each bowler actually faced.

 

[Days of Grace]

Yeah, I think by % of overs bowled would be a better reflection

The problem with doing that is sometimes a bowler with a very low rating will bowl the most overs in the innings. Take England's attack (of those who bowled 10% or more of the innings) when Gavaskar scored 221 at the Oval in 1979:

	Rating at completion of match	Career peak rating	Adjusted rating	Overs bowled
P Willey	35​	71​	47​	43.5​
PH Edmonds	552​	654​	586​	38​
IT Botham	884​	911​	893​	29​
RGD Willis	759​	837​	785​	28​

 

[Jarquis]

The problem with doing that is sometimes a bowler with a very low rating will bowl the most overs in the innings. Take England's attack (of those who bowled 10% or more of the innings) when Gavaskar scored 221 at the Oval in 1979:

	Rating at completion of match	Career peak rating	Adjusted rating	Overs bowled
P Willey	35​	71​	47​	43.5​
PH Edmonds	552​	654​	586​	38​
IT Botham	884​	911​	893​	29​
RGD Willis	759​	837​	785​	28​

I don’t think that’s a problem though? It’s a more accurate representation of the attack faced

 

[Days of Grace]

I don’t think that’s a problem though? It’s a more accurate representation of the attack faced

Would you like it for all bowlers in an innings, or just the bowlers who bowled 10% or more of the innings? Because if I include Slater and Hayden's ratings for Laxman's 281 then it will probably make the Australian attack look worse.

 

[srbhkshk]

Would you like it for all bowlers in an innings, or just the bowlers who bowled 10% or more of the innings? Because if I include Slater and Hayden's ratings for Laxman's 281 then it will probably make the Australian attack look worse.

It's always gonna be a statistical problem, it's not really that the attack was weak - more that they batted the main bowlers into the ground and Australia was forced to bowl part-timers. Sort of a thing that always happens with very long innings.

Probable solution is to reduce the weight of part-timers if the innings was beyond X length, but retain it for less than X (or maybe a more gradual weight reduction). This probably biases the results in some other way though so I am not gonna claim it actually improves things.

 

[Days of Grace]

Or limit it to five bowlers maximum, since a "bowling attack" normally consists of four or five bowlers and any extra bowlers are considered as part-timers. We shouldn't punish batsmen who successfully saw off the main threats and were able to score runs off part-timers.

 

[Jarquis]

Would you like it for all bowlers in an innings, or just the bowlers who bowled 10% or more of the innings? Because if I include Slater and Hayden's ratings for Laxman's 281 then it will probably make the Australian attack look worse.

Which it was! If those blokes are bowling overs.

It’s never going to be perfect unless you can get to the point of the number of balls each player faced from each batsman. But having a bowler who bowled 18/178 count for the same as a bowler who bowled 45 doesn’t make sense to me.

But others may disagree

 

[Days of Grace]

If I weight the top 5 bowlers (McGrath, Kasprowicz, Warne, Gillespie, Mark Waugh) by the number of overs they bowled, then Australia's bowling attack rating rises to 650 from the original 608. This would be enough to take Laxman above Gooch and into 5th place but of course there would be changes to bowling attack ratings for all innings.

 

[honestbharani]

You should not penalize batsmen for playing well enough for long enough to force opponents to bowl their part timers or lesser bowlers longer.

For example, the ATG Windies side usually had no spinners. If a batsman batted for a long time against them, it's more than likely the likes of Richards and Logie bowled a lot of overs at them. It should actually mean the innings was that much better, not worse.

 

[Spark]

Yeah that surprises me a lot to be honest. In general I think the way these innings are rated systematically should reflect how we qualitatively rate innings watching them live or in the immediate aftermath and I don't think "that wasn't as good because the fielding captain was forced to turn to part-timers" is a thought any reasonable person has ever voiced here.

 

[TheJediBrah]

Which it was! If those blokes are bowling overs.

It’s never going to be perfect unless you can get to the point of the number of balls each player faced from each batsman. But having a bowler who bowled 18/178 count for the same as a bowler who bowled 45 doesn’t make sense to me.

But others may disagree

That's still not going to be perfect. It's a fun exercise but let's not pretend it's an accurate measure. Good bowlers have bad days, bowl worse in certain conditions, may be more tired or not at peak fitness at certain stages of games/series etc.

Just because an attack has McGrath and Warne in it, or 4 Windies quicks, doesn't mean it was bowling great on that day

 

[Days of Grace]

This is the top 20 if I weight the top 5 bowlers (those who bowled the most overs) for each innings by the number of overs they bowled. Laxman moves up to no.3 and Sehwag drops to no.7. Clem Hill and Jayawardene drop out and Jayasuriya and Bradman move into the top 10. This is actually a better looking top 20, IMO.

1​	MDKJ Perera	153*	South Africa	750	Durban	2019​	21.07​
2​	BC Lara	153*	Australia	751	Bridgetown	1999​	20.00​
3​	VVS Laxman	281	Australia	650	Kolkata	2001​	19.79​
4​	GA Gooch	154*	West Indies	764	Leeds	1991​	19.63​
5​	IT Botham	149*	Australia	604	Leeds	1981​	19.56​
6​	Saeed Anwar	188*	India	642	Kolkata	1999​	19.53​
7​	V Sehwag	201*	Sri Lanka	648	Galle	2008​	19.36​
8​	ST Jayasuriya	253	Pakistan	553	Faisalabad	2004​	19.26​
9​	KR Mayers	210*	Bangladesh	523	Chattogram	2021​	19.16​
10​	DG Bradman	270	England	612	Melbourne	1937​	19.11​
11​	BKG Mendis	176	Australia	624	Pallekele	2016​	18.87​
12​	LD Chandimal	162*	India	555	Galle	2015​	18.58​
13​	BA Stokes	135*	Australia	659	Leeds	2019​	18.36​
14​	DPMD Jayawardene	180	England	674	Galle	2012​	18.23​
15​	AC Gilchrist	149*	Pakistan	641	Hobart	1999​	18.16​
16​	GL Jessop	104	Australia	710	TIe Oval	1902​	18.12​
17​	Azhar Mahmood	132	South Africa	690	Durban	1998​	18.01​
18​	CG Greenidge	214*	England	557	Lord's	1984​	17.94​
19​	C Hill	188	England	581	Melbourne	1898​	17.81​
20​	KJ Hughes	100*	West Indies	818	Melbourne	1981​	17.75​

 

[Coronis]

Of the innings I have checked so far:

SR Waugh	200	West Indies	648	Kingston	1995​	14.78​
AR Border	196	England	485	Lord's	1985​	14.52​
JH Kallis	162	England	664	Durban	2004​	14.12​

See, Border just cashed in on **** attacks.

 

[Spark]

That's still not going to be perfect. It's a fun exercise but let's not pretend it's an accurate measure. Good bowlers have bad days, bowl worse in certain conditions, may be more tired or not at peak fitness at certain stages of games/series etc.

Just because an attack has McGrath and Warne in it, or 4 Windies quicks, doesn't mean it was bowling great on that day

Ultimately the only way to get vaguely close to "accurate" is to do what CricViz does and use actual ball-tracking data for every ball bowled in the game, which clearly you can't do historically.

 

[Victor Ian]

Using 5 bowlers also seems wrong. 80s Windies and 2000 ustralia were just 4 bowler attacks. The others were just filling in time. If you survived those 4 you had done well. Even with all-rounders, and fifth bowler attacks, if you survived the first four, you're on your way to something special. No way should mark Waugh be figured into Australia's ranking.

 

[Days of Grace]

Waugh’s not really “figured into” the ranking so much because I have weighted it by overs bowled.

5 bowlers seems right, especially when you weight it by overs bowled. If it’s a four man attack then you won’t usually see much of the fifth bowler anyway.

 

[ankitj]

Which it was! If those blokes are bowling overs.

It’s never going to be perfect unless you can get to the point of the number of balls each player faced from each batsman. But having a bowler who bowled 18/178 count for the same as a bowler who bowled 45 doesn’t make sense to me.

But others may disagree

Should only count top 4 bowlers in the attack (by rating or overs bowled). Reason Slater and Hayden were bowling is because Laxman (and Dravid) were batting so well and main bowlers got exhausted.

Edit: looks like this has already been suggested.

 

[Days of Grace]

It’s rare to have five quality bowlers, and the system should award it. Batsmen should be rewarded for scoring runs when there was no respite. You could argue that Australia got punished for not having an allrounder at Kolkata in 2001.