How valuable is that wicket!

[ankitj]

This is a really interesting point that I hadn't thought of. It puts Marshall's greatness into even clearer perspective.

Yeah. It's interesting how often people say stats don't tell the whole story but the truly great ones really stand out. Marshall always comes out on top no matter by how little. That's why when host of a recent cricinfo talk said Marshall is a bit of a question mark in the ATW XI, I wanted to smash something into his face.

 

[ankitj]

There's a guy in cricinfo "It Figures" blogs section (Anantha Narayanan) who does a lot of analysis of this type. I've read a few of his articles but not any lately.

Actually I have a friend in Cricinfo. He does commentary often. Let me see if I can get in through to him.

 

[ankitj]

Harold Larwood and Hugh Tayfield, two bowlers I should have included in the original analysis:

[B]Bowler		Wickets	Bowl Ave	Ave value Discount Factor[/B]
Harold Larwood	78	28.35	34.22	82.8%
Hugh Tayfield	170	25.91	28.78 	90.0%

Interesting that Larwood has highest average value of wickets of all bowlers I have analysed. Getting Bradman 5 times helped! Remove Bradman and that value comes down by almost 7! Anyway, I will now* rate Larwood a little higher than I so far did.

EDIT: I will now (and not not) rate Larwood higher is what I meant to say

 

[vcs]

Harold Larwood and Hugh Tayfield, two bowlers I should have included in the original analysis:

[B]Bowler		Wickets	Bowl Ave	Ave value Discount Factor[/B]
Harold Larwood	78	28.35	34.22	82.8%
Hugh Tayfield	170	25.91	28.78 	90.0%

Interesting that Larwood has a highest average value of wickets of all bowlers I have analysed. Getting Bradman 5 times helped! Remove Bradman and that value comes down by almost 7! Anyway, I will not rate Larwood a little higher than I so far did.

Haha, the Don. Evil stat skewer.

 

[Migara]

In other word Larwood was a "Don" specialist!

 

[smash84]

Yeah. It's interesting how often people say stats don't tell the whole story but the truly great ones really stand out. Marshall always comes out on top no matter by how little. That's why when host of a recent cricinfo talk said Marshall is a bit of a question mark in the ATW XI, I wanted to smash something into his face.

In other word Larwood was a "Don" specialist!

Now that is a nice term

 

[Ikki]

Another flaw of the list is that it does not actually say how much a certain bowler conceded to a certain batsman. It just looks at their overall average and the average when they were taken out by a certain bowler.

For example, a player may have conceded most of his runs in an inning to the best batsmen but got the wicket of a taileender who'd made his runs elsewhere and built a score over his own batting average. Therefore the average of his wicket being taken by that bowler is high, without the bowler actually conceding runs. Since there is little movement discount wise it is going to certainly affect him negatively in this analysis. In reality, it could even be a great batsman who pasted your opening bowlers but is done by first change bowlers who'll be similarly disadvantages as the above explains.

Also, in a side where the bowling attack is strong, the big wickets will be shared around. The biggest gains you can make in this exercise is by taking a big wicket early. If you take a batsman's wicket for 5 runs, and he averages 50 that is some difference. But take a tailender's wicket on 1 when his average is 15 and there is very little movement.

 

[Migara]

The biggest gains you can make in this exercise is by taking a big wicket early.

No it's not. It does not matter much whether batsman is 0 or 235. If he averages 40 that will be only counted as 40. Not even that runs made in that innings would matter, becaue what is taken is the career average of the batsman.

 

[Ikki]

No it's not. It does not matter much whether batsman is 0 or 235. If he averages 40 that will be only counted as 40. Not even that runs made in that innings would matter, becaue what is taken is the career average of the batsman.

What? It completely matters. That is the whole point of the discussion: how well certain bowlers were at restricting batsmen with respect to the average of the batsmen.

 

[Migara]

What? It completely matters. That is the whole point of the discussion: how well certain bowlers were at restricting batsmen with respect to the average of the batsmen.

In cricketing terms and looking for the perfect model, yes it is important. But the scope is THIS model by Ankit. Yes, it has it's flaws I agree, but bloody impressive piece of work.

 

[Ikki]

So you're wrong. Again: The biggest gains you can make in this exercise is by taking a big wicket early.

 

[Migara]

So you're wrong. Again: The biggest gains you can make in this exercise is by taking a big wicket early.

Care to explain?

 

[Migara]

Value of Wickets

What the averages don't tell you however is how valuable were the wickets taken by a bowler. A wicket of Harbhajan Singh is not as valuable as that of Sachin Tendulkar. That fact is howerver hidden in the overall bowling averages. In order to put value to every wicket taken, therefore, we can use the career batting average of the batsman dismissed as a proxy. Therefore value of Tendulkar's wicket is about 57 and that of Harbhajan only 16.

That's how I read it.

 

[Ikki]

Care to explain?

If two bowlers take a wicket each; both only conceding 5 runs to their batsman, but one of the batsmen averages 5 and the other 50:

Bowler A averages 5; the average value of his wicket is 5; which makes his discount factor 100%.
Bowler B averages 5; the average value of his wicket is 50; which makes his discount factor 10%.

 

[Migara]

Probably you mean early in the spell? Early wickets better results non-shocker TBH.

 

[Ikki]

Technically, it matters not an iota when in a spell it is taken; only the runs conceded and the average of the batsman taken.

 

[vcs]

So you're wrong. Again: The biggest gains you can make in this exercise is by taking a big wicket early.

Actually, it doesn't matter when you get a wicket because only a batsman's final career batting average matters while deciding the value of his wicket. It's obviously a flaw but there are no perfect statistical models anyway.

Basically Peter George would have got the same value in the Bangalore Test match irrespective of whether he had dismissed Tendulkar on 0 or 214.

 

[Uppercut]

Another flaw of the list is that it does not actually say how much a certain bowler conceded to a certain batsman. It just looks at their overall average and the average when they were taken out by a certain bowler.

That's not a flaw at all! The entire point of the analysis is that an underlying problem in conventional averages is that all wickets don't have the same value. All runs do have the same value so it doesn't matter whether they're conceded to someone good or not.

I honestly think the OP's done a fantastic job here, not getting enough credit. The point of the analysis was to partially correct, a particular problem with conventional bowling averages, and thereby arrive at a more accurate figure. It's not supposed to be the be-all and end-all, so I don't see why there's been so much criticism based on the flaws that he hasn't addressed. All of which, obviously, were there for the conventional figures anyway.

 

[ankitj]

Basically Peter George would have got the same value in the Bangalore Test match irrespective of whether he had dismissed Tendulkar on 0 or 214.

This. Either I am not understanding Ikki's doubt or he has misunderstood the methodology. But what vcs said here is accurate description of the method.

 

[vcs]

Well, it is moreso IMO. An overall average is much more general and is looked at much more generally. This is specifically looking at the batsmen you took and if you took them whilst they've made few runs then it really makes you look like a champ. But if the batsman's already made 40 and you've only bowled 2 overs and conceded 3, and took his wicket...then frankly you've taken him in no time but it does not reflect that way at all.

In terms of general average though, every run you concede is accounted for, for every wicket you take. You don't get other people's run concessions sneaking into your overall average. They're the runs you conceded.

I don't understand your point. How do other people's run concessions sneak into your overall average? They only impact your discount factor positively, because your own bowling average is what the average value is measured against. If you are hardly conceding anything, that will be reflected in your own bowling average and you will get a good discount factor.

That was the whole point we were making earlier on while contrasting Steyn and Marshall's discount factors. In the '80s there were a lot of world class bowlers, bringing down batsman averages overall and cutting into each others' discount factors. Whereas nowadays there aren't many bowlers around at the level of Steyn, hence he picks up fantastic value for dismissing the likes of Gambhir and Samaraweera, who are considered more valuable wickets than Gavaskar in this exercise.