Suggestions for top 30 bowlers countdown

[venkyrenga]

Yes. What AVG tells is how many runs did a bowler concede to get a wicket combined with SR which tells how quickly did he get that wicket.
By getting wickets quickly, you prevent partnerships from developing.

By that same logic, a batsman who faces more deliveries helps to build partnerships. Would you be ok if we awarded points for balls faced in batsmen ranking?

Yeah but then again the amount of runs you concede for a wicket is directly proportionate to the amount of runs you concede in an over.

No, it is not directly proportionate. A bowler can have a high ER and a low average if his SR is low.

 

[Starfighter]

Strike rate is also perhaps more important in some eras than others. It didn't matter when 125 overs a day was the norm. I think overall strike rate is just as influenced by the batsmen as by the bowlers in any case.

 

[TheJediBrah]

By that same logic, a batsman who faces more deliveries helps to build partnerships. Would you be ok if we awarded points for balls faced in batsmen ranking?

Bro you keep saying that, it's not the same thing

 

[Engle]

By that same logic, a batsman who faces more deliveries helps to build partnerships. Would you be ok if we awarded points for balls faced in batsmen ranking?

No, I would not. You award points based on difficulty.
A batsman facing deliveries is not as difficult as a bowler getting wickets QUICKLY (SR).
Every player has to face deliveries. They could very well defend or leave aside deliveries and just hang around. Even tailenders can do this. And they have support from their batting partner.

A bowler on the other hand has to exert maximum effort to capture a wicket and to do so QUICKLY is even more onerous.

 

[Kirkut]

1. Marshall
2. Mcgrath

The rest play musical chairs.

 

[Fuller Pilch]

1. Marshall
2. Mcgrath

The rest play musical chairs.

Is it last 30 years like the batting? In that case it won't include Marshall and Hadlee both of whom would otherwise be top 7 with McGrath, Ambrose, Steyn, Murali, and Warne.

 

[venkyrenga]

Its hard to generalize if low-SR/high-ER is better than high-SR/low-ER. As it depends on the match conditions and team composition. But one thing I am clear about is this. Cricket is a game of ball vs bat. If we consider one metric as important for bowlers, mathematically the reciprocal of the same metric should be considered important for batsmen.

For now, I don't have a firm conviction either way. So I am doing away with both ER and SR. And stick with average & WPI as primary metrics. And as a secondary metric I will be rewarding bowlers for a high proportion of top/middle order wickets.

 

[Pap Finn Keighl]

Akram
Marshall
Hadlee
Steyn
Lillee
McGrath
Murali
Warne
Donald
Imran
Ambrose
Waquar
Holding
Garner
Roberts
Walsh

 

[Kirkut]

Is it last 30 years like the batting? In that case it won't include Marshall and Hadlee both of whom would otherwise be top 7 with McGrath, Ambrose, Steyn, Murali, and Warne.

Nope, I believe it's top 30 fast bowlers.

From the 90s and onwards, the top 5 are:

1. McGrath
2. Steyn
3. Ambrose
4. Akram
5. Donald

 

[TheJediBrah]

Akram
Marshall
Hadlee
Steyn
Lillee
McGrath
Murali
Warne
Donald
Imran
Ambrose
Waquar
Holding
Garner
Roberts
Walsh

Hmm. Judging by your last 287 posts, if my detective skills are on point I'm going to hazard the guess that you think Wasim Akram is good.

 

[Gob]

Hmm. Judging by your last 287 posts, if my detective skills are on point I'm going to hazard the guess that you think Wasim Akram is good.

Was he a crim

 

[venkyrenga]

I was just looking at the sheets and observed that bowlers with a good SR also have a higher WPI as they have taken wickets quickly. So they would anyways be rewarded without having to factor in SR.

And now the challenge I have is with awarding points for top and middle order wickets. Howstat has the career stats for wickets taken by 1-3, 4-7 & 8-11 but I can't give more weightage for 1-3 over 4-7 as most teams have their best batsman at no.4. It would be absurd to give more points for Malan's wicket than Root's.

So, does anyone know if it's possible to look up wickets taken by specific batting order? Any help is appreciated.

 

[TheJediBrah]

I don't think you need to differentiate between the top 7. Separating it into top 7 wickets and bottom 4 wickets is more than enough. Throughout history a lot of teams have had their best batsmen bat as low as 5.

 

[The Sean]

And now the challenge I have is with awarding points for top and middle order wickets. Howstat has the career stats for wickets taken by 1-3, 4-7 & 8-11 but I can't give more weightage for 1-3 over 4-7 as most teams have their best batsman at no.4. It would be absurd to give more points for Malan's wicket than Root's.

You've previously (and correctly) noted not to break this exercise down into individual examples but rather to take it as a statistical whole. By giving a higher weighting (and it might be only slightly higher) to top order over middle-order, you're not saying that Malan's wicket in and of itself is worth more than Root's, you're saying that as a historical and statistical whole there is more value to a bowler dismissing the top three over a bowler dismissing, for example, five-six-seven.

The option is just to keep them the same weighting, though that would be better if if was just for 1-3 and 4-6 as I feel including number 7 (notwithstanding outliers like Gilchrist) reduces that significantly.

 

[venkyrenga]

You've previously (and correctly) noted not to break this exercise down into individual examples but rather to take it as a statistical whole. By giving a higher weighting (and it might be only slightly higher) to top order over middle-order, you're not saying that Malan's wicket in and of itself is worth more than Root's, you're saying that as a historical and statistical whole there is more value to a bowler dismissing the top three over a bowler dismissing, for example, five-six-seven.

The option is just to keep them the same weighting, though that would be better if if was just for 1-3 and 4-6 as I feel including number 7 (notwithstanding outliers like Gilchrist) reduces that significantly.

Well, by individual examples I meant not to base it on the stats of specific bowlers as that will obviously lead to favoritism.

But historically most teams had their best bat at no.4 (Sachin, Lara, Kallis, etc). So it would be better to have a break down of 1-4, 5-7 and 8-11. There will still be some exceptions like Waugh and Chanderpaul but that is not a lot.

Otherwise I would just have to break it as 1-7 and 8-11. Probably with a weighting of 2:1.

 

[The Sean]

In previous eras a lot of teams traditionally had their best bat at three - and even Lara did his best work there, despite batting more often at four (not to mention Ponting, Sangakkara and Williamson even from recent years). I take your point though that the 1-3, 4-7 breakdown isn't ideal - and the reason I would say weight higher for 1-3 is because when you include number seven (and even number six) then historically that brings the average value of that group of wickets down. Again, you of course get blokes Sobers and Gilchrist who were phenomenal at six and seven, but historically the collective average of batsmen 1-3 is higher than 4-7.

I do sometimes see more specific position stats get quoted but sadly I don't know where they are drawn from, short of a complete searchable database.

I suppose my preference for weighting towards 1-3 (or 1-4) is also a personal thing - I have always though that a bowler who could regularly break the back of the top order and put immediate pressure on the rest of the line-up is a priceless asset.

 

[Bolo.]

You've previously (and correctly) noted not to break this exercise down into individual examples but rather to take it as a statistical whole. By giving a higher weighting (and it might be only slightly higher) to top order over middle-order, you're not saying that Malan's wicket in and of itself is worth more than Root's, you're saying that as a historical and statistical whole there is more value to a bowler dismissing the top three over a bowler dismissing, for example, five-six-seven.

The option is just to keep them the same weighting, though that would be better if if was just for 1-3 and 4-6 as I feel including number 7 (notwithstanding outliers like Gilchrist) reduces that significantly.

I think the weakness of 7 would be balanced out by the lesser weakness of the 2 openers (or rather the advantage the new ball provides). 7 shouldn't really be counted here, but it having it included does balance the top 6 out... 1-3 surely average less than 4-6 otherwise?

 

[TheJediBrah]

Just from an Aus-centric point of view we have often had our best bat at no. 5. Other than Ponting, Border and S.Waugh mostly batted 5. Clarke also batted mostly at 5 when he was the best in the team.

Well, by individual examples I meant not to base it on the stats of specific bowlers as that will obviously lead to favoritism.

But historically most teams had their best bat at no.4 (Sachin, Lara, Kallis, etc). So it would be better to have a break down of 1-4, 5-7 and 8-11. There will still be some exceptions like Waugh and Chanderpaul but that is not a lot.

Otherwise I would just have to break it as 1-7 and 8-11. Probably with a weighting of 2:1.

Sounds good to me

 

[The Sean]

I just did a quick statsguru - the all time historic batting averages are:

1-3: 37.01
4-7: 35.04 (4-6 is 37.28)
8-11: 15.27

So you're pretty much on the money - though even a 2:1 weighting is being generous to 8-11. Perhaps 2.5 to 1?

 

[venkyrenga]

Well, we are only doing the last 30 years so lets not go all time.

In the last 30 years,

1-3: 37.16
4-7: 36.67
Not much of a difference.

1-7: 36.9
8-11: 15.34
And yeah, 2:1 is generous. So may be 5:2.