1. No-one is saying that stats don't matter (certainly not me). What I'm saying is that your evidence for the assertion that Hadlee just pips Lillee just isn't good enough. Just because on one criteria Hadlee wins, doesn't mean he wins overall at all. Why? Because like any other statistician, I want to know whether the difference is due to factors other than random chance. The bare numbers don't explain anything and your misuse of the data shows up YOUR lack of understanding of how to use statistics in analysing a situation other than that which is easily predictable.
First, its not just ONE criteria. For the umpteenth time.
I have compared his performance against the best batting lineup of his time- same faced by Hadlee.
SJS made an attempt to downgrade the WI of the 80s but he forgot that while Viv was a lesser force in the 80s and Lloyd + Kalli were gone, Greenidge was at his peak, Gomes was there and Richardson was at his very best, making up for Lloyd's departure with the bat...ie, the difference was small.

Like any statistician, you are getting bent outta shape regarding the degree not the verdict.
Even one criteria is enough to pronounce superiority( assuming all others balance each other out) and nomatter how miniscule, the superiority is there.
Therefore, its not a question of if X is superior to Y but by how much.
And you can have a varience in every statistical analysis and varience is there for practicality purposes in attesting to a product or practicality between time management and risk(this i am sure you are very aware of, given the kind of statistical data you deal with).
But that does NOT negate the fact that there is a statistical superiority between products A and B, nomatter how insnignificant.

Applying the same logic to human systems is fraught with danger. Your conclusions are premature. My general theme hasn't been "You are proveable wrong" about Hadlee and Lillee because that's not the case. What I'm saying is that you have not proven yourself right and if you read back, you'll see I'm not questioning the logic but the VERACITY of the conclusion.
this isnt a quintissential human system like propensity of a rape case or the propensity of a user to buy a certain product. At worst, its a low level computer simulation, since there are several parameters ( ie figures) that are easily quantifiable and not arbitarily tied to anything.

As such, based on the evidence presented so far and accessable, the conclusion is that Lillee is inferior to Hadlee.
Ofcourse, the degree is in question.
But unless you can quantify accurately( and not arbitarily) the differential factor in pitches at WACA and Auckland or the mood of a batsman, that data cannot be considered.
It is a given that anything involving human beings is not 100% accurate system but you still use it, dont ya ? Your modelling of the propensity of a rape happening at a given location is not 100% accurate but it is considered.
That aspect needs to be kept in mind.

Why? You have to do the significance tests BEFORE you can determine whether the difference between population means IS significant. You can't just look at them and say 'they are significant', simple as that. This has been the crux of what I've been saying.
that is valid in either of the two scenarios:

1. You are comparing the varience with experimental error
2. You have an arbitary number assigned to 'significance' based on the requirement of the job (say your boss deciedes that if there is less than 5% chance of a rape happening in this nighbourhood, there wont be a patrol car there or that the company you work for specifies that your bolts can only vary by .01 mm at maximum-thus the 0.01 is the significance factor).

Since i am not experimenting but merely crunching numbers, there is no experimental error.
And since there is no arbitary significance condition assigned to cricket, any difference is significant.

Ie what i am saying is there is clear indication that the overall dataset affects Lillee more negetively than Hadlee. You are arguing the significance of those datasets, ie if the impact is significant enough.
Well, how DO YOU deciede significance in this case ?

Say you do a statistical analysis on Lillee and Hadlee.... and then proceeded to give them an aggregate score that takes into account all pluses and minusses.
you come up with something like this:

Lillee: 81.112

That is a 0.112 difference
Now how is it determined if that 0.112 difference is significant enough or not, since you do not have an arbitary governing number.

2. In the 2003/4 series vs England, Lara's batting return read 500 runs at 83.33.

Statistically brilliant but, in reality, a failure as 400 of those runs came from a self-indulgent innings played when the series was lost. When the chips were down, he had 6 completed innings for 100 runs.

Statistics are useful, but not definitive.

3. First, its not just ONE criteria. For the umpteenth time.
I know that. But you said IF there was that'd be enough which I believe to be wrong.

Like any statistician, you are getting bent outta shape regarding the degree not the verdict.
That's because every verdict has a degree of error (not experimental error) surrounding it and those degrees of error instruct as to how credible the verdict is. A verdict in isolation is never enough for anyone. You need to qualify that verdict with an understanding as to its significance. Again, this is what I've been saying all along.

Even one criteria is enough to pronounce superiority( assuming all others balance each other out) and nomatter how miniscule, the superiority is there.

But that does NOT negate the fact that there is a statistical superiority between products A and B, nomatter how insnignificant.
I never said it did. But whether that statistical superiority is significant (i.e. due to random chance or an indication of a genuine superiority) is the question one always needs to ask. Saying Hadlee is statistically superior to Lillee and then, by extension, saying Hadlee was a superior bowler are two vastly different things and is a HUGE leap to make without real analysis which you or I haven't done as yet.

this isnt a quintissential human system like propensity of a rape case or the propensity of a user to buy a certain product. At worst, its a low level computer simulation, since there are several parameters ( ie figures) that are easily quantifiable and not arbitarily tied to anything.
You can quantify anything within sexual assault scenarios or any crime type for that matter in the same way as measuring stats for cricket. You just need to identify the criteria you want to measure, analyse whether they accurately represent what you're trying to say and then use them accordingly. It's all been done before by Kim Rossmo and Jerry Ratcliffe. The concept of geographic profiling was invented by Rossmo too and is in fact a service offered by departments like this;

http://www.rcmp.ca/techops/geog_prof_e.htm

Doing what I do is equally as quantifiable as any analysis of cricket stats.

As such, based on the evidence presented so far and accessable, the conclusion is that Lillee is inferior to Hadlee.
And that it's eminantly questionable whether you can credibly draw such a conclusion from the evidence provided. I say you can't and could probably prove it.

But unless you can quantify accurately( and not arbitarily) the differential factor in pitches at WACA and Auckland or the mood of a batsman, that data cannot be considered.
Yet again, this is a routine problem for a base statistics graduate. One must just be able to justify the criteria is all.

For example, I was asked to be able to quantify a score as far as matching the Modus Operandi (how the offence is done) of any sexual offence to an offender's known MO. However, which MO factors should rate as higher priority than others? What's the relative importance, as far as tying an offender to an offence(s), of the car they drove vs knowing the hair colour of the offender? How is each factor scored? A perfectly justifiable way of doing it was to survey several profesional detectives as to what THEY thought were the important factors were and how much. You then analyse that data like you would any survey. Suddenly you can rank different MO factors and come up with a useful score, matching an offender to an offence based on MO. Again, this is perfectly justifiable and is not out of the ordinary in any way.

Why can't one rank or assign value to aspects of pitches or batsman mood in the same way once criteria is agreed upon? Short answer, of course you can! It's an every-day occurrance in fact.

1. You are comparing the varience with experimental error
Not talking experimental error here but variability about the mean.

2. You have an arbitary number assigned to 'significance' based on the requirement of the job (say your boss deciedes that if there is less than 5% chance of a rape happening in this nighbourhood, there wont be a patrol car there or that the company you work for specifies that your bolts can only vary by .01 mm at maximum-thus the 0.01 is the significance factor).
Of course. You can't test the difference between two populations without having some degree of significance. All tested stats have an end result along with a 'within a 95%/90%/66% degree of confidence'. Again, routine.

Since i am not experimenting but merely crunching numbers, there is no experimental error.And since there is no arbitary significance condition assigned to cricket, any difference is significant.
If you want to do analysis of the numbers, you have to assign significance and associated confidence intervals to the values. Otherwise you're not doing real analysis which impacts on the validity of any conclusions. It's unavoidable in real analysis. If you want to understand differences between any populations, you have to have to make decisions as to the significance of any numbers you deal with (are the numbers the levels they are due to random chance or another factor?). Again, this is unavoidable in analysis. The numbers in isolation tell you nothing.

Ie what i am saying is there is clear indication that the overall dataset affects Lillee more negetively than Hadlee. You are arguing the significance of those datasets, ie if the impact is significant enough.
Well, how DO YOU deciede significance in this case ?
Easy. Read back and see the null hypothesis I put forward. Run a one-way ANOVA (best to use a non-parametric test) to determine whether the difference in Lillee's and Hadlee's averages (or any other factor) show significance or are down to random chance.

Say you do a statistical analysis on Lillee and Hadlee.... and then proceeded to give them an aggregate score that takes into account all pluses and minusses.
you come up with something like this:

Lillee: 81.112

That is a 0.112 difference
Now how is it determined if that 0.112 difference is significant enough or not, since you do not have an arbitary governing number.
Why would you need an arbitrary governing number? We're not testing the significance of Lillee and Hadlee's stats against some other number but against each other. Again, easily done.

If you want to determine who is 'great' and the 'greatness index' was 81.210, for example then what you're saying makes sense but that's a totally different question and isn't within the context we're discussing. Comparing one population against another doesn't need that arbitrary governing number. I think I'm seeing where the misunderstanding is coming from now.

4. Statistics are useful, but not definitive
Exactly. Stats are a tool to be used in an explanation, not an explanation in of themselves. That's why the results of any analysis has to be qualified with an analysis of the veracity of the numbers.

5. Lillee and Hadlee : Some tributes to two great bowlers.
....from those who know better than us mortals I presume

If Dennis Lillee isnt the greatest fast bowler then he must come very close to being so.
David Gower in Heroes and Contemporaries

When ever or wherever fast bowling is discussed, the name Dennis Lillee is bound to be mentioned, for he unquestionably is one of the all-time great quick bowlers in the game. As far as i am concerned, he is the last of what I would describe as the classic fast bowlers.
Fred Trueman in Larwood to Lillee

Dennis has been blessed with all the attributes required to reach the peak of his particular art form. He is a fast bowling perfection.
Fred Trueman in Larwood to Lillee

Which brings me to Richard Hadlee whom I believe to be the finest and fastest of the New Zealand bowlers I have seen. There is no doubt that he is among the most respected and feared quick bowlers in the world.
Trevor Bailey in Larwood to Lillee

There have been many outstanding fast bowlers.....but none has quite matched Lillee who I consider to be the greatest of them all.
Clive Lloyd in Living For Cricket discussing the fast bowlers he has seen.

Dennis Lillee the Australian legend and our best ever fast bowler.
Mark Waugh in The Entertainers while explaining his choices for the dream team of the 20th century

I rate Dennis Lillee as the king of fast bowlers.
Doug Walters in The Entertainers.while explaining his choices for the dream team of the 20th century

Note : Mark Waugh's team includes Lillee and Miller as the fast bowling choices while Walters includes Lillee, Marshall and Akram.

Hadlee was very handy with the bat in his day. He carried the New Zealand attack single handedly and was an excellent performer on all types of wickets.
Walters in The Entertainers

He](Dennis Lillee) fights all the time, never shirks the responsibility. He came to Pakistan in 1980 and took just three wickets in the three tests but he bowled his heart out, trying every tactic and keeping up his aggression. Thats the sign of fast bowling greatness.

Unlike Andy Roberts, Holding and John Snow, great bowlers who needed to be cajoled. Dennis Lillee never gave up but made things happen by sheer will
power.

From the first day I saw him in 1972, he's been the best bowler of my era.

Imran Khan in his biography Imran talking about Lilleee

PS. I am sorry the tributes to Lillee are more in number than those to Hadlee but thats what I could find.

6. Originally Posted by SJS
Lillee and Hadlee : Some tributes to two great bowlers.
....from those who know better than us mortals I presume

If Dennis Lillee isnt the greatest fast bowler then he must come very close to being so.
David Gower in Heroes and Contemporaries

When ever or wherever fast bowling is discussed, the name Dennis Lillee is bound to be mentioned, for he unquestionably is one of the all-time great quick bowlers in the game. As far as i am concerned, he is the last of what I would describe as the classic fast bowlers.
Fred Trueman in Larwood to Lillee

Dennis has been blessed with all the attributes required to reach the peak of his particular art form. He is a fast bowling perfection.
Fred Trueman in Larwood to Lillee

Which brings me to Richard Hadlee whom I believe to be the finest and fastest of the New Zealand bowlers I have seen. There is no doubt that he is among the most respected and feared quick bowlers in the world.
Trevor Bailey in Larwood to Lillee

There have been many outstanding fast bowlers.....but none has quite matched Lillee who I consider to be the greatest of them all.
Clive Lloyd in Living For Cricket discussing the fast bowlers he has seen.

Dennis Lillee the Australian legend and our best ever fast bowler.
Mark Waugh in The Entertainers while explaining his choices for the dream team of the 20th century

I rate Dennis Lillee as the king of fast bowlers.
Doug Walters in The Entertainers.while explaining his choices for the dream team of the 20th century

Note : Mark Waugh's team includes Lillee and Miller as the fast bowling choices while Walters includes Lillee, Marshall and Akram.

Hadlee was very handy with the bat in his day. He carried the New Zealand attack single handedly and was an excellent performer on all types of wickets.
Walters in The Entertainers

He](Dennis Lillee) fights all the time, never shirks the responsibility. He came to Pakistan in 1980 and took just three wickets in the three tests but he bowled his heart out, trying every tactic and keeping up his aggression. Thats the sign of fast bowling greatness.

Unlike Andy Roberts, Holding and John Snow, great bowlers who needed to be cajoled. Dennis Lillee never gave up but made things happen by sheer will
power.

From the first day I saw him in 1972, he's been the best bowler of my era.

Imran Khan in his biography Imran talking about Lilleee

PS. I am sorry the tributes to Lillee are more in number than those to Hadlee but thats what I could find.
I'm sure their would be hundreds more....You'd need to go through 1000's of cricket biographies/autobios as well as other literature like mags etc etc.

Its simply comes down to one mans opinion vs the other.

Also it was unclear whether the starter of this thread was comparing them simply as "bowlers" or as "cricketers" as Hadlee was obviously classed as an all-rounder. If he mean't the latter, then its not so easy to compare.

7. Originally Posted by zinzan12
I'm sure their would be hundreds more....You'd need to go through 1000's of cricket biographies/autobios as well as other literature like mags etc etc.

Its simply comes down to one mans opinion vs the other.

Also it was unclear whether the starter of this thread was comparing them simply as "bowlers" or as "cricketers" as Hadlee was obviously classed as an all-rounder. If he mean't the latter, then its not so easy to compare.
Yes. Sure there are hundreds. But I would really be surprised if there are many calling Hadlee the best or the greates of his time while there are umpteen calling Lillee the best of his time and since their times overlapped a fair amount, that would say something.

8. Originally Posted by social
In the 2003/4 series vs England, Lara's batting return read 500 runs at 83.33.

Statistically brilliant but, in reality, a failure as 400 of those runs came from a self-indulgent innings played when the series was lost. When the chips were down, he had 6 completed innings for 100 runs.

Statistics are useful, but not definitive.
Nicely put. His value to the "team" was minimal in that series.

9. Originally Posted by social
Statistics are useful, but not definitive.
Absolutely spot on.

10. Originally Posted by SJS
Yes. Sure there are hundreds. But I would really be surprised if there are many calling Hadlee the best or the greates of his time while there are umpteen calling Lillee the best of his time and since their times overlapped a fair amount, that would say something.
I'm not arguing for either player in this one. IMO its too close to call.

I know Hadlee himself rated Lillee as the best and most complete fast bowler when he was playing. So he'd no doubt vote Lillee.

IMO Hadlee emulated Lillee superbly and statistically overtook him.

Still its too close to call IMO

11. Originally Posted by zinzan12
I'm not arguing for either player in this one. IMO its too close to call.

I know Hadlee himself rated Lillee as the best and most complete fast bowler when he was playing. So he'd no doubt vote Lillee.

IMO Hadlee emulated Lillee superbly and statistically overtook him.

Still its too close to call IMO
No arguements with that whatsoever.

12. So. Another great theory that we have goes like this .....

If a bowler gets only 3 wickets in 3 tests at 101 runs each then, irrespective of his record in another 66 tests, it must be concluded that he could not bowl well in the conditions presented in these matches. That he was not a good enough bowler to master those conditions.

Question : Isnt it possible that he just ran into a bad patch, an unlucky phase, a small time when edges did not go to hand or batsmen got beaten but didnt snick or whatever ?

Answer :NO. Satistically blah blah blah. It proves he could not bowl in these conditions.

Question: But surely if he had bowled more on those wickets he might have improved his record, dont you thin ?

Answer : NO. Statistically blah..blah..blah It is proven what I say.

I asked myself, what would we have done if the failure/statistically-poor-performance had come in the same conditions where the great successes had come? Whom would we have blamed then? Perhaps, it was a specific batsman or two in that team who made it impossible for this fallible bowler to bowl like other infallible greats. Interesting.

There once was a bowler who, in three successive tests, got a sum total of 5 wickets for 273 runs. An astronomical average of 54.6. Clearly a useless bowler and one who could not do anything against the opposition he faced in those conditions.

The we found that before these three games, he had taken 80 wickets at 14.2 runs each. And after these three games he proceded to take another 16 wickets at 16.2 runs each. More surprisingly, the opposition was the same and the conditions were the same.

Something funny here, no ?

This bowler, one Mr CTB Turner, holds many records including taking his 101 test wickets in just 17 tests at 16.5 runs each !!

Could it be just one of those quirks of sport that he did not get the same success in those three games ???

Some theories leak the moment you put something weightier than air in them.

Page 15 of 15 First ... 5131415