Highest averages in Tests.
DG Bradman (AUS) 6996 5uns @ 99.94
SPD Smith (AUS) 5796 @ 62.32
AC Voges (AUS) 1485 @ 61.87
RG Pollock (SA) 2256 @ 60.97
GA Headley (WI) 2190 @ 60.83
H Sutcliffe (ENG) 4555 @ 60.73
Smith has 22 centuries and 21 half centuries.
Super Smith's stats.
-
Donny
- Posts: 80170
- Joined: Sun Aug 04, 2002 6:01 pm
- Location: Toonumbar NSW Australia
- Has liked: 57 times
- Been liked: 25 times
From Pies4shaw:
Smith has just become the first batsman ever to make at least 1,0000 runs over four consecutive calendar years at an average of over 60 in each year. His cumulative average over the 4 years from 2014 to 2017 is presently 75.25. His 22 centuries in that period are the most ever by a batsman in a 4-year period. This is heady stuff.
Ricky Ponting still has the crown for most runs in a 4-year period (he's the only batsman ever to get over 5,000). With 3 (realistically, probably only 2) further innings available this calendar year, Smith is a chance to catch Ponting. Before the start of play, he is on 4,816. Ponting's effort was an astonishing 5,077 (at 72.52). He needs 159 more runs this year to catch Hayden's 3rd place and 171 to pass both Hayden's efforts and take second place.
Smith has just become the first batsman ever to make at least 1,0000 runs over four consecutive calendar years at an average of over 60 in each year. His cumulative average over the 4 years from 2014 to 2017 is presently 75.25. His 22 centuries in that period are the most ever by a batsman in a 4-year period. This is heady stuff.
Ricky Ponting still has the crown for most runs in a 4-year period (he's the only batsman ever to get over 5,000). With 3 (realistically, probably only 2) further innings available this calendar year, Smith is a chance to catch Ponting. Before the start of play, he is on 4,816. Ponting's effort was an astonishing 5,077 (at 72.52). He needs 159 more runs this year to catch Hayden's 3rd place and 171 to pass both Hayden's efforts and take second place.
Donny.
It's a game. Enjoy it.
It's a game. Enjoy it.
Have the ICC given any info anywhere about how they come up with those ratings? I'd really want to know basic things like who came up with them, what goes into them, how they justify their formula, etc. before I gave them much credibility. I mean, it's not even clear whether that's supposed to be a linear scale.
In the old days, Deloittes came up with their ranking system, presumably to get publicity for themselves.
In the old days, Deloittes came up with their ranking system, presumably to get publicity for themselves.
Go to 'About Us' or their FAQs.-
Donny
- Posts: 80170
- Joined: Sun Aug 04, 2002 6:01 pm
- Location: Toonumbar NSW Australia
- Has liked: 57 times
- Been liked: 25 times
A player, say, in Smith's case, a batsman, will score more ratings points by making runs against a highly rated team.
England's attack had, until recently, 2 bowlers in the top 10 (Broad now 15th.), including Anderson, at #1.
If Australia played the Windies, their batsmen would score fewer ratings points for the same number of runs because the Windies highest ranking bowlers are at #20, 21 & 22.
Not sure what you mean about predicting what a batsman will score, K.
England's attack had, until recently, 2 bowlers in the top 10 (Broad now 15th.), including Anderson, at #1.
If Australia played the Windies, their batsmen would score fewer ratings points for the same number of runs because the Windies highest ranking bowlers are at #20, 21 & 22.
Not sure what you mean about predicting what a batsman will score, K.
Donny.
It's a game. Enjoy it.
It's a game. Enjoy it.
They are pretending to have precise quantitative info, i.e. concrete numbers, that rank the players --- and presumably not just that but the precise gap between them, e.g. between best and second best.
We all know that the degree of difficulty affects how meritorious a performance is. We also in qualitative terms have a fair idea of some of the factors that determine this degree of difficulty, though even this we could argue endlessly over.
The problem is not just to make sure you take into account all the factors that are important but how you should combine them to get one number. The only way to have any justification for the way you combine them is for them to describe data accurately. The important point is the data it describes should not be the data that you used to come up with your recipe. Hence prediction.
Otherwise, you're giving yourself the illusion of having something meaningful and objective, which in many ways is worse than just having a well-informed opinion, because in the latter case at least you're aware that there is subjectivity and it's not a precise and infallible judgement.
How much you trust the formula is determined by how well it makes predictions. Again, some predictions are easier than others. How much Smith will make today is a harder task than just how much he'll score on average in the next calendar year. The predictions should thus really come with confidence intervals, or margins or error, as they are also called. e.g. in weather reporting, they now do this: today there was an x% chance of rain, we're told (beforehand).
We all know that the degree of difficulty affects how meritorious a performance is. We also in qualitative terms have a fair idea of some of the factors that determine this degree of difficulty, though even this we could argue endlessly over.
The problem is not just to make sure you take into account all the factors that are important but how you should combine them to get one number. The only way to have any justification for the way you combine them is for them to describe data accurately. The important point is the data it describes should not be the data that you used to come up with your recipe. Hence prediction.
Otherwise, you're giving yourself the illusion of having something meaningful and objective, which in many ways is worse than just having a well-informed opinion, because in the latter case at least you're aware that there is subjectivity and it's not a precise and infallible judgement.
How much you trust the formula is determined by how well it makes predictions. Again, some predictions are easier than others. How much Smith will make today is a harder task than just how much he'll score on average in the next calendar year. The predictions should thus really come with confidence intervals, or margins or error, as they are also called. e.g. in weather reporting, they now do this: today there was an x% chance of rain, we're told (beforehand).
-
Donny
- Posts: 80170
- Joined: Sun Aug 04, 2002 6:01 pm
- Location: Toonumbar NSW Australia
- Has liked: 57 times
- Been liked: 25 times
Still not with you, re. predictions. This is not what the ratings are about. Predicting anything, in sport, is a near impossibility.
The ratings are simply results of their (ICC) given criteria - with all possible variations - taken into account. ie. Runs and/or wickets scored/taken against a highly rated team will get the player higher points.
The ratings are simply results of their (ICC) given criteria - with all possible variations - taken into account. ie. Runs and/or wickets scored/taken against a highly rated team will get the player higher points.
Donny.
It's a game. Enjoy it.
It's a game. Enjoy it.
| 1902 | 1903 | 1910 | 1917 | 1919 | 1927 | 1928 | 1929 | 1930 | 1935 | 1936 | 1953 | 1958 | 1990 | 2010 | 2023 | The point is that it's subjective --- arbitrary --- how precisely to take the variations into account. Your point about predicting things in sport exposes this arbitrariness. If they made accurate predictions, that would show that there was objectivity in the process.Donny wrote:Still not with you, re. predictions. This is not what the ratings are about. Predicting anything, in sport, is a near impossibility.
The ratings are simply results of their (ICC) given criteria - with all possible variations - taken into account. ie. Runs and/or wickets scored/taken against a highly rated team will get the player higher points.
(Actually, betting companies do make pretty good predictions. It's why it's very hard to beat the odds. Not predicting exact scores, etc., but more general things, such as the line.)
Put another way, you (i.e. your model) have to decide (as just one example) how many more points runs against higher opposition should get. Should Pujara scoring 245 against Bangladesh in Bangladesh (I'm just making up a hypothetical scenario) earn more or fewer points than Smith scoring 100 against England at the MCG? This will affect whether Smith ends up with a higher overall rating than Pujara.
That's the problem.