cricrate: new cricket ratings website

[hendrix]

Yeah Craig is weird. I'm not sure on the formula but I suspect he's getting some type of boost from away matches + contributions towards winning matches + beating teams with batsmen such as Younis and Chanderpaul, which probably boosts the quality factor of the opposition.

Viriya - does your away statistic also help spin bowlers? Because if you're a NZ spinner you should automatically do much better abroad than at home anyway.

 

[viriya]

Yeah Craig is weird. I'm not sure on the formula but I suspect he's getting some type of boost from away matches + contributions towards winning matches + beating teams with batsmen such as Younis and Chanderpaul, which probably boosts the quality factor of the opposition.

Viriya - does your away statistic also help spin bowlers? Because if you're a NZ spinner you should automatically do much better abroad than at home anyway.

Away factor is flat for any away game, it's very minor though. He crept into the top 10 just because of the last Test - there's not much between #10 and #20 right now. It's the downside of lowering the discount for lower players - you have them showing up at the top earlier than you expect (they also can drop down as fast as they went up though).

 

[marc71178]

Sarfraz at number 2 at the moment is perfectly justifiable. Good players exist from countries other than Australia, England and India.

Good player maybe but better than the list of people he's ranked ahead of is bollocks.

 

[viriya]

ODI update after a month or so:

- Faf jumps to #3, Raina drops out of top 10: cricrate | Current Top ODI Batsmen
- AB pips Viv to hit #1: cricrate | Best ODI Batting Careers
- Morkel rises to #1: cricrate | Current Top ODI Bowlers
- Kohli with #29: cricrate | Best/Worst ODI Fielding Performances

 

[viriya]

@viriya

Kim Hughes' 100 at mcg 1981 isn't in the top 100 either. What's that innings missing? It's one of the all time great match winning knocks on a terrible pitch against the greatest pace attack ever assembled. Just makes me feel that you still put too much emphasis on runs scored in am innings. Apart from the fact that the innings isn't a massive one, it has everything else going for it. What lets it down in your formula?

Btw - Kim Hughes' 100* is at #86 now:
cricrate | Best Test Batting Performances

Only thing that's not pushing it even higher is that it was a "small" hundred. It's not a low score in the context of the match, which is accounted for with the high % score factor. I think this innings definitely warrants being in the top 100 and #86 might be still too low (and I'm sure there are some cases that don't deserve to be higher), so I'm still looking into what factors I might be over/underrating.

 

[viriya]

Test update after a couple months:

- Joe Root climbs to #2 with career best rating: cricrate | Current Top Test Batsmen
- Root debuts in top 100 at #65: cricrate | Best Test Batting Careers
- Shoaib Malik with the #43 rated innings: cricrate | Best Test Batting Performances
- Yasir Shah hits #1 for the first time: cricrate | Current Top Test Bowlers
- Milinda Siriwardene with #33 rated all-round performance on debut, Malik at #53: cricrate | Best Test All-Round Performances
- Kaushal Silva jumps to #3: cricrate | Best/Worst Test Fielding Careers
- Ian Bell with #28 worst fielding performance: cricrate | Best/Worst Test Fielding Performances

 

[Daemon]

Is Malik's knock rated that highly because Broad and Anderson were bowling at him?

 

[viriya]

Is Malik's knock rated that highly because Broad and Anderson were bowling at him?

Partly, but they only bowled 30% of England's overs so overall the bowling attack was average/slightly above average. The biggest reasons it's ranked so high is that it's a big score, it's ~50% of the total, he came to bat at 5/1, and the next highest score was just 107.

I do think it's slightly overrated though knowing that Broad/Anderson is not that good in UAE conditions, but it's very hard to account for that systematically. This kind of example makes me think I should increase the significance of the "bowling quality" factor a bit more actually.

 

[Howe_zat]

If you are counting Bell dropping him as a very poor fielding performance, shouldn't Malk's innings be rated lower as it was scored against that fielding performance?

I think neither really, but both would be better than one.

 

[OverratedSanity]

Btw - Kim Hughes' 100* is at #86 now:
cricrate | Best Test Batting Performances

Only thing that's not pushing it even higher is that it was a "small" hundred. It's not a low score in the context of the match, which is accounted for with the high % score factor. I think this innings definitely warrants being in the top 100 and #86 might be still too low (and I'm sure there are some cases that don't deserve to be higher), so I'm still looking into what factors I might be over/underrating.

Partly, but they only bowled 30% of England's overs so overall the bowling attack was average/slightly above average. The biggest reasons it's ranked so high is that it's a big score, it's ~50% of the total, he came to bat at 5/1, and the next highest score was just 107.

I do think it's slightly overrated though knowing that Broad/Anderson is not that good in UAE conditions, but it's very hard to account for that systematically. This kind of example makes me think I should increase the significance of the "bowling quality" factor a bit more actually.

These two posts make it obvious to me that you're still massively overrating the value of the size of a batsman 's innings.

 

[Prince EWS]

These two posts maje it obvious to me that you're still massively overrating the value of the size of a batsman 's innings.

It's hard not to do this statistically without ignoring the difference between making 20 and making a ton though if you're using linear models.

Perhaps incorporating the findings of the third graph I posted in this post would help a bit in this instance, smashing the linear model.

 

[viriya]

These two posts make it obvious to me that you're still massively overrating the value of the size of a batsman 's innings.

It's hard not to do this statistically without ignoring the difference between making 20 and making a ton though if you're using linear models.

Perhaps incorporating the findings of the third graph I posted in this post would help a bit in this instance, smashing the linear model.

Yea the reason runs scored is a big factor is because if it isn't, the resulting career rating list would be skewed by players who made 20-30s at 5 down a lot but don't do much else.

I just browsed your post PEWS and it's very interesting - i'll look into the thread completely and check whether I can use something along those lines.

 

[viriya]

I redid what PEWS did with just dismissed scores and based on that I came up with an "equivalence table" that had three tiers:
0-40: linear
41-200: discounted, with 50 => ~49, 100 => ~91, 150 => ~127, 200 => ~157
201-400: heavily discounted, with 250 => ~191, 300 => ~212, 400 => 223

0-40 is linear for ratings purposes, in terms of odds changes going from 0 to 10 is the most significant, but I wanted a cut-off to avoid overrated low scores.

http://www.cricrate.com/test/batting/performances.php

Some changes to top 100 innings:
- Lara 153* #14 => #6
- Sanga 156* #20 => #11
- Kim Hughes 100* #87 => #46
- Gooch 154* #51 => #35

The biggest change is the heavy discounting of 200+ innings and as a result <200 innings showing up more:
# of innings >= 300 went from 19 => 10
# of innings < 200 went from 38 => 56
Even 2 sub-100 innings show up.

Surprisingly, all this doesn't result in any significant changes in the career batsmen list, which is favorable - fixed innings ratings without affecting career ratings negatively.
http://www.cricrate.com/test/batting/career.php

 

[viriya]

If you are counting Bell dropping him as a very poor fielding performance, shouldn't Malk's innings be rated lower as it was scored against that fielding performance?

I think neither really, but both would be better than one.

Fielding is a separate exercise, and I've avoided using it as a factor when rating batting/bowling performances because there is a lack of data pre-2005 so it would make it harder to compare players between eras. The Bell "worst fielding performance" is a little extreme tbf, since this is since 2005, not really "all-time" like it suggests.

 

[weldone]

Why is Chanderpaul so ridiculously high? Is being part of a weak West Indies batting lineup for a long-time indirectly helping his rating?

 

[viriya]

Why is Chanderpaul so ridiculously high? Is being part of a weak West Indies batting lineup for a long-time indirectly helping his rating?

That and he benefits from coming into bat at bad situations a lot because of the WI top order weaknesses. I'd agree that he's overrated a bit but I think it's partly due to his reputation for selfishness and unattractive style.. if he had made those runs while looking good people won't mind him being rated as high I'd guess.

 

[G.I.Joe]

Why is Chanderpaul so ridiculously high?

woah, is that an accusation?

 

[viriya]

Test update:

- Cook and Warner return to the top 10, Faf drops out, Kane returns to the top 5:
cricrate | Current Top Test Batsmen

- Yasir and Ashwin reach career-best, Starc reaches top 10:
cricrate | Current Top Test Bowlers

 

[weldone]

woah, is that an accusation?

Do you seriously think that Chanderpaul is better than Richards, Dravid, Ponting, Border, Hammond, Headley, Hobbs, Gavaskar, Hutton, Chappell?

 

[viriya]

Do you seriously think that Chanderpaul is better than Richards, Dravid, Ponting, Border, Hammond, Headley, Hobbs, Gavaskar, Hutton, Chappell?

I think he meant high as in that purple kush, the good stuff.. On Chanderpaul, I don't agree with his position in the list either, but it's not obvious to me what he is overrated in as of yet. I do think he is generally underrated though (not top 10 worthy but top 25 imo).