cricrate: new cricket ratings website

[Spark]

That makes sense, but I would have the linear decay starting much earlier than 200 personally (albeit slower). Cribb and I have talked about this a fair bit.

 

[viriya]

That makes sense, but I would have the linear decay starting much earlier than 200 personally (albeit slower). Cribb and I have talked about this a fair bit.

I went with 200 after looking at the correlation between runs and win % as I recall. The (win + draw) % tended to increase with runs up to around 200. So it's based on actual historical numbers.

 

[Spark]

I went with 200 after looking at the correlation between runs and win % as I recall. The (win + draw) % tended to increase with runs up to around 200. So it's based on actual historical numbers.

Yeah, fair enough.

===

Smitteh is already #4 of all time for Australian batsmen, which is kind of scary tbh.

 

[viriya]

Smitteh is already #4 of all time for Australian batsmen, which is kind of scary tbh.

Yea forgot to mention that.. although I'm sure him being above Greg Chappell and Steve Waugh career-wise isn't going to be universally agreed upon.

The numbers are insane though. It's a bit like Ashwin.. seems premature but the numbers are just too good to ignore.

 

[Chewie]

Henry Nicholls jumped 40 spots to 14 lol

 

[viriya]

Henry Nicholls jumped 40 spots to 14 lol

A combination of a couple of very good innings and just getting past the early career 20 innings penalty. His rating is super-volatile though - if he gets out for a duck next innings he'll drop multiple spots.

 

[Chewie]

If he gets another century will he go top 10 then?

 

[viriya]

If he gets another century will he go top 10 then?

Depends on how good it is.. It gets harder to climb the higher he goes.

 

[viriya]

That makes sense, but I would have the linear decay starting much earlier than 200 personally (albeit slower). Cribb and I have talked about this a fair bit.

Actually found a conversation on this in late-2015 here (mid page):
http://www.cricketweb.net/forum/cricket-chat/62989-cricrate-new-cricket-ratings-website-31.html

TLDR on what I do now:

Equivalence table of 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

 

[Victor Ian]

How does the all time ranking work again? (sorry if you've answered this already)

I'm asking because Steve Smith is no 38 whereas on the cricinfo ratings he is no5. I recall that you had a nice factor for longevity. How many more times does Steve have to bat before he gets the longevity bonuses kick in and take him higher?

 

[Spark]

How does the all time ranking work again? (sorry if you've answered this already)

I'm asking because Steve Smith is no 38 whereas on the cricinfo ratings he is no5. I recall that you had a nice factor for longevity. How many more times does Steve have to bat before he gets the longevity bonuses kick in and take him higher?

The ICC rankings are a current ranking, and their all-time ranking is all-time peak ranking.

 

[viriya]

How does the all time ranking work again? (sorry if you've answered this already)

I'm asking because Steve Smith is no 38 whereas on the cricinfo ratings he is no5. I recall that you had a nice factor for longevity. How many more times does Steve have to bat before he gets the longevity bonuses kick in and take him higher?

What spark said.

You can get the equivalent to th ICC all-time peak rating if you toggle from Current to all-time in the current rating page:
cricrate | Current Ratings - Test Batting

 

[Victor Ian]

Thanks Spark

To Viriya: A real tiny idea, but on your excellent graph, could you paint it from 5 to 1 instead of 1 to 5. When they are dead close it is hard to tell. For instance, currently Kane looks above Quentin because his green line is painted over top of Quentin's red line and I imagine the line width covers the difference of 3.

 

[viriya]

Thanks Spark

To Viriya: A real tiny idea, but on your excellent graph, could you paint it from 5 to 1 instead of 1 to 5. When they are dead close it is hard to tell. For instance, currently Kane looks above Quentin because his green line is painted over top of Quentin's red line and I imagine the line width covers the difference of 3.

Good point - I'll look to change. It's harder to decipher because De Kock has made the graph's range larger than usual.

I was thinking that it might make sense to extend the graph to the last 2 years too.

 

[vcs]

You guys need to stop comparing De Kock to others'... size doesn't matter as they say.

 

[viriya]

Would it be better to consider Laxman's entry point as 1/-225, rather than 1/50?

Then factor exit point into it too -- an innings that takes you from 1/-225 to 9/-100 is still wholly irrelevant, whereas 1/-225 to 5/150 is hugely valuable.

Exactly. Because this isn't followed in viriya's methodology, Laxman's 167, which while an excellent knock but in the end meaningless, is rated higher than the far greater and more meaningful 281. The reason McCullum's 302 and VVS's 281 are amazing is because, as I said before, they dragged the team from a hopeless position all the way to a dominant one. It wasn't a high quality but futile hit and giggle session which was never going to change anything (ala Astle's 222 or the VVS 167)

I modified the Match status factor to adjust to how good the comeback was in the 3rd innings, also for 4th innings chases/defenses (applies to bowlers as well).
cricrate | Performance Ratings - Test Batting
Laxman's 281 #267 to #76
McCullum's 302 #84 to #19

Also, I calculate the significance of the innings not only for the team innings, but for the entire match. Therefore, a double century is seen as the significant innings of the match and not just a large score in a large team innings. Likewise, a 60* in a team score of 100 but in a match aggregate of 1500/35 loses a lot of points compared to if you just rated the significance compared to his teammates in the same team innings.

Therefore, with my method, one can see how Laxman's 281 would be rated significantly higher than his 167 which, incidentally, doesn't make my top 100.

Hope this makes sense.

I added a Pitch quality factor for Tests which takes other innings of the Test into account as well as the innings of the match (the later the harder run scoring is).

I changed the dismissed batsmen rating factor for bowling performances to adjust based on how many runs the batsman had scored already (If Tendulkar has already hit a ton, getting him out is not as impressive as getting him for a duck). Details here:
cricrate | Methodology - Test Batting

Also adjusted the weights of all factors to make more sense.

Some of the major changes from this update (pretty significant):

Batting performances:
cricrate | Performance Ratings - Test Batting
Kim Hughes's 100* jumps 39 places to #65
VVS Laxman's 281*jumps 190 places to #71
Graham Gooch's 154* jumps 57 places to #27
Donald Bradman's 270*jumps 33 places to #7
Brendon McCullum's 302*jumps 65 places to #19
Dinesh Chandimal's 162* jumps 157 places to #48
Hanif Mohammad's 337*jumps 56 places to #11
AB de Villiers's 278* drops 32 places to #45
Brian Lara's 153* jumps 11 places to #6
Derek Randall's 174*jumps 422 places to #49
Mark Waugh's 116*jumps 69 places to #9
Andrew Stoddart's 173*jumps 376 places to #56
Gordon Greenidge's 134*jumps 97 places to #16
Michael Clarke's 151*jumps 82 places to #14
Donald Bradman's 334*drops 25 places to #30
Donald Bradman's 299* drops 33 places to #40
Damien Martyn's 161*jumps 8 places to #2
Sunil Gavaskar's 221*jumps 47 places to #12
Dudley Nourse's 231*jumps 141 places to #37
Kusal Mendis's 176*jumps 15 places to #4
Kane Williamson's 161* drops 72 places to #93
Saeed Anwar's 188* jumps 27 places to #8
Kumar Sangakkara's 211*jumps 101 places to #32
Ian Redpath's 159* jumps 144 places to #46

Bowling performances:
cricrate | Performance Ratings - Test Bowling
Jermaine Lawson's 6/3 jumps 21 places to #4
Bert Ironmonger's 6/18 jumps 168 places to #33
J.J. Ferris's 7/37 jumps 212 places to #57
Imran Khan's 8/60 jumps 80 places to #24
Clarrie Grimmett's 7/40 jumps 142 places to #47

Career ratings changes:
cricrate | Career Ratings - Test Batting
cricrate | Career Ratings - Test Bowling

Viv Richards jumps 3 places to #6
Yasir Shah drops 45 places to #88
Bishan Bedi drops 27 places to #54
Shane Warne drops 3 places to #7
Imran Khan jumps 4 places to #6
Anil Kumble drops 5 places to #13
Hugh Tayfield drops 13 places to #34
Andy Roberts jumps 18 places to #32
Hugh Trumble drops 24 places to #71
Lance Gibbs drops 14 places to #42

I really like how the performance ratings looks now (especially for batting).

Appreciate any feedback!

 

[Athlai]

I've never noticed just how highly cricrate has Fleming placed. What are the reasons for that?

 

[viriya]

I've never noticed just how highly cricrate has Fleming placed. What are the reasons for that?

I haven't checked him specifically, but I would guess the same reason why Chanderpaul, Flower and De Silva are higher up than expected. They get credit for being the sole performer in a weaker team most likely (The percentage of total, point of entry, wickets at crease, support factors all help them naturally).

I do think they are overrated - haven't found an effective way to handle that at the innings level while still rating universally accepted great innings highly.

 

[Athlai]

Well Fleming wasn't that far ahead of his peers, compared to the other examples anyway. He on occasion played spectacular innings and was very consistent in getting a score (though had a terrible 50-100 conversion rate).

 

[viriya]

Well Fleming wasn't that far ahead of his peers, compared to the other examples anyway. He on occasion played spectacular innings and was very consistent in getting a score (though had a terrible 50-100 conversion rate).

Yea that particularly helps in the way some of the factors work. Actually this inspired me try a change where I only give Point of Entry/Wickets at Crease factor credit if the player makes a significant score. Currently it scales with your score and it really shouldn't. Let's see how that changes things.