Batsmen Impact Per Series

[Prince EWS]

Also, have you considered making the tiers more fluid?

eg. letting someone avergaing 45 get 2.5 instead of 2, and someone averaging 42 getting 2.2.

I think this would be probably more faithful to the idea of rating consistent series performances, as opposed to just arbitrary gatekeeping.

 

[vcs]

We can keep making it better. For example, more weightage for longer series.

Until it eventually becomes.. the batting average.

 

[NUFAN]

Can I post the full list now that you've posted the points?

Yes, go for it. Well send to me first please.

The tier issue is fine.

60 is an excellent series average, but it's a below average performance for a few bats and a barely above average one for just about everyone being considered. It's also not a number that you could say would impact on result with any degree of confidence, so it's pretty low

Below average performance for who? I totally get what you are saying that the highest tier could have been higher. Perhaps I was brought up watching more 4 or 5 Test per series, compared to the now more common 2 or 3 Tests and if you get to 60 for the series, you've certainly had a big impact.

 

[TheJediBrah]

We can keep making it better. For example, more weightage for longer series.

Until it eventually becomes.. the batting average.

lol exactly what I've been trying to say. It is what it is, and has the limitations it does, for a reason.

Sure you can fix it so that it more fairly represents the quality of the batsmen but then it's a pointless exercise

 

[Bolo]

For people thinking this regresses to batting average, look at the non-linear size of point categories in Nufans explanation. This system rewards consistency.

Awarding longer series increased points would involve something of a regression to career average. I don't think this is a good idea. A series is a series. Marquee series tend to be longer, and they can be more meaningful, but it isn't necessarily the case, especially in terms of competitiveness. Rather a high series average in a 2 match tour of Lanka than AUS beating up on England at home in a 5 match one. It's likely the difference between winning 1 series and two.

Also, have you considered making the tiers more fluid?

eg. letting someone avergaing 45 get 2.5 instead of 2, and someone averaging 42 getting 2.2.

I think this would be probably more faithful to the idea of rating consistent series performances, as opposed to just arbitrary gatekeeping.

The logical conclusion of this is simply to ditch the categories altogether and go for a form of exponential smoothing. It's counterintuitive to the vast majority of us though. Hard to set up and hard to digest. As it sits, at least we can make sense of it, and it shouldn't be too different on the whole.

Yes, go for it. Well send to me first please.



Below average performance for who? I totally get what you are saying that the highest tier could have been higher. Perhaps I was brought up watching more 4 or 5 Test per series, compared to the now more common 2 or 3 Tests and if you get to 60 for the series, you've certainly had a big impact.

A couple of players average over 60 career. Most players you are listing average pretty close to 60, so 60 is not a notably good series for them, merely a slightly above average one- the point range is one (60+) for an above average series and 6 for a below average one (below 50).

 

[NUFAN]

For people thinking this regresses to batting average, look at the non-linear size of point categories in Nufans explanation. This system rewards consistency.

Awarding longer series increased points would involve something of a regression to career average. I don't think this is a good idea. A series is a series. Marquee series tend to be longer, and they can be more meaningful, but it isn't necessarily the case, especially in terms of competitiveness. Rather a high series average in a 2 match tour of Lanka than AUS beating up on England at home in a 5 match one. It's likely the difference between winning 1 series and two.


The logical conclusion of this is simply to ditch the categories altogether and go for a form of exponential smoothing. It's counterintuitive to the vast majority of us though. Hard to set up and hard to digest. As it sits, at least we can make sense of it, and it shouldn't be too different on the whole.



A couple of players average over 60 career. Most players you are listing average pretty close to 60, so 60 is not a notably good series for them, merely a slightly above average one- the point range is one (60+) for an above average series and 6 for a below average one (below 50).

If you average 50 you are a great of the game. 60 is phenomenal. You are downplaying just how nice a series 60+ is. Most greats average 60 about 1 in 2.5 / 3 series and dont forget most cricketers arent as good as the guys ive chosen.

 

[stephen]

Yeah I've got no issue with the methodology. It's designed to reward consistency over the course of a career by looking at series level granularity. It has some limitations but it's not designed to act as a proxy for batting average, rather it looks at how consistently batsmen had great series'.