Leaving out the minnows...

[fredfertang]

How about the Ashes series between 1977 and 1985 - should they be expunged from the record? The 78/79 sides, Australia in particular, were lamentably weak due to WSC

 

[silentstriker]

Should India have been considered minnows away from home in the 1990s? 15 years, zero Test wins. Not even in Zimbabwe. Sounds like minnows to me, no?

 

[Uppercut]

Should India have been considered minnows away from home in the 1990s? 15 years, zero Test wins. Not even in Zimbabwe. Sounds like minnows to me, no?

They drew plenty though, they just had a weak attack.

 

[HeathDavisSpeed]

Should India have been considered minnows away from home in the 1990s? 15 years, zero Test wins. Not even in Zimbabwe. Sounds like minnows to me, no?

This is the issue, isn't it. What is a minnow?

Bangladesh at home, for example, nearly won test matches against Pakistan (lost by one wicket) and New Zealand. And didn't they dominate Australia for a couple of days in a home test, only to fall away?

Is there scope for including Bangladesh's games at home in a 'fair' analysis?

 

[HeathDavisSpeed]

They drew plenty though, they just had a weak attack.

So, runs scored by opponents could be left out because the Indians had the bowling attack of minnows?

 

[silentstriker]

They drew plenty though, they just had a weak attack.

Hence being minnows away from home. Should opposing batsmen not have their records counted against traveling Indian sides, while their bowlers records should because the batting might have been up to par?

 

[Goughy]

It is a simple cleave. No point getting overly concerned or complicated. If you want to compare like-with-like for modern players then Zimbabwe and Bangladesh should be excluded from the records.

Purely for trying to get to sets of figures that are remotely comparable. It is an artificial tool but one that is often needed due to certain players playing a disproportional amount of cricket against weaker or stronger teams.

 

[Uppercut]

Hence being minnows away from home. Should opposing batsmen not have their records counted against traveling Indian sides, while their bowlers records should because the batting might have been up to par?

Personally, cbf. Wouldn't bother taking it that far. It might be interesting to see that certain batsmen scored an excessive proportion of their runs against India at home, but I wouldn't bother taking India out completely.

It's not the same thing as removing a side that have pretty much lost 52 straight matches. That's a simple cleave (I like that phrase ).

 

[Goughy]

This is the issue, isn't it. What is a minnow?

Bangladesh at home, for example, nearly won test matches against Pakistan (lost by one wicket) and New Zealand. And didn't they dominate Australia for a couple of days in a home test, only to fall away?

Is there scope for including Bangladesh's games at home in a 'fair' analysis?

No. There is no evidence. In the last 2 years Bang have only played 6 Tests at home and lost 5 and drawn 1.

Over that 2 years 10 opposition wickets average out to equal 19 Bangladesh wickets.

If that is the basis for a round of applause and a change of status then you have lower standards than me.

 

[Maximus0723]

Bangladesh was minnows all the time. Zim and Kenya after '00.

 

[Top_Cat]

There's no statistically defensible reason to exclude a player's record against 'minnow' teams. You might weight them differently if you wish but they're not outliers.

 

[Flem274*]

Last night when I was stuffing around on statsguru (rarely visit it) I found Ross Taylor's average goes up to 45 excluding Bangladesh.

Wondering if there are other players that improve with minnows excluded (along with Harbijan (sic))

 

[Top_Cat]

Last night when I was stuffing around on statsguru (rarely visit it) I found Ross Taylor's average goes up to 45 excluding Bangladesh.

Wondering if there are other players that improve with minnows excluded (along with Harbijan (sic))

Exactly.

Without checking, Mark Waugh's average would surely go up if you exclude Sri Lanka (who were borderline minnows until the early 90's)

 

[Uppercut]

There's no statistically defensible reason to exclude a player's record against 'minnow' teams. You might weight them differently if you wish but they're not outliers.

By that logic, is there a statistically defensible reason to exclude a player's FC record in Australian cricket?

 

[The Sean]

Exactly.

Without checking, Mark Waugh's average would surely go up if you exclude Sri Lanka (who were borderline minnows until the early 90's)

Actually reckon Warney's would improve slightly off the top of my head - though as he only got 3 Tests out of 145 against "minnows" it's not much of a sample.

 

[Top_Cat]

By that logic, is there a statistically defensible reason to exclude a player's FC record in Australian cricket?

Look at it this way; the usual (but not only) guide for an outlier is when an observation is at least two standard deviations away from the mean, yeah? Murali's record against, say, Bangladesh is well within that. Zimbabwe, within 1 std.

Sure you can use a player's records against other countries to give context for their performances and certainly you can qualitatively rate them as low as you like but there is no numerical basis for exclusion.

 

[Uppercut]

Look at it this way; the usual (but not only) guide for an outlier is when an observation is at least two standard deviations away from the mean, yeah? Murali's record against, say, Bangladesh is well within that. Zimbabwe, within 1 std.

Sure you can use a player's records against other countries to give context for their performances and certainly you can qualitatively rate them as low as you like but there is no numerical basis for exclusion.

Yeah- but are Shane Warne's performances against Tasmania two standard deviations away from the mean? I don't suppose they are. So, numerically, should we include those too?

 

[bagapath]

Exactly.

Without checking, Mark Waugh's average would surely go up if you exclude Sri Lanka (who were borderline minnows until the early 90's)

But SL had already won test series against India and Pakistan by then. Bangladesh is still pits.

I cant believe one can rate a 50 against Australia and a 50 against Bangladesh equally. As far as I am concerned, Sachin's highest score in test cricket is 241 n.o. against Australia in Sydney. His highest score in first class cricket is 248 vs Bangladesh.

 

[Top_Cat]

Yeah- but are Shane Warne's performances against Tasmania two standard deviations away from the mean? I don't suppose they are. So, numerically, should we include those too?

As it happens, he averages around 25 against Tas. If you're talking about his FC performances, why would you not?

A full analysis of a player should include their FC records, really. You just have to weight them appropriately. Exclusions, unless in some seriously extreme circumstances, should be fairly minimal. It's not a big thing.

 

[Top_Cat]

But SL had already won test series against India and Pakistan by then. Bangladesh is still pits.

I cant believe one can rate a 50 against Australia and a 50 against Bangladesh equally. As far as I am concerned, Sachin's highest score in test cricket is 241 n.o. against Australia in Sydney. His highest score in first class cricket is 248 vs Bangladesh.

Whoa, no-one has said that.