First thread - Who will be the top run scorer and wicket taker?

[BeeGee]

Cook and Pattinson.

Clarke is due a poor series and I think the English bowlers will share the wickets around pretty evenly, but Pattinson will take big hauls.

 

[Spark]

am i the only one who finds the whole "due a poor/good series" logic inherently weird?

 

[Jono]

It has as much logic as believing a coin is due to land on tails due to the fact its landed as heads the last 4 times.

 

[Daemon]

It has as much logic as believing a coin is due to land on tails due to the fact its landed as heads the last 4 times.

It's fair enough to believe someone's been playing way above his actual standards and it's only a matter of time he comes back down to earth.

Though I guess it also works the other way though, someone in good form is more likely to score in an upcoming series than someone who isn't.


carry on

 

[Valer]

It has as much logic as believing a coin is due to land on tails due to the fact its landed as heads the last 4 times.

Batsmans' cricket scores aren't IID so this is incorrect. Can't be assed to import all that data to show this however. IIRC when I've seen this done before they don't have the markov property either.

 

[Prince EWS]

It's fair enough to believe someone's been playing way above his actual standards and it's only a matter of time he comes back down to earth.

Yeah but that's not really saying "due a poor series"; that's saying "isn't as good as their performances so far have indicated and will start to fall back gradually as they play to their ability." If you're played above your mean then you're not suddenly more likely to play below your mean; that's nonsense. What you are likely to do is play to your mean and have your average converge on it.

It's the Monte Carlo fallacy, gambler's fallacy, maturity of chances myth - whatever you want to call it - but applied to cricket.

 

[Spark]

Batsmans' cricket scores aren't IID so this is incorrect. Can't be assed to import all that data to show this however. IIRC when I've seen this done before they don't have the markov property either.

Yeah that stands to reason, but it doesn't follow that batsman doing well for ages ---> batsman has a higher-than-even chance of underperforming next season. That doesn't make sense.

 

[Valer]

Yeah but that's not really saying "due a poor series"; that's saying "isn't as good as their performances so far have indicated and will start to fall back gradually as they play to their ability." If you're played above your mean then you're not suddenly more likely to play below your mean; that's nonsense. What you are likely to do is play to your mean and have your average converge on it.

It's the Monte Carlo fallacy, gambler's fallacy, maturity of chances myth - whatever you want to call it - but applied to cricket.

Form is predictive ergo the gambler's fallacy does NOT apply...
pappubahry: The concept of form in cricket has predictive value, though its sign varies with the batsman.

Spark its sign varies as seen here. Note this is NOT my blog.

 

[Daemon]

jesus christ you ****ing nerds

 

[Prince EWS]

Form is predictive ergo the gambler's fallacy does NOT apply...
pappubahry: The concept of form in cricket has predictive value, though its sign varies with the batsman.

Spark its sign varies as seen here. Note this is NOT my blog.

That's the complete opposite of the "due for a failure/success" after the opposite has happened theory though.

That blog is suggesting that form is real and a batsman is more likely to score runs when in form and less likely to score runs when out of form - the complete opposite of saying someone is due for a failure because they've used up all their runs (which really is exactly like the Monte Carlo fallacy).

 

[Valer]

That's the complete opposite of the "due for a failure/success" after the opposite has happened theory though.

That blog is suggesting that form is real and a batsman is more likely to score runs when in form and less likely to score runs when out of form - the complete opposite of saying someone is due for a failure because they've used up all their runs (which really is exactly like the Monte Carlo fallacy).

Once again tho even if you've go a postive correlation to previous score its still not the Monte Carlo fallacy -- it requires IID trials. This doesn't suggest that the person making the failure claim is correct in positive correlation situations.

Did you actually read the post? Hes suggesting that you can actually have both. Some quotes.

Justin Langer got most of his big scores after mediocre runs.

When Bradman had a rut for five innings, England must have been very, very worried. Conversely, if he'd made 750 runs in his last five dismissals, they would have been confident that he was due for a low score.

Fwiw I think some form of modified garch would work well here but I haven't tested it. I could even trying adding in team mates scores (obviously multivariable) to see how it changes things. Probably too lazy tho.

 

[Jono]

How long is "too long" whereby you're due for a failure? 10 innings of straight 50+ scores? Or 5?

 

[Adders]

jesus christ you ****ing nerds

Was just about to post something similar. Got NFI what they are talking about but I'm sure they're right.

I'm due a worthwhile post now me thinks

 

[Jono]

Adders

 

[Valer]

How long is "too long" whereby you're due for a failure? 10 innings of straight 50+ scores? Or 5?

Its not clear what you're asking. Unless you have a negative correlation (or zero) you're going to more likely to score runs after success in one particular innings.

It should be noted David Barry (the author) only looked at a group of 5 outs prior.

However something like Pr(# [innings with runs <50 after n innings] >=1) will converge to 1 (in most cases).

 

[honestbharani]

Yeah but that's not really saying "due a poor series"; that's saying "isn't as good as their performances so far have indicated and will start to fall back gradually as they play to their ability." If you're played above your mean then you're not suddenly more likely to play below your mean; that's nonsense. What you are likely to do is play to your mean and have your average converge on it.

It's the Monte Carlo fallacy, gambler's fallacy, maturity of chances myth - whatever you want to call it - but applied to cricket.

Man.. it is real trouble when statsmongers start equating a person playing cricket to a coin toss and make comparisons..

 

[honestbharani]

Man.. it is real trouble when statsmongers start equating a person playing cricket to a coin toss and make comparisons..

Does any of the great stats guys who are coming up with these theories even understand the fact that how a batsman bats depends on a how the bowler bowls and vice versa.. It is almost as if every batsman plays against bowling machines and are supposed to get out or make big runs every Xth, Yth and Zth inning... Jeez.. Talk of overcomplicating a simple business. This is where my biggest gripe about having Performance Analysts telling professional cricketers what to do comes in.. Sure,as a performance analyst, I can scour through data and tapes and figure out that the release shot for a particular batsman when kept quiet is a slog over cow corner but how exactly does that help if said batsman is actually good enough to hit the ball over cow corner of either foot and off any line from outside off to outside leg? Theorizing on Human Beings as if they were predictable elements is just the biggest pile of **** going on today...

 

[Jono]

But that's not the point PEWS is saying. All he is saying is that people saying someone is due a poor series is a stupid statement with no genuine argument or even remote evidence behind it. Its a figment of everyone's imagination.

In fact there was nothing stats monger about that post you quoted of PEWS. He may have used stats terminology (i.e mean) but really, what he said was actually just common sense.

 

[Valer]

But that's not the point PEWS is saying. All he is saying is that people saying someone is due a poor series is a stupid statement with no genuine argument or even remote evidence behind it. Its a figment of everyone's imagination.

In fact there was nothing stats monger about that post you quoted of PEWS. He may have used stats terminology (i.e mean) but really, what he said was actually just common sense.

But (some) batsman do tend to be more likely to have a poor innings after a string of good ones....

This makes no attempt at explanation tho. It could be just that batsman that do this minow basher and end up facing good bowlers after the string of good results, or in the other direction had to face a very good attack and got back to beating up easier attacks.




As an aside I'd suggest that flawed models are NOT a good excuse for not trying to fit models.

 

[uvelocity]

But (some) batsman do tend to be more likely to have a poor innings after a string of good ones....

This makes no attempt at explanation tho. It could be just that batsman that do this minow basher and end up facing good bowlers after the string of good results, or in the other direction had to face a very good attack and got back to beating up easier attacks.




As an aside I'd suggest that flawed models are NOT a good excuse for not trying to fit models.

don't mind if i do