***Official*** New Zealand in the West Indies 2014

[flibbertyjibber]

5 wicket bags and 4 wicket bags
southee 4,7
boult 3,3
wagner 1,4

All of them need more 5 wicket hauls.

Be like the West Indies early 80's side, won't get them as they all get wickets hunting in a pack.

 

[Blocky]

Braces is way down the current pecking order for mine. Whilst I agree that Southee & Boult are a major reason for NZ's recent test improvement, I'm excited by the guys on the fridge such as Henry, McClenaghan, Bennett & Milne. I can't recall NZ have such promising pace options on the sidelines tbh.

Agree - also really interested to see the development of guys like Kuggeleijn, Ben Wheeler and Daryl Mitchell this season as they're all coming off their best seasons to date. I think Mitchell is one good season away from starting to knock on the door of selection and if Kuggeleijn keeps improving at the rate he has, he'll be another all rounder to throw into the mix.

 

[Hurricane]

Be like the West Indies early 80's side, won't get them as they all get wickets hunting in a pack.

Yeah even still though Southee is averaging 3.2 bags per 100 wickets while Marshall averaged 5.9

Yours and Steve's point is taken but would be nice to see any of them get one on occasion.

 

[HeathDavisSpeed]

Kugglejuggle is on the improve, but is still well away from where he needs to be with the ball to be in contention for Test honours. Whilst he's clearly a very useful batsman, I don't think he's going to have the goods to compete with Neesham and Anderson so really he needs to provide more with the ball and become someone who can be heavily relied on with the ball. Then he can target Wagner's spot.

 

[Blocky]

Kugglejuggle is on the improve, but is still well away from where he needs to be with the ball to be in contention for Test honours. Whilst he's clearly a very useful batsman, I don't think he's going to have the goods to compete with Neesham and Anderson so really he needs to provide more with the ball and become someone who can be heavily relied on with the ball. Then he can target Wagner's spot.

Agree - but two years ago I'd have said he wasn't anywhere near a first class players asshole and he's had a pretty good season and his batting statistics haven't really paled in comparison to either Anderson or Neesham at first class level.

 

[Mike5181]

Anyone else rate Tom Blundell?

 

[RossTaylorsBox]

Ah, the conversation moved on.

 

[Hurricane]

A Personal Improvement Stat
* Since August 2012, Tim Southee has taken 81 wickets in 17 Tests, at an average of 22.0. Prior to that, he had taken 42 in 17 matches, at an average of 44.7. Of the 44 bowlers who have taken 15 of more wickets since August 1, 2012, Southee's average is the third best, behind Mitchell Johnson (74 at 17.4) and Ryan Harris (56 at 21.6).

* In his last 21 Test innings, since May 2013, Brendon McCullum has reached 40 just three times. On those occasions, he has scored 113, 224 and 302. In the other 18 innings, he has averaged 13.2. (In his previous 39 innings, he had reached 40 on 14 occasions, but had a highest score of 84.)

Blogs: Andy Zaltzman: Stats gems you couldn't unearth even if you made them up | Cricket Blogs | ESPN Cricinfo

 

[Zinzan]

Looking forward to WW's post.

Predictions:

Concedes defeat.
Hates on the selectors
Lays out a challenge for next time.
Talks up the pace attack.
Hates on the selectors + Shillingford.
A wish for Narine.
Talks up one of Southee or Boult
Hates on the selectors

I'm going for 3 smileys as well.

You missed;

1) The ICC/Umpiring/DRS world-wide conspiracy to keep West Indies cricket weak.
2) Roach not being quite match/series fit
3) References to 2012 (and what Narine did as if nothing else with the sides have changed)
4) Predictions of several new 'kids' who will form part of a great Windies side next year
5) A "We'll see what happens 'next time we play NZ, shall we" comment
6) A grudging congratulations to NZ.

I see your 3 smiles and raise you an eye-roll.

 

[Blocky]

Ah, the conversation moved on.

You missed the point mate - you're trying to claim "Cricket can't possibly use statistics in the same way Fortune 500 companies invest millions upon millions doing" by showcasing one style of statistics that isn't even used widely by businesses anymore. There are many models that Cricket uses that are absolutely the same as what businesses use, WASP is a great example of a tool that businesses would also use to forecast the likely result of something based on current position and situation. You also miss the point that average is one indicator in Cricket out of literally hundreds that can be used to marry up against average.

In the case of WASP - it's even averaging the likely possibility of something "untoward" happening in a match, that is built into it's algorithm. It predicts how likely events like that are to happen based on current match situation and uses simulation (many thousands of times over) to come up with what is likely to happen. It's main flaw isn't that it couldn't take into account the quality of a batsman, it simply doesn't - you could program WASP to take into account the overall career performance of a player, combined with his recent performance, combined with his performance of players in the opposition, combined with performance on that ground, combined with performance during the time of day - at the moment, WASP isn't differentiating at all between a #4 batsman and a #4 batsman but that's more down to certain IP being held by NZ Cricket as they funded the PHD in the first place.

I know for a fact that they're not far away of translating the co-ordinates of ball pitching with the velocity and deviation to a metric that can be used to work out plans to certain types of batsman.

You're attempting to argue that a game so rich in concrete statistics that help us understand performance and allow us to know what "a good score is" cannot be used to forecast or simulate match results, which is completely contradictory to how most of the modern world now views Cricket. Go and watch Moneyball if you want to find out how statistics in a comparable game gave an under resourced team an advantage.

 

[Athlai]

I'm mostly upset that the Russian dude who put like $500 on 200+ runs @ $1.30 to be scored on day 5 got paid out.

This was not a great bet.

 

[Howsie]

Braces is way down the current pecking order for mine. Whilst I agree that Southee & Boult are a major reason for NZ's recent test improvement, I'm excited by the guys on the fridge such as Henry, McClenaghan, Bennett & Milne. I can't recall NZ have such promising pace options on the sidelines tbh.

I'd love to see the CD four of Bracewell, Wheeler, Milne and Small finally hit their straps with an injury free (or drama free in Bracewell's case) season. I still can't believe a region like that produced four bowlers of that talent at the same time.

 

[Zinzan]

I'd love to see the CD four of Bracewell, Wheeler, Milne and Small finally hit their straps with an injury free (or drama free in Bracewell's case) season. I still can't believe a region like that produced four bowlers of that talent at the same time.

Agree that would be an exciting 4-some at domestic level, but would you rate Braces as one the top 5 seamers in the country at the moment?

 

[Howsie]

Agree that would be an exciting 4-some at domestic level, but would you rate Braces as one the top 5 seamers in the country at the moment?

Nope. Not at the moment. From what I've seen, if I was to rate the bowlers in this country at their best I'd go:

1. Tim Southee
2. Trent Boult
3. Doug Bracewell
4. Ben Wheeler
5. Matt Henry
6. Neil Wagner
7. Mitchell McClenaghan
8. Adam Milne
9. Hamish Bennett
10. Bevan Small

I was a massive fan of Small a few years ago, injuries really seem to have taken a toll on his body and overall action. Hopefully he can get back to what he looked like when he was 19/20.

 

[auditor]

I'm mostly upset that the Russian dude who put like $500 on 200+ runs @ $1.30 to be scored on day 5 got paid out.

This was not a great bet.

I have every day market 200 + runs @ 1.1 - 1.3 @
This is a great bet, I often use it

 

[harsh.ag]

Wioderful series to watch this one, and I glad it was as close as it was. Though I wish the last day was slightly more competitive than it turned out to be. Tbf, I only watched this in between the FIFA coverage and Wimbledon, but it was as close to a dream sports night as it gets

 

[RossTaylorsBox]

You missed the point mate - you're trying to claim "Cricket can't possibly use statistics in the same way Fortune 500 companies invest millions upon millions doing" by showcasing one style of statistics that isn't even used widely by businesses anymore. There are many models that Cricket uses that are absolutely the same as what businesses use, WASP is a great example of a tool that businesses would also use to forecast the likely result of something based on current position and situation. You also miss the point that average is one indicator in Cricket out of literally hundreds that can be used to marry up against average.

In the case of WASP - it's even averaging the likely possibility of something "untoward" happening in a match, that is built into it's algorithm. It predicts how likely events like that are to happen based on current match situation and uses simulation (many thousands of times over) to come up with what is likely to happen. It's main flaw isn't that it couldn't take into account the quality of a batsman, it simply doesn't - you could program WASP to take into account the overall career performance of a player, combined with his recent performance, combined with his performance of players in the opposition, combined with performance on that ground, combined with performance during the time of day - at the moment, WASP isn't differentiating at all between a #4 batsman and a #4 batsman but that's more down to certain IP being held by NZ Cricket as they funded the PHD in the first place.

I know for a fact that they're not far away of translating the co-ordinates of ball pitching with the velocity and deviation to a metric that can be used to work out plans to certain types of batsman.

You're attempting to argue that a game so rich in concrete statistics that help us understand performance and allow us to know what "a good score is" cannot be used to forecast or simulate match results, which is completely contradictory to how most of the modern world now views Cricket. Go and watch Moneyball if you want to find out how statistics in a comparable game gave an under resourced team an advantage.

I don't think you understand my argument at all. I'll quote myself:

"You say that cricket statistics are the same as business statistics, so why don't we have Bayesian models for analysis and prediction purposes in cricket? It's because the simple metrics we use for cricket aren't the same as statistical models businesses use."

I'm not saying cricket statistics or business statistics are useless, I'm saying that they are different. Surely you can see that using an average to predict future performance is not the same as using hundreds of variables to construct a prediction model. I mean, technically they are both models, but one of them is more complex. By the way, I said Bayesian models, but this is completely irrelevant. Replace Bayesian with any statistical model you like. The point was to compare very simple mathematics with more powerful models. That's why I said the metrics we have for cricket (average, strike rate, etc.) aren't the same as models. As such, I also think they're not a tool with which you should base all your decisions on, although it may be effective, as Moneyball has shown.

And I still stand by my point on the efficacy of models. Outliers can cripple algorithms. No model will tell you that if Southee is playing in the next match he will probably take 0/100 and Sodhi will pick up 10 wickets. I would also argue that variables used in a business climate are different than those in cricket, where things like the state of mind of individual payers have an effect on the game. The only cricket model I've seen is WASP, which predicts the outcome of the game. I assumed we were predicting the performance of players and I'm guessing teams have their own statisticians who use their own models to do that for them, but I haven't seen anyone else do anything more complex than assessing players against varying teams, bowlers, etc. Even you compared bowlers simply by using a simple "x% is higher than y%". Again, this isn't a necessarily a bad thing, but it shouldn't be the only part of the decision-making process.

Anyway, I don't want to derail this thread with this argument, I'd rather bask in a series win.

 

[Blocky]

The point was to compare very simple mathematics with more powerful models. That's why I said the metrics we have for cricket (average, strike rate, etc.) aren't the same as models. As such, I also think they're not a tool with which you should base all your decisions on, although it may be effective, as Moneyball has shown.

The problem with this is that you seem to lack a basic level of understanding around what a model actually is. I work in this field professionally as a consultant and have done projects for NZ Cricket, NZ Police, Ministry of Health, Australian Federal Police, New Zealand Defense Force, Australian Defense Organisation, NZ Warriors and many of the banks and insurance companies in AsiaPac.

A model is simply a series of complex statistical calculations that take a number of variables - i.e metrics - into account and apply them to a dimension such as time, relative position or category. In many instances, these models utilise variables to assess correlating cause and effect factors of other events. There is no difference between the City Council using time variance modelling on the rate of decay of water pipes, combined with smart sensors that provide further variables such as water flow, temperate, weight of a pipe section etc and the WASP model - they're both using a form of the Poisson distribution method. The City Council then uses the insight gained from this model to create its strategy around asset management, the best case and worst case scenarios of the requirements of the city. No less than the WASP model is currently used to show a percentage chance of the team batting second of winning a match based on their current match position, enabling said team to create its strategy around which match position gives them the best chance of victory in any particular time interval.

The results of either one of these models give you valuable insights that do forecast the probability of a future outcome. The results of either one of these models also give you variables which can be used in further modelling to understand other factors such as the rate of population increase and how that will impact that piece of pipe - or in crickets case, utilising sensors in the wicket to understand soil and moisture content, taking into account individual performances of players versus other players, etcetera, etcetera.

So, coming back into your line of thought - outliers do not cripple algorithms, because they are often built into the algorithm and most modern day algorithms adapt and adjust as the event occurs. "The only cricket model I've seen is WASP" - then you haven't really got much of a leg to stand on, considering WASP is a very small piece of a highly complex model that was generated for NZ Cricket. The WASP you see on TV is simply one small variable of an aggregated strategy model that assists NZ Cricket in match situations. John Bracewell even assessed this as giving his team a massive competitive advantage over the opposition in limited overs cricket, but still required the players to be able to execute the strategy.....

And there in lies historic performance and other mechanisms - more and more data is captured on players in their training to help understand what their likely performance will be within a match.... as smart sensors become more prevalent, this information is helped to promote consistency in athletes through understanding what they require for peak performance - again, I've got first hand knowledge and experience implementing some of these models with sporting agencies.

Coming back to Cricket, cricket is awash with concrete statistics that make up absolutely great variables for any number of models in assessing any number of different cause and effect factors, that's before you start taking into account pitch maps and translating those to relative location that can be used in simulating bowling plans to certain batsmen, etc.

Quite frankly - in the case of this series, all of the statistics both career and also recent backed New Zealand as the likely winners and required the Windies to play out of their skins (which they kind of did, as previously discussed) and NZ to play poorly (which they kind of did, as previously discussed) for this series to be close. So going back to how you assess teams and players, do you assess Don Bradman and Brian Lara on their one peak performance? 401 vs 333? Or do you assess them on their history of performances to understand what their likely value to a side will be in winning a test match.

Which leads me back to your original statement " Why don't we have predictive and analytical models in Cricket?" - we do. You just don't have access to them.

PS - every bookie in the world is using statistical analysis to set their odds. You'd have more of a leg to stand on if you said that "statistics and models by themselves are just numbers and without subjective opinions on how we interpret those numbers, the do not tell a full story" - I'd agree with that, but then I've been using interpretations to prove my point - i.e Wagner vs Boult.

 

[hendrix]

I don't enjoy it, because I also realise how flawed it is and how much better it could be - if this was developed in a nation outside of NZ, it would have had millions thrown at it in investment, as it stands, this was just one clever guy doing a PHD thesis on Duckworth Lewis being a highly flawed method to set targets in a game of cricket. WASP started its life as a way of simulating what likely score would be struck if a team knew how many overs they were going to bat, based on their performance to date within an innings.

It still has this function built within it, the amount of models and analysis that it could provide, versus what SkyTV actually shows you is laughable - it's really only the fact that most people just don't get how to use statistics or even data visualisation to explain points.

I absolutely understand how it works. While statistics isn't my trade, I'm a research biologist so I pretty much have to use stats for my work (no matter how much I avoid it).

I enjoy WASP because it seems I look at it differently from most people. I just take it as "historically, if a team were in this position, they'd likely set this score" or "if a team were in this position, they'd have x chance of chasing the total".

Of course that chance is going to fluctuate ball by ball. That's the point.

 

[Blocky]

I absolutely understand how it works. While statistics isn't my trade, I'm a research biologist so I pretty much have to use stats for my work (no matter how much I avoid it).

Cool - outside of Cricket discussion we'd probably get along.. :P

I enjoy WASP because it seems I look at it differently from most people. I just take it as "historically, if a team were in this position, they'd likely set this score" or "if a team were in this position, they'd have x chance of chasing the total".

Exactly how the guy who wrote the model intends for it to be taken.