Monday 28 March 2011

Head vs heart

People have asked me who I'm backing in the New Zealand semi-final match. Unfortunately my head is going against my heart on this one.

I've got a couple of statistical explainations for your geeky perusal.

I made a bit of a fool of myself in a discussion recently saying that it would be stupid to bat first in Colombo in a day/night game, because the dew makes it very difficult to bowl in the second innings. While it might make it more difficult to bowl, Premadassa is one of only three grounds in the world where we can be statistically confident that it is better to bat first. (The other two are Newlands and Sydeny, interestingly all are at sea level). So there is actually a significant advantage to any team batting first in Colombo.

The historical statistics are not good for New Zealand. Over the past 5 years only India have tormented New Zealand more than Sri Lanka, with New Zealand wining 3 and losing 8. Likewise of the test playing nations, only Bangladesh and Zimbabwe have a worse record against Sri Lanka than New Zealand.

Sri Lanka also have a better record in semi-finals as well. New Zealand have won just 2 of 12 semi finals, while Sri Lanka have won 3 of 8. (this includes all tournaments that have semi-finals, not just world cups). Only South Africa and Kenya have a worse record than New Zealand.

Also, while teams from Asia have only won 44% of matches against non-asian opponents outside Asia in the last 5 years, they have won about 60% of the equivalent matches in Asia. (hence it's not really surprising that all 3 top teams have made the semi-finals.

Oh and also 5 of Sri Lanka's 6 bowlers have averaged under 21 against New Zealand over the last 5 years, with only Mendis (43.00 at 4.03 rpo) dragging the chain.

So those are some of the reasons that my head is disagreeing with my heart as to what is going to happen in the first semi-final.

Saturday 26 March 2011

Odds, probability and winning records

As a task for my statistics students I often use tennis results to create probabilities for players to beat each other, and then get my students to construct a probability tree to find out the probability of each player winning a tournament. I normally tell them the semi finalists, or one game before the semi-finalits are found, to avoid the tree taking too long to construct. (It's better to spend that time thinking rather than writing in meaningless numbers).

The system is flawed in a number of ways, but mostly because the true probability of a player winning is much more complicated than just the sum of their previous results. Things like fitness, form, and surface make a big difference. And yet in some ways the data is the most accurate and most quantifiable that we have. Also often the person that we find the greatest probability for ends up winning the tournament.

I was having a conversation with someone about how I wanted England to beat Sri Lanka, because New Zealand had a much better record against England than Sri Lanka. It got me thinking that perhaps I could do what I have had my students do.

So I got the data for the last 10 years (the results from before that not really very relevant) and used them to generate some probabilities. The results were actually quite interesting. Here are the 5 most one sided match ups in terms of recent history:

Fot those of you for whom probability is a long way in the past, I've included odds on the favourite, which may be a more familiar way of measuring it. (I've assumed an even chance is at odds of $1.85, this differs between bookmakers)

MatchupProb.Odds
NZ vs Eng0.714$1.30
Ind vs Eng0.667$1.39
SL vs NZ0.615$1.50
Ind vs SL0.589$1.57
Pak vs NZ0.581$1.59


The closest matchup is India vs Pakistan, with both teams having an even record over the last 10 years.

I then put that data into a probability tree, and came out with some interesting data.

WinnerProb.Odds
Ind0.296$3.13
Pak0.278$3.32
NZ0.239$3.87
SL0.123$7.49
Eng0.063$14.66


I also found out the probability of making the final:

FinalistProb.Odds
NZ0.564$1.64
Ind0.5$1.85
Pak0.5$1.85
SL0.28$3.31
Eng0.156$5.94


The last step was to check these against the odds that actual bookies are offering. I used Bet365 and TAB New Zealand. Here were what they were offering:

TeamBet365 WinBet365 FinalTAB NZ WinTAB NZ Final
Ind2.51.52.41.5
Pak52.552.5
NZ7.52.446.52.6
SL4.552.2542.15
Eng94.594.5


There are obvious differences between these numbers and my ones. Part of that is due to the bookies having much more sophisticated methods to make their predictions than just historical wins/losses. Part of it also is that they are having to cover existing bets. A brief look at the numbers suggests that Bet365 are running lower margins on the winner than the NZ TAB, but that is reversed for the finalists.

I'm not sure that I want to make any predictions based on this data, except to suggest that if England beat Sri Lanka, back New Zealand to make the final.

Friday 25 March 2011

Contributions made by openers

In a couple of previous posts, Partnerships that made big contributions" and A different kind of hundred partnership I looked at a different way of measuring the impact of a partnership.

I used the method from the contributions post to look at opening partnerships in this world cup. There have been a couple of outstanding partnerships this time round, not least of which was the innings by Dilshan and Tharanga of 282 vs Zimbabwe at Pallekele.

Here is the list of all partnerships that contributed 100 or more:

NamesForAgainstRunsContribution
TM Dilshan, WU TharangaSLvs Zimbabwe282307
BJ Haddin, SR WatsonAusvs Canada183246
WTS Porterfield, PR StirlingIrevs Netherlands177241
MJ Guptill, BB McCullumNZvs Zimbabwe166217
V Sehwag, SR TendulkarIndiavs South Africa142201
BJ Haddin, SR WatsonAusvs New Zealand133189
Kamran Akmal, Mohammad HafeezPakvs West Indies113180
RW Chakabva, BRM TaylorZimvs Sri Lanka116165
KP Pietersen, AJ StraussEngvs Netherlands105149
CH Gayle, DS SmithWIvs Netherlands100141
HM Amla, GC SmithSAvs Bangladesh98139
MJ Guptill, BB McCullumNZvs Kenya72139
KP Pietersen, AJ StraussEngvs Ireland91125
S Chanderpaul, DS SmithWIvs Ireland89124
TM Dilshan, WU TharangaSLvs Pakistan76105


A couple of names appear there more than once. Dilshan & Tharanga, Haddin & Watson, Guptil & McCullum and, despite often looking as comfortable as a pin in the bottom, Devon Smith.

That leads to the next obvious question: Which country has got the most value from their openers. Here's the list sorted by country and average contribution. (for this average I am dividing by innings, not dismissals, as there is already a bonus for getting a not out)

TeamContribution
Aus104.33
SL99.00
NZ78.17
Eng75.50
India68.29
SA62.67
WI57.14
Ire45.00
Zim38.00
Ban31.67
Neth27.14
Kenya11.83
Can10.00


There's some obvious distortion in these numbers caused by the ability of some teams at minnow bashing. The usual suspects for beating up on minnows (NZ, SA, India and SL) all feature high on the chart. I will at some point produce a table that removes games against lesser teams, but the pub is calling.

Happy World Cup watching everyone.

Thursday 24 March 2011

India vs Australia

Just a quick thought.

Throughout this series Australia have taken their second wicket inside the first 12 overs every single match. They have been exceptional at taking the top off their opposition. Only twice (vs Canada and Pakistan) did the third wicket fall after the 15th over.

India have had great success with their top 3. They have also had great performances from Yuvraj Singh. But Dhoni, Raina and Pathan have all had tournaments to forget so far.

I would suggest that the crunch point in this game will be the first 10 overs of India's innings. If Australia manage to break through early, the pressure may be too much for the Indian middle order to handle.

If they don't break through early, then their plans will all turn to custard very quickly. (for those unfamiliar with this Kiwi expression, despite custard being good, things turning to custard is bad. Especially if it is things like bridges and buildings.)

The other key point may well be the toss. Both teams are bat first teams, and are not nearly as comfortable chasing as defending.

Sunday 13 March 2011

Double after 30?

There is an old adage that to get a good gauge of the score that a team is going to get, double their score at 30.

Last night India were at 197/1 off 30 overs last night, and yet they only managed to get 296.

As I mentioned in a couple of previous posts, India have not managed to score double their 30 over score once in the last 2 years. Which makes me ask, is double after 30 realistic.

I've gone through the first innings of every game in the world cup and looked at the halfway point in terms of runs.

Of the 27 matches in which the first inning was completed 9 times the total score was double or more than double the 30 over score.

The team that left their run the latest was New Zealand, in the Taylor blitz against Pakistan, where they scored the same in the last 12 overs as they did in the first 38. I remember that clearly, as I went to bed after 34 overs and woke up expecting to hear that New Zealand had lost.

The worst was Bangladesh against West Indies, where they were looking reasonable at 30/2 off 6, but then collapsed to 58 all out.

Overall the mark is actually normally close to 28: 13 times it was between 26 and 31. All of the top 8 teams except England and India have had at least one game that they have doubled after 30, but of the other 6 only the Netherlands have managed it. (although Canada haven't batted first in a match yet).

If we were to look for a guideline, perhaps we could say double after 29 for a test playing nation against another test playing nation, but go under that for a non-test nation, and over that for a test nation against a non-test nation.

Saturday 12 March 2011

Keiron Pollard and the World Cup activity ratings

Last night we saw an amazing innings for Keiron Pollard. When he arrived at the wicket, West Indies were in trouble. It was 130/3 off 32 overs. Devon Smith was looking about as comfortable as a penguin in the Sahara, and Ireland had the sniff of blood.

Pollard hit 5 sixes and 8 fours, and scored 94 off 55 balls. But interestingly he scored 62 runs in boundaries, off 13 balls, meaning that off the rest of his innings, he scored 32 off 42. Even if you remove the boundaries he scored at a strike rate better than 75.

His running between wickets had an effect on Devon Smith also. When Pollard arrived at the wicket Smith had a strike rate of 72.2 During the time that they were batting together Smith scored 29 runs off 25 balls. Pollard took the pressure off Smith, and he responded by batting a lot better. (Obviously he had got his eye in also, but the change in his attitude was quite evident watching the game.)

We all know how powerful Pollard can be. We think of the big sixes that he has hit. But the subtlety that he showed last night showed us that he is capable of being much more than a big hitting circus performer.

In his whirlwind 60 against the Netherlands, there were only 18 balls that he didn't hit to the boundary, and he managed to get 16 off those. The next step for him will be to repeat this against a test playing nation.

His activity rate (runs per ball not hit for a boundary) is far above anyone in the World Cup, but there are a couple of other interesting names in the list:

NameRunsBallsRuns from boundariesActivity rate
KA Pollard 154831060.787
DPMD Jayawardene 134123580.697
V Sehwag 2541951460.667
AB de Villiers 2662471200.664
Younis Khan 128159280.658
HM Amla 169196480.658
IR Bell 193227600.624
KC Sangakkara 252281920.618
AJ Strauss 2982991360.609
IJL Trott 289369760.609
Tamim Iqbal 152158660.606
Misbah-ul-Haq 192259440.594
V Kohli 154161680.593
G Gambhir 128143520.585
RN ten Doeschate 148165620.57
T Taibu 117140440.566
BRM Taylor 140154620.561
KJ O'Brien 1641251100.535
SR Watson 141153680.533
RS Bopara 106140360.53
Shakib Al Hasan 111144400.53
TM Dilshan 2832791580.521
Umar Akmal 167193760.52
CR Ervine 113133520.504


The three players who have been near the top of a lot of the activity rate charts are Sehwag, Dilshan and Tamim Iqbal. All are in this list, although Dilshan is a lot lower down than usual.

The bottom of the list features a number of players from associate nations, but there are a couple of very good batsmen who have made it into the list also.

Sorted from lowest to highest

NameRunsBallsRuns from boundariesActivity rate
AS Hansra 142264720.282
W Barresi 99154520.333
A Bagai 102180440.343
CO Obuya 144212720.367
WTS Porterfield 106169460.38
BJ Haddin 84116440.381
MJ Guptill 1922511040.388
DM Bravo 112149580.397
TLW Cooper 140213600.404
Imrul Kayes 111149560.407
CH Gayle 119149640.41
Kamran Akmal 118169520.423
P Utseya 86150240.431
EC Joyce 136207520.433
KP Pietersen 131136800.436
DS Smith 2022531020.439
WU Tharanga 2582921440.445
SR Tendulkar 2132221260.451
GC Smith 87151200.459
PW Borren 86101440.462


Haddin has managed to feature in the bottom list for other formats also, showing that he is very reliant on hitting boundaries for his usually good strike rate. Also at the wrong end of the table is Martin Guptill. Watching Guptill he doesn't seem to have a reliable single scoring shot against spin bowlers. Accordingly the more that spin bowlers bowl against him, the lower his activity rate falls.

It wil be interesting to see if Pollard manages to keep up his extrodinary rate as the competition heats up, and also if the likes of Tendulkar, Smith and Gayle can get their rating up towards 0.5.

Thursday 10 March 2011

Time to bet on Goliath?

The story of the young boy David felling the giant Goliath has resonated across many cultures for hundreds of years. The thought of the underdog triumphing over the gloating giant somehow warms our hearts. It has inspired artists to paint pictures, authors to write books, and musicians to write songs. One of my favourite is by Steve Apirana here, which I would recommend you play in the background, to provide some theme music to this preview.

So can Zimbabwe's David defeat the Goliath of Sri Lanka? Probably not.

However they do have some things going for them.

1. The match is at Pallekele. This ground looks like the real deal. A genuine Cricket ground, up in the mountains, with a nice bank, good facilities and an excellent pitch. It's a pitch that favours batsmen that take their time to get their eye in, and one that allows poor bowling to be punished.

Zimbabwe's batsmen certainly know how to take their time. 5 of their players have scored more than 100 runs at a strike rate of less than 60 in the last 2 years. But this big thing here is that it's a ground that seems to favour spin bowlers that get more bounce than spin. Bowlers like Price, Utseya and Lamb. Despite Sri Lanka being very good against spin, these are different spinners than they are used to.

2. Zimbabwe beat Sri Lanka recently. Within the last 12 months, the games have been 2-1 to Sri Lanka. And the one that Sri Lanka lost was convincing. Zimbabwe know that if things go well, then they are capable.

3. Chigumbura might actually fire. In the last 10 matches against Sri Lanka, Elton Chigumbura has scored 35 runs. 35 is a reasonable average, but a very bad total. However he has been beaten by spin in 7 of the 8 times that he's got out to a bowler. With spin less important in these conditions, It could be his turn. Finally.

However things are not all rosy. Here's some reasons that Sri Lanka can take heart.

4. The exception to the "spin bowlers don't take wickets at Pallekele" rule in Sri Lankan domestic cricket in Ajantha Mendis. He has not seemed to have the same worries. He also has an amazing record against Zimbabwe, averaging in single figures in 10 matches.

5. Nuwan Kulasekara Sri Lanka are known for their unorthodox bowlers, Mendis, Murali, Malinga, but Kulasekara hs the technique and patience that is perfect for undoing lesser players. And he bowls the sort of length that is likely to be rewarded in Pallekele. Watch out for him.

6. Upul Tharanga. Tharanga has seen Zimbabwe bowling as his meal ticket. He's scored 412 runs in 10 matches, almost 200 more than any other player in either team in those matches. And he looks hungry again tonight.

So my summary. There are reasons that Zimbabwe could win this game, but they are not convincing. If they are to win, they will need a good dose of luck.

If you are looking for somewhere to put your 20c, here are my suggestions. One almost sure bet would be for Sri Lanka to have the highest score after 15 overs, at $1.16. A slightly more risky, but at better odds is Tatenda Taibu to get 50 or more at $4.50. Taibu is good at playing when there is a bit of extra bounce, and so this pitch is likely to suit him.

Wednesday 9 March 2011

3 reasons why New Zealand fans should and should not get excited

Last nights game in Pallekele had a very convincing result. However it's not so clear that the performance was quite as convincing. There are my three reasons that we shouldn't get excited, and the three reasons that we should.

Reasons not to get too excited.

1. Ross Taylor got dropped twice early on.

If Pakistan didn't have Kamran Akmal behind the stumps the outcome of this game could have been vastly different. What are the chances of let-offs like this occurring in the semi-final? Much less than in a meaningless game where both teams have all but qualified for the next round.

2. We were playing Pakistan.

As I pointed out in my previous post Matching Mercury Pakistan and New Zealand have a tradition of beating up on each other. Both teams are as stable as the Italian Parliament and collapse to crushing defeats quite regularly. As a result a 100 run win really means not much more than a 100 run loss.

3. Vettori got injured.

The loss of Vettori is a more serious blow to New Zealand's chances than any boost in confidence that the team got by thrashing Pakistan.

Reasons to be excited:

1. Ross Taylor got dropped twice early on.

Taylor is the sort of player that can really capitalise on getting his eye in. In a way he's similar to Matthew Hayden. In the lead up to the last world cup there were calls to drop Hayden from the Australian side. The final warm up game agains New Zealand was possibly his last chance. He looked scratchy, was dropped early, and took 82 balls to get to 50. (similar to the number that Taylor took last night). However he started finding the middle of the bat, and went on to dominate all that came before him over the next 12 matches (the 2007 World Cup) scoring 840 runs at 93.33 and a strike rate of over 100. Could this be the start of a run like that for Taylor?

2. It was Pakistan.

This was not a thrashing of a minor team, or a team on the rocks. Pakistan are a powerful test side, who were riding high after beating Sri Lanka. New Zealand have lost their last 6 games to Pakistan, including 2 semifinals. And they have beaten a sub-continent team, in the sub-continent.

3. Vettori got injured

New Zealand managed to win this game convincingly without their no 1 star. The managed to beat a good team, on a sub-continent wicket, without relying on 10 overs from Vettori. Without Vettori's captaincy. Without him even facing a ball. If they are this good without Vettori, how good could they be with him.

Tuesday 8 March 2011

Matching Mercury

The dictionary defines mercurial as: changeable; volatile; fickle; flighty; erratic.

While we are still 16 days away from Mercury being at its farthest point from the sun, we are seeing two of the most mercurial teams playing in Sri Lanka tonight. New Zealand and Pakistan are both teams that can look like a Rolls-Royce or a rubbish truck.

Between the two teams in the last 3 years they have been involved in 20 matches that have been either won or lost by 100 or more runs, and 13 that have been concluded with more than 15 overs remaining.

In the 10 most recent matches between the two teams, 2 have been close. New Zealand won by 7 runs at Abu Dhabi and lost by 2 wickets with 6 balls remaining at Napier. But 2 have been two of the most one sided matches in recent memory between test teams: NZ won by 9 wickets just after the first drinks break at Wellington, but lost another match at Abu Dhabi by 138 runs.

As a result it is almost impossible to say who will win this match. The 10 matches are 5-4 to NZ with one N/R (that Pakistan were looking quite good in). It is a strange situation where a match is too close to call, and yet will probably be very one-sided.

To add to the mystery the game is the first ever ODI at Pallekele. Looking through the previous games at lower levels, the one thing that seems constant is rain. Almost every match was decided by messers Duckworth and Lewis. However, even this can't be relied on, as the forecast is for a pleasant sunny day, with negligible wind.

One thing that has been common on this ground is lots of wickets and runs for quick bowlers, and not may wickets or runs for spinners. In one match a quick took 5/57 off 7, while a spinner got 0/17 off his 7 overs. This suggests that it is a pitch that gets good bounce.

Given all this here is my (somewhat hesitant) things to watch for:

1. Tim Southee. Southee has really shone under Allen Donald and John Wright, taking 14 wickets in 8 matches at 21.71. There have only been two bowlers to have taken more wickets in that time, both of them are also (likely to be) playing, Shahid Afridi and Hamish Bennett. The difference here is that those other two players have been as mercurial as their teams, taking 4 and 5 wicket bags, and then nothing, whereas Southee has been difficult for his opponents in all but one match (on the road at Napier). Southee loves a little bounce, and a little greenness. Look for a big performance from him, especially if New Zealand bowl first.

2. Brendan McCullum. John Wright has been trying to get McCullum to slow down a bit. McCullum has averaged near 60 recently in tests, and near 20 in ODI's. If he manages to score some more runs, even at a lower strike-rate he may become much more of an asset for New Zealand.

3. Misbah-ul-Haq. This year Misbah has scored 388 runs at 77.60. And he's scored a lot of those runs against New Zealand. If they are going to win this game, they will probably need to find a way to get past him.

4. James Franklin and Wahab Riaz. In the few games that we do have records of at Pallekele left arm fast medium bowlers have been the most effective. This could just be coincidence or it could be something about the make up of the pitch. This game may just tell us.

5. Shahid Afridi. He has taken lots of wickets so far in the tournament, but his average against New Zealand is a pedestrian 47.78, which is particularly odd given that he is a leg-spinner and generally New Zealand batsmen have struggled against leg-spin.

If I was going to be betting on this game (and I probably won't be, it's too hard to call) I would be looking at putting money on top innings run scorer. I'd look at someone who was good at playing pull, hook and cut shots, given that the pitch is likely to bounce a bit. Perhaps someone like Umar Akmal might be a good look.

Saturday 5 March 2011

Dance of the Titans?

Australia have looked ominous. Sri Lanka have looked like they are capable, and now they meet. The first match between two really good teams.

The Sri Lanka Pakistan was merely shadow boxing. Without Malinga and Mendis Sri Lanka lose one of the things that makes them what they are.

Here is my guide to things to look for in this match:

1. Ajantha Mendis. He has been one of the big stories of the last 4 years. When he burst onto the scene in 2008 everyone was talking about him. And yet he has never played an ODI against the Australians. We don't know if they will be able to read him or not. It's one thing to play him in a T20, where you don't have the luxury of time, but the decision making process for a batsman is different in an ODI, and we will finally get to see him in action. Interestingly his stats when bowling first and batsman don't have scoreboard pressure are much worse than when bowling second. so the toss may be vital to his success.

2. Michael Clarke. It was hard to watch or listen to any Australian sport show without hearing someone calling for Clarke to be dropped. Everyone was saying that he was so badly out of form that he shouldn't be there. There was even a teenage fan manage to get into a press conference and tell him so. Well the advice seems to have worked. In his last 5 matches, Clarke has averaged 109 at a strike rate over 90. That's not bad. And Sri Lanka are close behind Pakistan as his second favourite opponent. His form has returned at the right time, and now we get to see if it is truely good form, or just a statistical blip.

3. Captaincy. Ponting has been criticised for the way that he captains his teams, for being too conservative, and not attacking enough. During the last 5 years, Ponting has got less wickets per run out of his bowlers than any other player to have captained Australia other than Michael Hussey. Likewise Sangakkara has not been nearly effective at gaining wickets for his bowlers as Jayawardene was. However there is a significant difference. When things go wrong for Ponting, he tends to get more conservative. Sangakkara tends to get more unorthodox. It will be interesting to see what the pressure does here.

4. Batting First. The team that bats at Premadasa normally wins. More than 75% of matches over the last 5 years have been won by the team batting first. However, in day/night games the scores are almost even, so it will be interesting to see what happens near the end of the first innings. Which ever team finishes the first innings best is likely to go on and win the game. If you are someone into in-game betting it might be a good idea to back the underdog if it's at something like 2.50 at the 40th over. There is a high chance that it will change to their favor over the next 10-15 overs.

5. "Sri Lanka haven't lost until Angelo Mathews is out." Tony Grieg famously said, just before Mathews went on to single-handedly destroy Australia at their last meeting. Can he repeat the dose tonight?

Sneaky 20c bet: Michael Clarke to top score for Australia at $4.75

These are the things I'm watching for. What are you looking out for?