Category Archives: predictions

Mahela Eats His Words

Posted on by

You should always be careful writing things like this when you still have games left to play:

[Cook's] got his own game and he can’t really get away from that. He just has to keep believing in himself and keep playing that role, and probably improve on it as well. Over a period of time, once you’ve played a certain number of matches, you find different ways of playing in a better way without getting away from your own game, and that’s something Cook has to find by himself.

He has to stick to his own game rather than try to hit big shots and get out of shape, as he did towards the end of his innings at Lord’s. You have to trust your ability and trust the shots you’ve been playing throughout, and hit those gaps – simple as that. It’s tough when a guy like Lasith Malinga is coming at you, but every player has to trust his own ability and not try to be someone else.

So after this was written, Cook scored 95* from 75 balls to secure a crushing 10-wicket defeat while Mahela scores 9 off 12. Surely, someone who has scored 18 Test centuries before the age of 27 deserves a bit more respect. What is more, the man had already made 48 off 52 in the second match and had just scored an ODI century against your team while the rest of his team was failing (119 versus the next highest score, 41) and his strike rate matched the innings run rate! Cook undoubtedly has some talent and even if one-dayers are not his strong suit, he doesn’t need this kind of patronizing.

The Sangakkara Speech

Posted on by

Unlike Ducking Beamers, I would recommend just downloading the PDF of the transcript of Sangakkara’s speech as you might be able to get through it in a few minutes than spend an entire hour. Having said that, I think it is a good speech. I was fascinating by much of the Sri Lankan history (both political and cricket) and there are some things I didn’t know about and have marked down for some Wikipedia explorations.

Now, everyone is talking about Sangakkara’s blasting the cricket authorities in his country. But to be honest, I think there is a fundamental disconnect between many parts of the speech. On the one hand, he criticizes the politics. On the other hand, he says that Sri Lanka has been successful and they have produced amazing cricketers and been successful. From a few readings, I think the point he is trying to make is that whenever the Sri Lankan team failed (as in 1999), it was the fault of politics, and whenever they succeeded, it was despite the poor administration or because the administration at that time (before 1996) was excellent.

You really cannot have it both ways. Look, the idea behind a cricket administration is to produce results. Whatever their dysfunctional behavior, if they produce results, then they are doing something right. If they are not, then they should be replaced. I guess Sri Lanka is not perhaps the greatest cricket team on Earth right now, but their achievements are not trivial. Just to give one example, Sri Lanka reached the last two finals of the ODI World Cup. They were up against a world class Australian team in the first one and had to play a “we owe it to Tendulkar” India in Mumbai. But suppose Sri Lanka had won that single match earlier this year. They would have been World Champions. Would this Sangakkara speech make sense if that were the case? After all, it is the same board and it is the result of one match.

Also, in the Test arena, Sri Lanka has come a long way. They are extremely strong at home and, I think, haven’t been beaten by anyone other than Australia there. They have also achieved many victories abroad in recent times. Unfortunately, they haven’t quite made it to the top tier, losing some important series (versus India and England) recently.

I think that is perhaps what is behind Sangakkara’s outpouring. After their World Cup win in 1996, I think many Sri Lankans were very optimistic about their team, but their progress has been more gradual than spectacular. There have been achievements, but also setbacks, which is to be expected. Right now, no team can take a match against Sri Lanka for granted unlike in the early 1990s. Their Test team is also strong, but hasn’t become competitive yet against Australia or India abroad. Some of their star players (Murali, Jayasuriya) are retiring without worthy replacements, but they forget that both Murali and Jayasuriya took time to become legends.

Also in the long run, I think Sri Lanka will hit limits it cannot overcome, namely GDP and population. Sri Lanka is a country of population 20 million. It has a higher per capita GDP than India and Pakistan. But long-term, both those countries will develop faster and catch up and their vast population (more of whom can afford to play cricket at a younger age) will make a huge difference in terms of talent availability. This is not something that can be fixed by a better cricket board, something that a country like New Zealand is slowly discovering.

The Qualification Story of England, Pakistan and Netherlands

Posted on by

Achettup and Q beat me to the qualification analysis for what Pakistan need to do to reach the semis. I will try to add a little more context and explanation to the whole situation.

So, why is everyone saying that England is through to the semis? Won’t the last match affect England’s chances?

To understand this, let us take a look at the current Points Table

Team          M-W-T-L   Pts     NRR
England       2-1-0-1     2    1.18
Netherlands   1-1-0-0     2    0.05
Pakistan      1-0-0-1     0   -2.40

There is only one match left: Pakistan v Netherlands. If Pakistan lose the match, they have zero points and are out of the tournament. If Pakistan win, all teams will be on 2 points. However Netherlands’ NRR will decline and will be below England. So England will at least be No. 2. The question is whether Pakistan can improve their NRR to go above Netherlands.

Why does Pakistan need to win by at least 25 runs even though they lost to England by 48 runs?

Because they don’t need to overtake England’s NRR. They only need to overtake Netherlands. And because they are both playing the same game, Netherland’s NRR will decline at the same time as Pakistan’s NRR increases because of a greater margin of Pakistan’s victory. So even if Pakistan’s NRR stays negative, it is okay as long as they can overtake Netherland’s NRR.

Here is the equation for those mathematically inclined. In case Pakistan bats first and then beats Netherlands, it would mean that 20 overs would be counted against both teams. Being bowled out is considered playing out 20 overs. Overs are only not counted if a team wins batting second with balls remaining. Let us assume Pakistan makes “P” runs and Netherlands makes “N” runs

Pakistan Net Run Rate, PNRR = (137 + P – (185 + N)) / 40

Netherlands Net Run Rate, NNRR = (163 + N – (162 + P)) /40

We need PNRR > NNRR

Solving, we get: P > N + 24.5

Since there are no fractional runs (!), Pakistan has to score at least 25 runs.

I want a graph of how quickly Pakistan has to score the runs, not a graph of the required run rate.

Here you go. Netherlands’ target is the X-axis and the overs is displayed on the Y-axis.

pakistan

Can I have the data used to populate the graph?

Here it is, the first column being the Netherlands score and the second being the maximum number of overs that Pakistan should bat.

Neth	Overs
0       14.3
10      14.5
20      15.0
30      15.2
40      15.3
50      15.4
60      15.5
70      16.0
80      16.1
90      16.2
100     16.3
110     16.4
120     16.4
130     16.5
140     17.0
150     17.0
160     17.1
170     17.1
180     17.2
190     17.2
200     17.2
210     17.3
220     17.3
230     17.4
240     17.4
250     17.4

But I want to know what if Netherlands score 133 runs?

In any case, use the following equation to derive the number of overs beyond which Pakistan will be out of the tournament if they don’t hit the target. That value is equal to (40 (301 + 2N) / (348 + 2N)) – 20, where N is the number of runs Netherlands scored.

You can derive it from using PNRR > NNRR with

PNRR =  ((137 + P)/(20 + x)) – ((185 + N) / 40)
NNRR = ((163 + N) / 40) – ((162 + P) / (20 + x)

where “x” is the number of overs. Now substitute P = N + 1 (one more than the Netherland’s score)

Will Pakistan achieve the target given all we know?

Netherlands’ better bet would be to bat second, given their performance against England. Remember that England had reached 100 for no loss before Netherlands pulled them back. So they would be better chasing. Pakistan would also like to be chasing since they know what they have to achieve. Considering that, the toss is extremely crucial. At this moment, I think both captains will choose to field first, but you never know based on the pitch and weather conditions.

All said and done, Pakistan have the better team man-to-man. They should be able to back themselves to win this match and by the required margin, regardless of whether they bat or bowl first.

Chennai Surge to the Top

Posted on by

Have the Deccan Chargers peaked too soon? After winning 4 matches on the trot, they have now lost three in a row. Meanwhile, Chennai have recovered from a poor start to the tournament with their third consecutive victory. Has the IPL turned on its head or what?

Here is the Projections table once again: Delhi, Chennai, Deccan and Punjab are now projected to enter the semis. The most important figure in the chart is the 14.15 value which every team has to cross to each the semis. Since we only have whole numbers for points, teams wishing to qualify must gain at least 15 points, i.e., 8 wins or 7 wins+abandoned match. The top five teams are half-way there. Chennai still needs to win 3 of their remaining 6 matches..

Team        M-W-D-L  P   P/M   FRR   ARR    NRR  Prob  Final

Delhi       6-4-0-2  8  1.33  8.12  8.04   0.08  8.98  16.98
Chennai     8-4-1-3  9  1.13  8.39  7.03   1.36  6.35  15.35
Deccan      7-4-0-3  8  1.14  7.64  7.60   0.04  7.02  15.02
Punjab      7-4-0-3  8  1.14  7.37  7.28   0.09  6.15  14.15

Bangalore   8-4-0-4  8  1.00  7.18  7.38  -0.20  6.04  14.04
Rajasthan   7-3-1-3  7  1.00  6.22  7.25  -1.03  6.90  13.90
Mumbai      7-3-1-3  7  1.00  7.68  6.91   0.76  5.94  12.94
Kolkata     8-1-1-6  3  0.38  6.62  7.80  -1.18  5.08   8.08

Chennai did themselves a lot of good with this win. First, this was one of their tougher remaining matches and they aced it. Secondly, their big victory have given them the best Net Run Rate in the League, a huge advantage if NRR comes into play to decide the semi-finalists. They have made Deccan a little less complacent, which may make it tougher for other opponents to ease to a big victory over Deccan. (That is not a given – look at Kolkata)

In light of today’s victory, a look at the next few matches:

  • Chennai’s victory has put more pressure on Rajasthan (lying just above Kolkata) to register a win against Punjab. A loss for Punjab is not critical, but they are running out of easy teams to beat. Rajasthan’s weak batting is there to be taken advantage of, as long as Punjab can get rid of that pesky Yusuf Pathan first. Once again, it is the battle of the Pathans.
  • The last couple of results have helped inch Delhi upwards. Delhi are the best placed team now with 8 matches in hand, among them two easy ones against Kolkata. They should close out those matches and if so, the battle will be between the remaining 6 teams in the League.
  • Like Rajasthan, Mumbai have to win one against the Top Four and no better time to beat someone (Deccan) than when they are down. That loss against Bangalore was unexpected and Mumbai have to compensate somewhere.

Mid-point in this tournament, one would have expected a clear leader. But this has been the weirdest tournament ever. The only team whose fortunes we can be sure of is Kolkata who will formally exit the tournament after their defeat to Delhi tomorrow. Unless, they win to prolong their pain or make the greatest comeback in sports history. Not gonna happen.

Punjab Squeak a Win, Kolkata Almost Out

Posted on by

Poor Kolkata Knight Riders! This is the second time they had a great chance to win in the last over and fluffed it. Their list of losses includes this last ball loss, the Super Over loss and another defeat by 9 runs, which, if they had closed out, would have made them the current leaders in the Points table. But that was not to be and Kings XI Punjab is now tied with the leaders Delhi Daredevils and Deccan Chargers, albeit with one fewer game in hand.

One more look at our Projections Table:

Team        M-W-D-L  P   P/M   FRR   ARR    NRR  Prob  Final

Delhi       6-4-0-2  8  1.33  8.12  8.04   0.08  8.58  16.58
Deccan      6-4-0-2  8  1.33  8.11  7.39   0.73  7.98  15.98
Mumbai      6-3-1-2  7  1.17  7.72  6.67   1.05  7.14  14.14
Chennai     7-3-1-3  7  1.00  8.31  7.37   0.94  6.98  13.98

Punjab      7-4-0-3  8  1.14  7.37  7.28   0.09  5.83  13.83
Rajasthan   7-3-1-3  7  1.00  6.22  7.25  -1.03  6.79  13.79
Bangalore   7-3-0-4  6  0.86  7.04  7.37  -0.33  6.63  12.63
Kolkata     8-1-1-5  3  0.38  6.62  7.80  -1.18  5.14   8.14

Punjab is still not projected to reach the semi-finals. The problem is that they only have one more game against the Bottom Half of the Points table, that too against unpredictable Rajasthan. They have two matches against Chennai and Deccan each. At this time, they are projected to gain just below 6 points from their remaining 7 matches.

Kolkata is almost out. Our projections now give Kolkata no chance of being No. 1 or No. 2 in the Points table because they can only gain a maximum of 12 points from their remaining 6 games: a total of 15 points. Another loss means that they are out as they will only end up at 13 points. Practically, they are certainly out because it is difficult to envision a world where Kolkata goes on to win the next 6 matches.

When we talk about Projections, it is important to keep in mind that for each team, it is not about its own victories and defeats that matter. The other matches also concern it because each match has 2 points that will be distributed among those teams – victory/defeat or abandoned. So, for example, Chennai is now projected to be No. 4 with around 14 points. But if Chennai loses the next few matches, it may not make much difference to the No. 4 position because another team will takes its place and be projected to get a similar number of points as Chennai.

By the time we come down to 2-3 matches each for every team, Projections will be less of a help because it only deals in probabilities which do not work so well with a small data set. At that time, we can switch to combinations to identify the path that every team needs to take to qualify for the semi-finals. At this point, it is not feasible as we have to consider 2^29 combinations which is kind of overkill. With 3 rounds left, it will be a more reasonable 2^12 ~ 4000 combinations. Not too bad.

It does seem very likely that Net Run Rate will be more important than wins and losses. But at this stage, teams would perhaps be better served by ignoring it and winning more matches. Teams that win at least 7 matches in total will have placed themselves in a very good position to qualify. And winning matches tends to improve the Net Run Rate anyway.

The Graveyard of T20 and IPL Predictions

Posted on by

I sincerely feel sorry for Ron over at cricketanalysis.com. He has been trying very hard to get his predictions for the Indian Premier League matches correct by tweaking the necessary parameters, but he seems to have hit a dead end, one reason being the illogical decisions that teams make. As he says,

I have learned quite a lot from doing the predictions over the first 11 games, even though my record is pretty dismal. The first thing is to never assume that teams will play their best lineups, by “best” meaning the lineup that I think is best, or the lineup that I think common sense would dictate. [...]

Another thing I have learned is that teams will never stick to five bowlers only if they have other people in the lineup that can bowl, no matter how much better or worse their part-timers are than their regular bowlers.

UPDATE: Apparently, he quit making IPL predictions around the same time I was writing this post – I noticed this after I put up this post. Here he goes:

No more predictions for IPL games. [...] Probably I will go 10-0 from now but whatever. Actually if things keep going against the model so accurately, it might have some value for fade purposes. [...]

One big problem I have is that fielding isn’t measured well by stats. For international cricket this isn’t a problem, because bowlers bowl with their international team consistently and so the collective fielding abilities of the side are incorporated into the bowling statistics. But the IPL is much different obviously. Maybe I will look at a way to measure fielding. I have some ideas.

In any case, Twenty20 is incredibly difficult to predict, primarily because there are not enough overs which prevent us from applying standard cricket ideas. T20 is evolving right in front of eyes and even the players and captains are making their best effort to understand what strategies and tactics work.

tragedy-v-statistics

In T20, one over can change the entire complexion of the game. We saw this in clear terms during the Rajasthan-Kolkata match. Rajasthan was easily in front when Munaf Patel effectively gave away 13 runs in one ball. Kolkata should have won the match in the final over, but Kamran Khan bowled a tight over against the flow of events and took Rajasthan to the Super Over where they won. In 50-over matches, a bad over is not necessarily a cause for surrender. Teams have easily recovered from the depths because of the large number of overs.

The smaller number of T20 overs means that wickets have an additional significance in terms of balls lost. If a team loses 10 wickets, that is almost 2 overs lost. It could mean the difference between a tough score of 180+ and an achievable 160-odd. This means that wickets in the last quarter of an innings can effectively kill a team’s ability to post a challenging target.

Finally, don’t understimate the effect of the “tactical” timeout in the middle of the innings. In quite a few matches, it has made a tremendous difference in the final score. What we have seen repeatedly is that big starts do not seem to materialize into huge totals.

Nice to know that Ron is a fan of Baseball Prospectus which uses statistical analysis techniques. I would like to see more of that to cricket, but I believe that Twenty20 is not mature enough to allow that possibility. Test and ODI cricket are much better avenues for predictions.

[Photo licensed from disrupsean]

Preview of India – New Zealand Test Series

Posted on by

Based on the strength of the teams on paper, the India-NZ Test series, which starts on Wednesday, should be a cakewalk for India. The batting order of India is one of the strongest in the world (Sehwag, Dravid, Tendulkar, Laxman, Dhoni, to name a few) and their bowling is not bad either when compared to the New Zealand team.

However, the history of India in New Zealand has been poor, especially the much-vaunted Indian batting which crumbled like a house of cards in an earthquake. While India has been the only team to consistently perform in Australia this decade, they have been unable to replicate the same at the Aussie’s neighbors. The problem has been the inability of the Indian batsmen to adapt to the pitches and conditions in New Zealand. This may not be as much of a problem as usual because the team has been playing in New Zealand for almost a month. However, six of the Test team were not chosen for the ODIs and T20s and do not have any live match practice. 

The recent Indian successes have been attributed by many to Dhoni’s captaincy and management style. It should be seen how he manages to inspire this team to perform better and win the series. Many of the youngsters in the team have much to prove. For Dravid, this is a do-or-die series. He was under pressure in India’s last series, and scored a century to ease some pressure off. Now he has to show that the century was not a flash in the pan.

This is a 3-Test series unlike recent series with other teams in New Zealand.  That will provide more excitement over a longer duration. Also it will allow a team to come back if they fail to start off the series properly. I suspect that would be India, which is most likely to crash and burn in the first Test. And if that happens, we will have to see if and how they come back.

Will SA Choke Again?

Posted on by

As the history of SA’s World Cup difficulties show, they come ever so near the prize and then self-destruct. After managing to defeat Australia at home, all they had to do was hold their nerve against an inexperienced new Australian team at home and lay claim to the No. 1 title without anybody objecting. At the end of the 2nd day in the second Test against the Australians, they are already looking at a demoralizing defeat (Test and series), their captain out for several weeks and perhaps even a whitewash.

What a change from just a few weeks back when South Africa had beat Australia in the second Test to clinch the away series and then beat them again in the one-day series. Even without their captain, Graeme Smith, they won the ODI series with a match to spare. Australia, on the other hand, seemed beaten down and failed to even beat the New Zealand team in the following one-day series.

The key to Australia’s comeback has been the excellent performance of their debutants and Mitchell Johnson. Even without his bat talking in this match, he has made an impact, striking fear into the hearts of the South African batsman, like we have not seen in some time. Today has been a ferocious advertisement for hostile fast bowling.

South Africa’s hopes lie in some amazing heroics by Jean-Paul Duminy and Dale Steyn. They have done it before, but history rarely repeats itself and not so soon either. With only 10 batsman in the team after Smith’s injury, they would first have to get very close to Australia, then bowl them out for nothing and then not do the same as the Australian batsmen. Tall order. Possible, but improbable. If it happens, that would be one of the greatest comebacks in Test cricket. 

Kridaya Prediction Meter: Why England Will Beat West Indies

Posted on by

On this blog, we are starting predictions for various cricket tournaments, starting with Test series between major Test-playing nations. The Kridaya Prediction Meter will start with basic statistical information and then we will move towards a more rigorous statistical analysis using more data. The first analysis will be about the England tour of West Indies.

west-indies-england-cricket

For the first installment, we are going to simply look at the batting averages of all the players in the two teams and see what we come up with:

England and WI Batting averages

Strauss 42.37 Gayle 39.60
Cook 42.70 Smith 24.30
Bell 41.48 Sarwan 39.58
Pietersen 50.48 Chanderpaul 49.71
Collingwood 42.04 Marshall 22.09
Flintoff 32.23 Nash 54.00
Prior 40.62 Ramdin 20.60
Broad 33.90 Taylor 16.66
Sidebottom 15.64 Benn 11.00
Harmison 11.63 Powell 7.10
Panesar 5.50  Edwards 4.45

As you can see, man-for-man, every person in the England batting lineup (except for No. 6) has a better batting average. The single exception (Nash) has only played 2 Tests in the recently concluded WI-NZ series. A better comparison would be the first-class batting averages of Flintoff (34.41) and Nash (30.93), where Flintoff comes out ahead. So, there is simply no comparison between the batting orders of the teams.

If you add up the batting average for both teams, here is what we get:

   England     359
   West Indies 282

That is a difference of 75 runs per innings and 150 runs per Test. A huge gap between the two teams. If West Indies have to win, England has to perform way below their on-paper strengths and West Indies have to rise above. The only bright point is that Chanderpaul has been playing well above his average in recent times, and the West Indies seem slightly resurgent with a strong draw in New Zealand. The top five in the England team have not played a single Test in the West Indies before this series. But that is about it.

We confidently predict England beating West Indies in this series.

[Photo licensed from nicksarebi]