A few big innings and many small; Ratings 12th August
Russell Degnan

4th TestEnglandvIndia
Pre-rating1039.11161.7
Form-40.2+20.3
Expected MarginEngland by 6 runs
Actual MarginEngland by an nnings and 54 runs
Post-rating1053.21140.5

Six ducks.

That more or less sums up India's opening day effort, losing four wickets for eight runs, all caught behind with poor footwork and hard hands. Anderson had the ball on a string, though it was Broad who took the bulk of the wickets, ending with 6/25. Dhoni's counter-attacking 71 was ugly, but somewhat effective, and only a partnership with Ashwin - almost certainly in India's best six batsmen, if not their best six bowlers - kept the scoreboard from being embarrassing.

India bowled quite well, for the most part. But they lost either side of the new ball, when Root and Buttler turned a handy lead into an impregnable one. They are also, somewhat inexplicably, unable to play Moeen Ali, whose bowling is handy, accurate, but not something to repeatedly collapse to, though the damage at the top was done by Anderson again. As hinted after the last test, India under Dhoni, when the game is drifting, don't fight very hard.

There is little else to add. History suggests India will again fail to turn up at the Oval, although they are not incapable of winning, nor England of losing, as was shown at Lord's. England's confidence is up now though, including, most importantly, that of Cook in his young bowlers, which both lessens the load on Anderson and Broad (if he plays), and prevents the sort of crises of confidence that afflicted Kerrigan and Borthwick. If India do lose at the Oval, it will be a sad end to a series that promised much, but ended as a mauling.


1st TestSri LankavPakistan
Pre-rating1038.41093.9
Form+15.3-19.1
Expected MarginSri Lanka by 22 runs
Actual MarginSri Lanka by 7 wickets
Post-rating1043.81086.8

Revenge for Sri Lanka, after Pakistan's heist earlier in the year, and in nearly identical circumstances. Younis Khan's magnificent 177 provided the guts of a competitive 451, but he was matched by the insatiable Sangakarra's 221 and Mathews 91. The 82 run lead gave Sri Lanka just enough to press for victory after a productive first session on the final day. Herath, naturally, provided the pressure, bowling taking 6/48 from 30.2 overs. They were held up late by Sarfraz Ahmed, whose 52 not out almost gave Pakistan enough time and runs to survive until the rain came. It did, almost literally as Sri Lanka walked off, having chased 99 in 17 overs.

The win doesn't change the rankings, but tightens the group of sub-continental rivals and England, with another two test matches to potentially shake things up. A series win for Sri Lanka would round out an impressive sequence of results in the past three months. And the impressive career of Jayawardene. His home record has always vastly exceeded that away from Sri Lanka, making it harder for his batting to be appreciated by anything beyond the raw numbers filtered through the scorecard. It is a record that speaks to a vulnerability to bounce, and to bowlers who could provide it; but that is a minor quibble against almost 12 thousand runs. Few players have ever dominated attacks when conditions were in his favour more than Jayawardene, and few have ever made it look as easy either.


Only TestZimbabwevSouth Africa
Pre-rating560.21295.4
Form+25.2-19.6
Expected MarginSouth Africa by 318 runs
Actual MarginSouth Africa by 9 wickets
Post-rating570.71289.1
Series rating723.31132.4

Even the dodgy internet streams couldn't get me coverage of this match, but the general consensus is of a game where Zimbabwe fought hard, without actually ever looking like doing anything but losing by a significant margin. Taylor remains their only batsman of class, although Mutumbami made useful contributions, and they'll be pleased with the wickets of Nyumbu and the discipline of Chatara. Zimbabwe have so many structural problems it is hard to know where to start: an inability to keep players; a lack of funds to run domestic cricket, pay players and train consistently; on-going governance issues and debt so deep they are almost insolvent.

This could have been much worse, but South Africa came to do the job, and did it, nothing more. What they didn't do is achieve the margin, which slips them closer to Australia in second place, with relative form likely to drag them closer still when each next plays.

Only Women's TestEnglandvIndia

There are no ratings for women's test matches, but it is worth discussing such a rare event. The format has been almost exclusively played by Australia and England for the past eight years, and the BCCI ought to be commended for getting their players to play the format. What will happen is less clear, as with little test cricket form to digest from either side, the only known factor is that India has struggled in recent tournaments, while England have defeated Australia in two Ashes contests, even if they've failed to win major trophies.

Not that finding out what is happening will be easy. The ECB has put money into professional contracts, but failed to invest in even basic streaming to help supporters follow the game. The BCCI haven't even done the former, and the game stagnates there, as does so much Indian women's sport. South-east Asian women's sides on a shoe-string are more likely to challenge the anglo-hegemony in a decade or two, as they have the willingness to push the sport to their female population. But with the other test nations investing and growing women's cricket, it may take several more decades to restore a sense of parity. This contest might be rarer still in another eight years.


Rankings at 13th August 2014
1.South Africa1289.1
2.Australia1273.7
3.India1112.0
4.Pakistan1086.8
5.England1068.1
6.Sri Lanka1043.8
7.New Zealand937.8
8.West Indies870.2
9.Bangladesh600.6
12.Zimbabwe570.7

10.Ireland594.8
11.Afghanistan587.6
13.Scotland430.3
14.Namibia383.4
15.Kenya276.4
16.U.A.E.257.3
17.Netherlands182.4
18.Canada147.9

Shaded teams have played fewer than 2 games per season. Non-test team ratings are not comparable to test ratings as they don't play each other.

Cricket - Ratings - Test 14th August, 2014 00:34:44   [#] [0 comments] 

Turning luck, turning pitch; Ratings 4th August
Russell Degnan

2nd TestSri LankavSouth Africa
Pre-rating1032.71302.6
Form+15.9-21.0
Expected MarginSouth Africa by 85 runs
Actual MarginMatch Drawn
Post-rating1038.41295.4
Series rating1091.11244.1

One of a string of outstanding games that Sri Lanka has played of late, this time coming down on the wrong side of a draw, having had South Africa against it for the final three days. Hashim Amla was the hero, batting for an incredible 541 deliveries, merely 164 runs, and one dismissal. It would nevertheless have been for nought had the tail not fought through 17 overs with as many as all 9 players around the bat, and the spinners working away. Herath and Perera performed their own herculean feats, the former bowling 90 overs in the match, taking 9/111, the latter 85.5 overs, taking 8/129.

The scoring rates show clearly both how difficult were South Africa finding it to make runs, and how comparatively easily the Sri Lankans scored of Imran Tahir (3/197 off 51). Sri Lanka batted a mere 175 overs, making 650 runs, while South Africa survived, and not much else, scoring 441 off 246 overs. Mathews declaration leaving 369 runs was, in retrospect, excessive, but with 107 scheduled overs to play, he could not depend on either rain or South Africa shutting shop.

Ultimately, he tourists took the series, and the mace, for which they must be commended. They retained number one spot on these rankings too, which seemed in doubt had they lost; their next fixture against Zimbabwe requires a monumental victory, but it would surprise if they failed to provide one. After flirting with passing England, Sri Lanka will need to beat Pakistan and hope India regain some form to climb into fifth.


3rd TestEnglandvIndia
Pre-rating1039.11161.7
Form-40.2+20.3
Expected MarginIndia by 11 runs
Actual MarginEngland by 266 runs
Post-rating1053.21140.5

After flirting with playing well enough to win, then ultimately failing all summer, England finally got the luck it needed in one fell swoop. This was a thorough smashing in the end, having declared twice for the loss of only 11 wickets, and wrapping up the match early on day 5, when even the optimists were predicting a day of hard slog. Anderson got his lines and more importantly, his lengths right, taking 7 wickets, and Moeen Ali took 8 in a performance that can only lead to future disappointment.

But the important thing was runs, loads of them, from Cook (dropped early), Bell (lucky to survive an lbw on nought), Ballance (who looks genuinely classy) and as they accelerated, Buttler (lucky, but with a license). That gave the bowlers some added rest before they got to work, and the time later to have a breather before setting into India again. By contrast, India were listless after a grinding first day that only saw two wickets, and positively generous while batting. The way England climbed all over them when they were behind was reminiscent of 2011.
Youth is not an excuse for a lack of backbone, not poor catching (particularly in the slips). It would seem inexplicable that Ashwin remain out of the side, as they need a spinner who can be relied on to rest their seam attack. His batting isn't a bad bonus either, for spectators and India. Both sides have gifted away wins to the other, which will make fools of those predicting the tests to come. This test, like the two that proceeded it, might have been quite different. England though, appear slightly superior with the ball, which should be enough.

2 TestsSri LankavPakistan
Pre-rating1038.41093.9
Form+15.3-19.1
Expected MarginSri Lanka by 22 runs

Another series predicted to be very close. Sri Lanka have a lesser home advantage against Pakistan, who can draw on the talents of Saeed Ajmal, and are accustomed to pitches of a similar nature. In their last match, Sri Lanka's negativity threw away a safe position, but Mathews seems to have learnt (a little) from this, and it will be a tighter tactical battle. If a turner, such as the one found for South Africa, is procured, both sides are capable of winning from a favourable toss. One a slightly quicker deck, Pakistan probably have a slight edge, if they can keep Sangakarra quiet.


Only TestZimbabwevSouth Africa
Pre-rating560.21295.4
Form+25.2-19.6
Expected MarginSouth Africa by 318 runs

A victory by more than an innings is predicted for this match, and given Zimbabwe's struggles against Afghanistan, it would be a surprise if it wasn't provided. Zimbabwe could surprise, and did so a year ago against Pakistan (who beat South Africa in their subsequent test), but turmoil seems to have enveloped their setup again, as finances are strained, and the players are unlikely to be anywhere near prepared enough to face the South African attack. Getting their southern neighbours to bat twice will be an achievement.


Rankings at 4th August 2014
1.South Africa12955.4
2.Australia1273.7
3.India1140.5
4.Pakistan1093.9
5.England1053.2
6.Sri Lanka1038.4
7.New Zealand937.8
8.West Indies870.2
9.Bangladesh600.6
12.Zimbabwe560.2

10.Ireland594.8
11.Afghanistan587.6
13.Scotland430.3
14.Namibia383.4
15.Kenya276.4
16.U.A.E.257.3
17.Netherlands182.4
18.Canada147.9

Shaded teams have played fewer than 2 games per season. Non-test team ratings are not comparable to test ratings as they don't play each other.

Cricket - Ratings - Test 4th August, 2014 22:54:28   [#] [0 comments] 

A sports draft without tanking
Russell Degnan

The odd thing about sports drafts is that, as currently constructed, they are really about who is bad. Because the best picks go to the worst teams, they reward and even encourage poor play; particularly at the back-end of the season, but increasingly across multi-year rebuilding strategies, across a whole season.

Needless to say, owners who aren't bad don't like it. Partly for financial reasons: in an NBA environment where 5 or 6 teams are tanking, a fifth of regular season games are pointless, which hurts teams both good and bad. Partly for fairness, as the draft is increasingly perceived as a lottery: a way to beat the sensible construction of a team by throwing the dice.

Tanking itself, does and doesn't work, depending on the specific situation.

Deliberately losing games (either by resting players or odd rotations) at the tail-end of the season will work a bit: it improves draft position which is worth more wins (though not necessarily a lot more), and if young players are tested, it improves knowledge of a roster's capability.

Drafting players who can't play - as the Sixers have - with the aim of staying early in the draft for multiple years probably can work, if the right players turn up. But it is unlikely trading good contracts for cap space (through short-term bad contracts) improves wins in the future. Free agents rarely provide more wins than their contracts, and a bad team is much more likely to have to over-pay than one in contention.

Unless you are Cleveland. Cleveland's third lottery victory in four years is the other driving force behind proposed changes to the NBA Draft. LeBron may of returned without Wiggins, but it made it an sensible choice for a player looking to escape from an old team to one he could build on. The combination of repeatedly making poor draft choices, and poor trade choices, and an unusual streak of luck sits poorly with NBA owners. And not because hey believe in competitive balance, but because they don't - and nether do fans.


When the league talks about competitive balance, they want every team to be a possible contender every year. Baseball almost gets to that, but basketball isn't close, and is unlikely to ever be close, because there just aren't enough star players. But like the concept of equality - on which it is based - it isn't true that owners want balance. They want dynasties and the chance to prove they are the smartest and canniest owners. In other words, not equality of outcomes, but equality of opportunities.

Any solution that encourages a team to play badly has a downside. It isn't clear that the current weighted odds are optimum. But as Zach Lowe points out, playing with the percentages of lottery odds on the margins incentives a different set of teams to play badly. But the alternative proposal, of a wheel that gave each team a set rotating pick, isn't necessarily better either, because it defeats the reason he draft exists in the first place: teams are not born equal.

Teams are unequal for many reasons: big markets have more cash, even with a salary cap limiting their options; teams with better players and coaches can attract better free agents; some teams are in better cities to live in. The point of the draft is to mitigate those effects by helping teams that are struggling (ie. losing). But helping down the track is a second order effort, equivalent to providing welfare, but no education. The alternative to a draft that incentivizes losing, is one that mitigates the effect of the most important starting assets: cash and draft picks.


Read Wages of Wins and it will tell you that wins are poorly correlated with spending. This is not because in a free market the biggest team couldn't buy titles - football proves they certainly can - but because in the NBA the market for players is distorted by rookie and maximum contracts and the salary cap, so a lot of the money goes to players providing fewer wins than they ought. Nevertheless, you can measure those distortions, and by extension, calculate which teams have benefited from circumstance, and adjust the lottery accordingly.

Big market spending is the easiest to calculate, because it basically comes to how much a team has spent over and above the salary cap. Every $1.6 million corresponds to one extra win, and therefore can be offset against expected value in the draft lottery.

The Draft is an uneven distributor of talent, because some years are better than others, but on the average, the value of each pick over and above their salary can be modelled, and adjusted over the previous four years. Both Arturo Galletti and Nate Silver both calculated similar numbers for the value of each pick: around $30 million over 4 years for number 1 picks, and $2-4 million for number 30 picks. Traded picks would still apply to the original team, as the trade was valued at the pick (presumably). That's the tedious but easy part.

Star players limited by max contracts, on the other hand is more complex, because although the number of wins earnt above their salary can be calculated, not all players in this category are on the max (LeBron for example) and a team shouldn't necessarily be penalised for finding the right guy. One option is to just ignore it, as it affects relatively few players (albeit the most important ones); another would be to apply it to designated players (either designated or acquired through trade) at the difference between the maximum 25% and the players salary. This is, broadly speaking, close to the expected difference in value, and subtly penalises teams with star players.

Playoff revenue advantage is the final piece. Teams earn more for making the playoffs, as well as gaining a free market advantage. At the same time, excessive offsets here would provide incentives to miss the playoffs. Keeping every team in the lottery means every team would look to make the playoffs, but adjusting for series played, reduces the advantages gained by being consistently good. Technically this money isn't used unless it goes into salary spending, but it covers teams who have attracted stars at below maximum salaries, closing a potential loophole.

Put it together and it produces something like this[1]:

The draft rank comes out roughly where it should, given the recent history of those teams: Utah, Philadelphia, Atlanta and Milwaukee have neither picked early, nor spent heavily, nor had deep playoff runs, and therefore find themselves high in the draft. Future years would see them drop, effectively replicating the wheel system for teams spending near the cap.

Conversely, it would be a difficult system to game, as there is a lower limit to improving draft lottery position: not making the playoffs, not exceeding the cap, not having a designated player. Many teams are in that position, so there is no value in completely blowing up a side to chase the number one pick. For borderline playoff teams like Atlanta, the value in reducing salary slightly, or skipping the playoffs would be a half a percent improvement in lottery odds. Not nearly enough to play for.

Equalization involves adjusting the lottery odds to offset the expected deficit from spending and recent drafts. Nine teams last season exceeded the average value of a draft selection, and therefore fall out of the lottery. The remaining teams have between a one and nine percent chance of winnings, as follows:

In the long term, such a selection would encourage teams to stay close to the cap, unless they can win - which is broadly what the NBA wants - and even out draft picks, rewarding teams who picked well, and giving no solace to those who don't. Teams would have a pretty good idea of their draft position several years in advance, because they would know their salary, their likely playoff position, and their previous picks. Trading for picks becomes slightly dangerous, because it allows the trader to up their salary without consequence. Or traded picks would likely be protected at both ends.

There is no perfect draft method, but this one would at least end tanking, and reward sound management. Which for owners, who are in the game to prove their savviness, and for fans, who want their team to win, not to lose, would be a big improvement.

[1] Draft value is taken from Nate Silver's piece, divided by four, for a yearly value. Designated players are listed in Larry Coon's cap FAQ. Team salaries from Story Teller.

Basketball 3rd August, 2014 14:23:15   [#] [0 comments]