Eight points really stood out to me as interesting and applicable to other areas of life. Beneath this short summary, I've highlighted four ideas in more detail.
You can put analytical methods into four categories, ranked in descending order of sexiness:
I want to spend a little time talking about the middle two. Whenever anyone reads Moneyball, or some other funky version of analytics in sport, the thing they then want to look for is the counterintuitive truth. They want to unearth the stat or piece of data that turns conventional wisdom on its head. It’s the question journalists always ask you, it’s what the coaches and players who buy into these methods want you to come up with. Now those little gems do often exist–we’ve covered a few of them in this book. But to fixate on that method of using analysis is to miss out on a simpler and often more useful means of gaining a competitive advantage from data, which can be summed up as ‘the ranking of the obvious’. I often explain this idea using a thought experiment I call ‘The Obvious Blackboard’.
Let’s say you are a coach and you want to make your team better. You decide to start by taking a blackboard and scribbling down everything you can think of that helps teams win.
You could fill that blackboard with ideas within a few minutes; with longer to think and a few colleagues to help generate ideas, you could easily fill multiple blackboards. And here’s the thing, pretty much all of those ideas will work. If you run the numbers, almost everything that sounds like it will help a team win, does. Taking more chances in the field correlates with winning. Scoring more runs in the PowerPlay correlates with winning. Taking more wickets in the middle overs correlates with winning. (Occasionally one of the obvious things doesn’t correlate with winning, so that’s your counterintuitive truth, and if you’re really lucky it will be something that you can use to your advantage as a tactic, game plan or selection strategy.)
But that doesn’t mean that all the other items on your blackboard are equally valid, and it certainly doesn’t mean that you can do them all. You can’t set up a side to do everything you’ve written on your board. You’ve got to choose which two, three or four ideas to focus on. So then the value of good analysis is to identify the things that are disproportionately important; to put a circle round the things on the board that are relatively easy for you to achieve but make a significant difference to your chances of winning.
It’s worth noting that this may not be the same for every team. Some teams will have resources available that allow them to pursue a certain strategy, whereas for other teams that same strategy is a non-starter because they simply don’t have the players to pursue it. There isn’t one correct solution to how to play winning cricket.
Playing poker is, at its core, an exercise in risk-management. The aim is not to win every hand, or even to only bet on winning hands; it is to play each hand in such a way as to make sure that you are maximising your long-term returns.
‘All decisions are investment decisions. Even if you are only deciding what to do with your time, you are making a decision about how to invest your finite resources so as to produce the best outcome, and generally doing that in an environment of uncertainty and insufficient information.’
Let’s look at three different betting options.
Bet A has a high chance of success and breaks even. Bet C is a long shot; it pays off handsomely if it comes off, but is actually a pretty poor gamble that will lose money in the long run. Caspar says that these two, Bet A and Bet C, typify how most of us live most of our lives. We tend to stick to ventures that have a high probability of success, even if the rewards are not particularly exciting. Then every now and again we flick into gamble mode and take on a low-odds long shot at a big prize that has very little chance of success. We put our savings away safely in the bank, and periodically buy a lottery ticket.
The only value bet here is Bet B. It fails more often than it succeeds, and the potential rewards aren’t large enough to excite us into a gamble, so it is the type of option that we tend to overlook or shy away from. But, partly because of this, it is exactly the type of option that will succeed in the long run.
Looking at the risk–reward curve, you can see that every point on the line represents a bet that will in the long run break even. Any point above the line represents a venture that will on average make money. The further above the line, the more money it will make.
So competition drives you downwards on the graph whenever you are above the line. Risk aversion drives you to the right. The higher you are, or the further to the left, then the stronger these forces are. This means that the majority of sporting strategies and techniques lie in the shaded area, the ‘status quo’, at the bottom right-hand corner of the graph.
This being the case, whenever you innovate, whenever you produce a new approach, you are likely to have to move to the left (and hopefully, if it is a successful strategy, upwards), i.e. most genuine innovations almost inevitably carry a higher level of risk than the established methods. ‘And here is the duality,’ explains Caspar Berry [former Poker player, now business consultant]. ‘Of course we all want our ventures to succeed, we want to turn the handle . . . and for the client to get the right advice, for the ball to hit the target, for the burger to be served without anyone getting food poisoning. But the effect of that perfectly natural instinct towards success is that we all operate in this pretty small area, which in sport you would call received wisdom, which I would call the comfort zone . . . and the visual impact of the graph is to see this enormous empty space, which is unexploited opportunity. ‘ROI [Return on Investment] is measured by your distance above the curve, and so there are enormous areas of potential ROI that go unexploited.’
Caspar explains the process of breaking out of this status quo using what he calls the innovation triangle. ‘If standard technique, or your comfort zone, are in the bottom right-hand corner, most successful innovations will necessitate a move up and left. This is the adoption of a new strategy (which is riskier but brings greater rewards). Then there is a movement to the right, which is honing and perfection (the success rate of the new method rises with practice and refinement) . . . and then the irony is that it is only when it has been refined that it is stolen. While it exists in a volatile state, it looks like the guy who won the poker tournament, it looks like they’re just getting lucky.’ But as soon as it starts to look like a reliable, profitable venture the forces of competition are brought fully to bear on it and it starts to move downwards.
For Caspar, the key concept about innovation and risk taking is the idea of the negative metric. ‘Whenever someone innovates in business or in life, they almost inevitably do so by accepting a negative metric that other people were unwilling to accept.’ The batsmen who started playing the reverse sweep (and the reverse hit) in Test cricket were not the players who invented the shot and first played it. Nor were they the only players of their generation who knew about the shot and could play it. The innovators were the ones who were willing to accept the negative metric that went with the shot. Initially this was the greater risk of failure, but more importantly the criticism that would follow if they got out doing it.
Studies have shown that ice-skaters who fall more often in training are more likely to win. If they are willing to take more risks in training, ultimately it will make them better skaters. But the negative metric that they have to accept is the increased number of training accidents and the greater risk of injury that comes with it.
In many spheres of life, as with Ryanair, the negative metric is public criticism, or, in other cases, a negative impression of your competence among those that matter to your continued success or employment. The conservative instincts that buttress received wisdom are often very powerful. And this is particularly true in professional sport. If you go against the traditional methods of success you may well be seen by many not as an innovator or pioneer, but merely as someone with suspect judgement, or someone who clearly doesn’t understand the game very well. The very scrutiny and attention that professional sportsmen perform under puts pressure on them to conform.
We have mentioned already that where the theoretically optimal tactics differ from what the professionals do in practice, then the received wisdom almost always carries less risk.
We are all loss averse, genetically programmed to be, and straying from the accepted path carries a twofold risk, firstly the increased failure rate of the innovation, and secondly the increased criticism and loss of standing that will accompany any failure. It is no wonder that there are so many Tethered Cats still being tied up each night across professional sport. Any Tethered Cat you find might turn out to be one of Chesterton’s Fences. So what does this look like in other sports?
As a rule of thumb, whenever there is a difference between what theory suggests is the optimal strategy and what the professionals do in practice, you will find that the professional strategy is more conservative than the statistically optimal.
One of the keys to using analysis successfully to improve performance is a willingness to take risks, because the evidence-based approach is almost always riskier and more aggressive than the received wisdom. There are any number of examples from lots of different sports, but some of the most well-documented are: shooting three-pointers in basketball; running on the fourth down in American football; we cover in Chapter 7 the sizeable, if counterintuitive, advantage of bowling first on flat wickets; we looked in detail at England’s use of hitherto unseen levels of batting aggression to win a World Cup in the previous chapter. Eoin Morgan’s willingness to accept the negative metric was the key to building the team that could do it.
As training programmes improve around the world, and become ever better at giving young players secure basics and accepted models of success, the other thing worthy of consideration is this: the comparative advantage of being unorthodox increases as the proportion of orthodox players rises.
Throughout Test cricket’s history, umpires have consistently given more touring batsmen out LBW than home batsmen. With the exception of the 1940s (an anomaly, given that due to the Second World War only six Tests were played), every decade up to the 1990s saw LBWs make up a greater proportion of dismissals of away batsmen than home batsmen.
In athletics, the march of progress is obvious. Across the board, in all the different disciplines, comparing performances today with performances from 40 years ago makes that progress clear. Times get quicker, weights get heavier, and distances get longer. The Men’s 100-metre sprint record has been beaten in every decade since the 1890s (save for the 2010s), among the most clear and visible yardsticks for proving sporting improvement. As standards improve (conditions, sports science, professionalism) so do performances, but the other thing that happens when standards rise is that the amount of variation falls. As overall standards rise, the margins between performances of elite athletes get smaller.
In sports such as swimming and athletics, we can see the improvement in times and distances. But, in other sports, such as football and rugby, it is harder. The raw score tells us little about the standard of performance. A rugby match finishing 20–12 could just as easily be a Premiership final as an Under-15s school match. For those sports, the more obvious indicator of increasing standards is the decrease in variance. You can see a very clear example of this in baseball. Scoring levels have stayed relatively constant in baseball for 100 years. A good batting average in the 1930s is still a good batting average. But if you look at the standard deviation (a measure of variance, of how spread out the individual scores are) of players’ batting averages by season, it has fallen steadily year on year and decade on decade. Players have collectively averaged the same, but the averages of the best and the averages of the worst each year have got closer and closer together.
It is exactly the same effect that causes Olympic sprint times to crowd together. The better you get at something, the harder it is to improve further. As you push closer and closer to the limit of what is humanly possible, each incremental improvement becomes smaller and smaller and harder and harder to make.
You can see a similar pattern if you look at the distribution of batting averages in Test cricket. Although overall scoring levels have not changed much since the 1930s, the variance between players’ averages has fallen steadily. As the standard of cricket improved, the difference in batting average between the best and the rest in the Test game has fallen.
Decade on decade it got progressively harder for a batsman to average over 50 in Tests. That is, until about 1996. Then suddenly there was a reversal in the trend, and variance in Test batting averages increased. This, however, wasn’t a sudden reversal in the march of progress. The increase was most likely the first unforeseen consequence of the improvement in the quality of umpires and their decisions.
What this ‘experiment’ shows is that poor umpiring is a leveller. It makes it harder to distinguish statistically between good and poor batsmen. It reduces the variance in measured performance between the best and the worst. And as you would expect, the converse is true as well. Good umpiring increases variance. The better the umpire, and the more accurate their decision-making, the wider the gap between Player B and Player C appears. If you improve umpiring, the best average more and the worst average less. Which, of course, is exactly what happened when neutral umpires were introduced.
In the modern era, the advantage of winning the toss seems to have disappeared. This is, of course, very counter-intuitive. After all, one team bats first, then the other, and the two teams’ chances of winning are not equal. The team batting first has different requirements for victory to the team batting second, and the pitch changes over the course of the match, affecting the balance of power between bat and ball. Therefore, we would assume, teams that win the toss can choose the best conditions and so gain an advantage. But they don’t. How can that possibly be?
This is the problem with Tethered Cats. Sometimes, a perfectly reasonable response to current circumstances becomes a habit, then a tradition, then an article of faith that outlives the circumstances that created it. We rarely question what we know to be self-evidently true, particularly when everyone else is doing the same thing. And so the bias towards batting first seems to have outlived the circumstances that created it by several decades.
So captains’ behaviour at the toss, in particular in subcontinental conditions, seems to be a particularly clear example of a Tethered Cat. In truth, there are many such examples in cricket, where received wisdom doesn’t concur with the evidence: where what teams do doesn’t seem to maximise their chances of winning. Why is this the case? Part of the story involves how our brains handle information.
There has been a great deal of research into memory and perception, and the results are both surprising and illuminating when it comes to our decision-making in sport. For a start, our memories don’t work as you might expect. They are not akin to a videotape; we don’t record a series of events and then play them back as and when they are needed.
The disturbing truth is that our unaided recall is not very good. The human brain encodes less than 10 per cent of what we experience; the rest it simply makes up. Our minds construct a narrative around the coded memories we do have that fills in the gaps with a plausible story.
Faced with a huge number of random or near random events (a cricket match, for instance) our brains pattern-spot, even when there is no pattern. Our minds look for those events that they can form into a pattern or story, and that becomes the meaning or lesson that we take away from the match. Even if the vast majority of events that occurred didn’t fit the pattern, we disproportionately remember the ones that did.
At their best, our memories work like Albert Camus’s description of fiction; they are ‘the lie through which we tell the truth’. What we remember didn’t actually happen; what we remember is a story that our brains have fabricated, but one that we hope contains the essential truth of what happened in a way that we can understand and retain.
And it’s not just our recall that is faulty. It’s also our perception of events. We see everything through the prism of our previous experiences and beliefs. We look for what we expect to see and we are far more likely to spot examples that agree with our expectations than contradict them. There is a wealth of hard scientific data showing just how easily manipulated our perceptions are by context and emotion. We are, unfortunately, very poor witnesses to our own lives.
This is how data can help us when used well. It acts a bit like a video camera. Video analysis, with slow motion and freeze frame, has enabled coaches and players to see things they couldn’t otherwise see. It reveals aspects of technique that the unaided eye can’t spot. Used well, statistics can do the same, revealing patterns and trends over seasons and decades that would otherwise be all but impossible to spot.
And there of course you have the central problem of much decision-making in cricket. This pitch is slightly different to all the other pitches that there have ever been. And you don’t know for certain how it is going to play, or how that will influence the balance of power in the match. There are those who would argue that this is why stats are useless, or at best very limited. We would agree entirely that stats are never sufficient to make a decision.
There is nuance and subtlety to weigh; the brain and eye have access to information that the laptop doesn’t. The feel and instincts of coaches and players, the hardwired learning from decades in the game, contain incredibly valuable information and will always be the mainstay of decision-making that must be flexible and fluid through changing match situations.
But if we are honest, we must also accept that the sheer weight and tonnage of what we don’t know about how cricket works would sink a battleship. To use stats and nothing else to make decisions would be incredibly foolish, and as far as we are aware no one ever has. But equally, to insist on making decisions on incomplete information, without ever reviewing the effectiveness of those decisions, would seem almost equally perverse.
.svg)
