Stop playing badly!

Why do we criticise a player? Why are we so negative about our team-mates’ mistakes?

Because we’re idiots, and because we’re selfish.

There is not a lot of evidence that punishing mistakes aids learning or improves performance, and much, much better evidence that rewarding what people have done well does actually work¹. So why are we so negative with each other? Why do we scream at that guy for turning over that huck?

There are two main reasons – because it makes us feel better, and because we think it works. The first is selfish, and the second is wrong.

It makes us feel better because we can vent our frustration. We’re annoyed that our team turned over, and venting performs a number of useful functions:

• General release of frustration (the equivalent of punching a wall; just a release of emotional energy).
• Passing the blame (letting everyone know who the mistake was made by – particularly prevalent when you threw the pass that was dropped or made the cut the turfed throw was aimed at. This has an internal effect as well – you may just be reassuring yourself it wasn’t your fault as much as telling others.)
• Emphasising your own usefulness to the team (‘I was wide open for a reset when you threw into triple coverage…’)

It’s selfish to do all of that unless the venting also helps your team. A good team-mate wouldn’t vent unless he had reason to believe it would make the team perform better from now on. Venting is usually an ego problem.

But what about those of us who choose to vent (or who justify our uncontrolled venting) because we truly believe it helps our team? Generally, it seems like it does – most times when someone does something terrible and gets screamed at, they don’t do it again as badly, or as often. So how can we say that venting doesn’t work?

Well, we’re idiots.

Regression to the mean is a simple concept that is often completely at odds with our intuition. Let’s imagine we’re guessing higher or lower with a pair of dice. If we roll a seven, it’s obviously an even bet whether the next number will be higher or lower; but if we roll a three it’s equally obvious that the next number is much more likely to be higher.

When something ‘bad’ happens, like rolling a 3, or something ‘good’ like rolling an 11, it’s overwhelmingly likely that the next roll will be closer to the average. There are just so many more ways in which that could happen.

Imagine you shouted at someone every time they rolled a low number, or praised someone whenever they threw high. I’m sure no-one would argue that praise or criticism would affect their dice-rolling skills, but you would nevertheless see a clear pattern that bawling them out for low numbers resulted in an improved performance, and praising their success made them worse². Whenever they randomly did something bad, your screaming intervention appears to sort them out – but of course it we only shout at them for things that are surprisingly bad, then it should be no surprise when the next random event is ‘better’.

Throwing frisbees is also largely random³. So is catching, or choosing a throw to make, and all the rest of it. We only shout at someone when they disappoint us, when they do something worse than we expect. So we don’t need any explanation for why they got better the next time – regression to the mean will cover that. But we all automatically assume that our advice/exhortation/screaming is the causal factor, and we develop a belief that our venting helps the team.

The things that rile us are the outliers, the extreme events, the surprises. And by definition that makes it more likely that the next thing they do will be more like what they normally do. And what they normally do got them picked for this team…

That’s not to say that you can’t work with people at practice to help them improve⁴; and it’s not to say you can’t give useful advice even in a game situation (‘They have a deep poach, careful on the hucks,’ or ‘From this corner it’s straight downwind, it will be hard to float that,’) because of course it’s not totally random – some of the factors that affect particular mistakes are in our control and can be taken on board even in the heat of battle.

It’s not even to say that people never need to be made to face up to their inadequacies – a stern talk from a respected coach, captain or peer can be very helpful⁵.

But the non-constructive, heat-of-the-moment attack on someone who makes a mistake is selfish, muddle-headed and detrimental. Players in a game situation who are thinking about their failings, or feeling their ego threatened, or focusing on past mistakes, are wasting cognitive resources on things that won’t help them perform. If you think it is helping your team when you attack players – and there are people out there, including established managers and coaches in serious professional sports, who do entirely believe this – then you need to think a bit about regression to the mean.

We’ve all seen people get shouted at for throwing a risky turnover. Most of the time they randomly get better, but there are two common non-random outcomes and they’re both bad:

1. The player decides to prove you wrong, to prove his value, and throws yet more risky stuff.
2. The player goes into his shell, second-guesses every choice, and starts turning over dump passes. He delays every throw while he consciously evaluates whether it’s ‘safe’ enough, and ends up too late in everything he does. More mistakes come, and the whole thing spirals downwards.

If you have a player who consistently does things that are not good enough, then don’t pick her. If you’ve decided to pick her, then come game-day you need to support her and let her play with confidence. Don’t put her in a situation where her conscious brain is trying too hard to control her decisions and her skills, because the conscious mind is too slow; good performance comes from the unconscious execution of the skills you’ve practised.

Support your team; foster an environment where they all feel valued and confident, and you will actually make a difference.

 

¹ If you haven’t read Daniel Kahnemann’s ‘Thinking, Fast and Slow’, then you should – and it has a nice story about this issue.

² Perhaps they got ‘overconfident’? I bet that’s what you think after that young player you praised does something less good afterwards… but the situation is pretty much identical to the dice. When you are surprised by the quality of someone’s execution, it’s vastly more likely that they will do less well next time. Don’t be afraid to praise people (though do work to make it ‘growth-mindset’ praise – see the work of Carol Dweck).

³ It’s a weighted randomness – there is of course an underlying level of skill that varies from player to player – but any particular instance could nevertheless be much better or worse than that average level. There’s still a possibility of occasionally throwing a ‘2’ even with loaded dice…

⁴ People do improve over time of course – that average level of skill, the base level about which our execution randomly fluctuates, can be moved. And we can improve consistency, meaning that the throws that are far away from our current average level are more rare. But the randomness is still going to be very significant occasionally –  the worst throw you made this year is certainly worse than the best throw you made in your first year playing. Improvement is real, and our average ability increases hugely over time, but there remains uncertainty about any particular instance of a skill.

⁵ But should normally happen away from a game situation. If you need to intervene at game time, you will need to be even more careful to impart the information in a way which does not threaten the player’s ego. Specific, actionable advice is often useful, but this is no time to make them feel like they’re letting everyone down.