One book that's had an especially big influence upon me is John Kay's Obliquity. He argues that in a complex and unpredictable world our goals are often best achieved indirectly:

Happiness is not achieved through the pursuit of happiness. The most profitable businesses are not the most profit-oriented. The wealthiest people are not those most assertive in the pursuit of wealth...Soviet planners managed the economy far less successfully than the adaptive, disorganized processes of market economies. (Obliquity, p8)

People tend not to fully appreciate this point because of the outcome bias. We tend to look at success and infer that those who achieved it had a conscious strategy for doing so when in fact they might just have gotten lucky. As Alasdair MacIntyre said:

When imputed organizational skill and power are deployed and the desired effect follows, all that we have witnessed is the same kind of sequence as that to be observed when a clergyman is fortunate enough to pray for rain just before the unpredicted end of a drought. (After Virtue p75)

This brings me to a question: does obliquity apply to the Labour leadership context?

What I mean is that the direct, non-oblique approach is to get a leader who is more "electable" - one who plays by the Westminster rules and who fits the mould of a "credible" leader. This is the appeal of Owen Smith.

By contrast, the Corbynistas' approach is the oblique one - rely less upon a leader and more upon a mass party, and hope that disenchantment with conventional politics leads to support for radical alternatives.

You might reply that Labour's poll ratings suggest this particular oblique strategy has failed so a change is needed. But this is not certain. For one thing, Corbynism is a long game: it relies upon a slow process of changing the ideological climate from the bottom up. And for another, there's no assurance at all that Mr Smith would greatly improve things. Like any leader, he'll face massive opposition from both the media and ideology. And he might well suffer big drop in party membership.

(There's also the possibility that a "coup" against Corbyn would alienate tens of thousands of bright, energetic young people whom he has inspired to become political. I don't like the potential long-term effects of this.)

It could be, therefore, that if Labour changes leader it will see the same results as when a failing football team changes coach - which is to say no change in the longer-run.

The fact that you're losing does not turn a bad bet into a good one.

Now, I don't say all this to defend Corbyn. My point is that, given the fluidity, complexity and unpredictability of politics, I just don't know which of the direct and oblique strategies will work.

But there's something I do know. One lesson of Corbyn's victory last year, the rise of Trump and the vote for Brexit is that those who talk about "electability" know much less about the future than they pretend. As Adam Kotsko said, "electability" is "a purely speculative property." All statements about the future must be heavily discounted. When pundits blather about "electability" they tell us little about Corbyn, but plenty about their own overconfidence, failure to learn from past mistakes and under-appreciation of the importance of complexity and obliquity.