Never tell me the Odds: Is your CEO Han Solo?
Probabilistic strategy, tactics, decisions require a mindset shift: stop hiding from the odds; start moving them.
Does heroism require the denial of reality? Can you inspire, while being honest about the low odds of success? Can you succeed, when the odds are against you?
‘The Empire Strikes Back’. Han Solo piloting the Millennium Falcon. Under attack and outgunned by the Empire’s TIE-fighters, Han is warned by Princess Leia of a hazard ahead:
‘Asteroids’ she shouts, alarmed.
To the horror of his companions Han Solo spots an opportunity to lose those attacking him…
But Sir, C-3PO opines, the possibility of successfully navigating an Asteroid field is approximately 3720-to-1.
Han Solo, eyes fixed determinedly ahead, growls Never tell me the odds.
Why? One view, is that knowing the odds would undermine his confidence making it less likely he’d succeed. It would be self-fulfilling.
One CEO recently worried that deploying Cassi, and seeing the odds of success, might “crystalise the doubts of his management team”, reducing his company’s chance of success.[i] It is not the first time we’ve heard the Han Solo objection, but it was put with impressive clarity and inspired us to explore the issue for a wider audience.
Defining the “Han Solo” Objection
In high-commitment ventures from start-ups to sports-fields, statecraft to Star Wars ‘never tell me the odds’ is an attractive one-liner, and superficially, seems to capture something important.
In analytical form the claim made in the the Han Solo objection can be described as:
The probability of success, viewed collectively - from my team, my company (etc), may be low.
If I confront that probability, my confidence, and/or my team’s, will fall.
Therefore I am better off not knowing the odds, because ignorance preserves the confidence required to achieve my desired outcome.
The romance is obvious - the triumph of human courage over maths, people overcoming the odds - and that has its own appeal. But the logic does not hold. The quote misreads the scene, and the objection does not stand scrutiny. Three failures follow.
First, information-avoidance can sometimes temporarily preserve motivation, but unless you are about to fly into an asteroid field - and perhaps, as we will see Han shows - even then - it is a leadership failure. Han Solo doesn’t have time to fully convince his comrades their odds are wrong, but you do.
Second, confidence bolsters a sound plan; it does not substitute for one. The argument is clearest at the extreme: even if you are galvanising your people towards a heroic sacrificial defeat that achieves nothing - you are still convincing them that their least-bad outcome at that point is to die gloriously and be remembered well. You are not deceiving them by hiding the odds.
Third and finally, sometimes the Han Solo CEO is right, and the crew are wrong. He has access to information they don’t, and should act on it, convincing them as they go.
1. Information Avoidance is Leadership Failure
A leader’s job in motivating others is surely to convince their people that their plan has a sufficient chance of succeeding to be worth pursuing. They should usually be able to explain the strategy, tactic or decision sufficiently well to convince their team it has sufficient odds of succeeding to be worth it.
Han Solo communicates his calculation of the odds through his confidence - the Asteroid field is a better bet than continuing to fight out-numbered and out-gunned. He says as much:
Leia: What are you doing? You’re not actually going into an Asteroid field?
Han: ‘they’d have to be crazy to follow me in there’.
Sure, the odds of survival aren’t great in the Asteroid field, but Han Solo knows the base rate C-3PO quotes “3720 to 1” doesn’t account for Han’s skill and experience, nor the worse odds of fighting in open Space - this is the best decision with highest expected value under the circumstances. His companions are alarmed, they check his reasoning ‘if you are doing this to impress me…’ but they don’t dissent.
Your job as CEO isn’t to suppress the odds, but to convince people your calculation of them is right, with reasons.
Furthermore, you must remember that odds are not a constant; they change as the facts change, they respond to action. Each step that removes a risk, reveals information or compounds an edge, changes the number.
In a race for survival through an asteroid field, every dodge of a space rock that takes Han and friends closer to the far side shortens the exposure window; conditional on survival, the forecast can legitimately improve. A 20% forecast of survival on entering, and an 80% forecast later are consistent. Ideally, every decision would be taken knowing which most increased the odds of survival.
The point generalises: when we founded Cassi, the base rate - 80-90% of start-ups fail - was C-3PO sounding the alarm. Our odds of success started from 10-20%. Accounting for what we know of ourselves, what we wanted to do, and the need for it, we raised those, with our median Founder forecasts of the company surviving for 5-years hovering around the 50% mark (vs more ambitious goals, like achieving £15M revenue in FY26, where odds were lower). As time has gone on - winning customers, gaining traction, deploying the platform, those odds have consistently risen.
2. Confidence is no Substitute for a Sound Plan
In most fields, where leaders have a different assessment to their team they can communicate it with more than just their own confidence. They have time. They can explain with reasons, evidence, their own judgement of the odds.
If this doesn’t work, the CEO might change their mind (or not - see next section). After all, those people around the Board table are there because their expertise and judgement in particular domains is judged to be of sufficient calibre and calibration to make their advice valuable to the CEO, who can’t know everything. This is particularly true where they forecast that some particular factor in their domain is highly influential on the strategic outcome.
Moreover, since we know that most human forecasts are no more accurate than chance – we’re usually better off deferring to the median crowd forecast, doings so over trusting our own gut, or relying on a single forecast from another - even if they are an expert in the area.
But you would still want to understand that expert view. For example you might weight a forecast due to expertise where:
the HR Director and the Board are predicting that the probability of their being able to hire thirty world-leading ML Researchers by end FY was 10%; and
hiring thirty world-leading ML researchers is the difference between a 60% probability of the company achieving its strategic objective; and
failing in recruitment reduces the chance of achieving that strategic objective to below 50%…
…so strategic failure is then shown to be more likely than not. At the very least the CEO should be interested in:
(a) why the probability of hiring the talent needed is so low,
(b) why the collective view is that this is so influential on the desired strategic objective and
(c) what might be done to alter that forecast - how could resources be differently assigned, or what actions taken, that would optimise for success?
The result of this should be either the CEO updates their forecasts, the HR Director and Board update theirs, or the gap remains. In either case - each should try to persuade the other of their judgement, to figure out what would change the view of one or other or both, and focus effort and resources on those factors that most move their odds. Hiding the odds is no substitute for optimising them, and likely to have the opposite effect than the one intended.
3. Sometimes the Han Solo CEO is Right
The CEO might see the odds from the collective intelligence of their team - the Board, the company - and disagree. If this is the case, they must have the courage to maintain their beliefs independently. The disagreement of others is not per se evidence to change their own beliefs.
In fact, a CEO or equivalent should, at key moments, *expect* to disagree with the crowd. Whether we think of that in terms of principled forecasting or informal disagreement around the table or during an online meeting, we should expect that the CEO will regularly disagree with the crowd, and be correct to do so.
If this was not the case, then organisations would not need CEOs, and decisions and courses of action would tend to be obvious.
Choosing the moments for this disagreement is in itself a key component of CEO prediction. Every major decision contains two forecasts: one about the external environment, and one about the organisation’s ability to shape that environment. The CEO’s must judge when their internal forecast justifies overriding the external and internal consensus – and the best way to do this, is to have a clear knowledge of their forecasting performance – do they consistently outperform others? Is there a reason this time to believe they might be right when others think they are wrong?
A slightly imperfect analogy is in investment. It is recognised that most people should just invest in tracker funds. Tracker funds are essentially the crowd wisdom of the investment community. But a CEO should be the equivalent of a trader who regularly beats a tracker forward by taking actions of whatever sort, equivalent to investments, that beat the tracker.
The analogy is imperfect only in mechanics. In both cases, someone is making probabilistic bets against the collective forecast. In finance, the ‘bet’ is a literal trade; in strategy, it’s an organisational action. A tracker fund is the best collective intelligence forecast; fund managers who beat that are super-forecasters beating the crowd average. CEOs, like super-forecasters, should also know when to go with the collective forecast, and when to not forecast at all (because they have no relevant insight or information, and they would just be praying for luck).
CEOs are selected to be capable and well-informed individuals, operating in areas they have much relevant experience. They have the ability to make forecasts, and to shape the world to meet those forecasts in certain areas. When a CEO chooses a path, they are predicting both the external and internal environment and their own ability to shape it.
They can and should beat the crowd: that is their arbitrage against the competition, and the rest of the world.
We must be clear here: we are not saying “trust the hero; to hell with the odds”. In Star Wars, we can see enough in the scene to suggest Han Solo was taking a calibrated risk, broadly aware of the baseline. But his throwaway line “Don’t tell me the odds” is dangerous. Just taking a wild punt, and getting lucky is likely to lead to catastrophe in time. A CEO should employ calibrated contrarianism – and that calibration means they *must* know the odds – they need to know how often they are right, and how right they are (e.g. do things they forecast at 80% confidence happen 80% of the time). They need to know the odds their team is putting on things, they should seek out the best possible odds. The contrarianism should be legible – Peter Theil’s question makes the point:
“What important truth do very few people agree with you on?”
Cassi would make you go further: why do you believe in this truth? What would change your mind? How likely do you think it is you will be proved right?
Over time, Cassi would help you improve your judgement, and enable you to learn how much you should trust your ‘instincts’ over the crowd. For the vast majority of humans – including CEOs, the median of the crowd forecast – or the median of a group of superforecasters - will give the best odds. At Cassi, we are showing that AI forecasting can often match or exceed the crowd – the CEO would be wise to test their contrarianism and refine it. Wild bets are not leadership, “don’t tell me the odds” is sexy but negligent.
Governance of the CEO – the Board, the market - will find the collective nature of the crowd forecast very helpful in calibrating the perceived riskiness of the CEO’s strategy, checking that heroic has not devolved into delusional, and, not least, the degree of reward the CEO should get for beating the collective baseline. Using an internal forecasting platform or prediction market allows the direct comparison of the CEO with crowd consensus over time.
We think organisations will go further - hiring and promoting partly on a proper scoring system that identifies:
who is the most effective in identifying the most important factors in achieving success;
who is the most accurate and well calibrated in predicting the odds of achieving them;
A good CEO might be such a person, consistently outperforming the ‘crowd’ that constitutes their rivals for the job.
Doing that takes judgement:
· knowing the odds of your achieving your desired outcome,
· knowing which factors will most influence your probability of achieving your outcome
· understanding where your odds, and the factors that you think matter, differ from those of your team, the unbiased external view, the analysis of AI and its forecasts.
· Knowing when and how much (if) to trust your judgement over that of your team, your crowd, your AI.
Your CEO should copy the wider lesson in the Han Solo scene, and ignore the misleading one-liner: you must know the odds. With Cassi, you know both the number, and what you most need to do to succeed. You can optimise your odds. At that point, as Han later said “You’re all clear, kid. Now let’s blow this thing and go home!”
Failing that, head over to our website, www.cassi-ai.com - currently in the early stages of a makeover - and learn a little more about what we do.
[i] Conversely, at a recent Conference, “Katie” made the argument that surfacing the odds, and the factors that most influence them, might consistently improve the odds. This is not just Cassi’s argument for more efficient management, but rather one suggesting that the concentrated shared consciousness such a method creates allows people to collective visualise an outcome – and perhaps this ‘mind share’ will move the odds – ‘manifesting’ the outcome as it were - beyond just what would be expected from more efficient resource allocation. We won’t explore this further – but I did say I would mention it on the blog as a counter-point to the view that surfacing the odds and factors would crystallise doubts!