How you can use AI to know when to sell

Recently, the Minotaur team has been thinking a lot about confirmation bias. We humans, despite all our strengths, have an underlying tendency to focus on and put our faith in evidence that fits with our existing beliefs. In investing, this is particularly pertinent, as we tend to seek out information that supports our thesis on the market or individual stocks, rather than contemplating why we may be wrong.

So, what if we could use our AI superpower, Taurient, to not only hold us accountable but keep confirmation bias at bay, helping us to move quickly when our thesis no longer has legs? As Mark Twain penned, 'It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so.'

With this in mind, we developed an AI-generated "Thesis Validation/Invalidation Report", which enables us to quickly understand the aspects of an earnings report that either support or challenge our thesis on any given company. Taurient then provides a recommendation of what it thinks we should do (i.e. close position, add to position, etc), which we can choose to ignore or take on board. 

For example, we shorted Zillow (NYSE: Z) in August 2024, America's answer to REA Group. We believed that changes in real estate brokerage commissions in the US would put downward pressure on buyers agent commissions, negatively impacting Zillow's business (which relies heavily on revenue from buyers agents' purchasing leads). 

We were wrong.

Taurient, unlike an analyst, isn't afraid of hurting our feelings, outlining that Zillow's update in November strongly challenged our thesis. The recommendation? That we "abandon our short on Zillow"... Which we did, and was ultimately the right decision. 

We have now incorporated this into our research process for every stock in the portfolio - helping us to maintain conviction in our positions, but also keep us humble enough to quickly move when we get it wrong.

More recently, we sold out of our position in MongoDB (NASDAQ: MDB) on its result using this report. The stock's share price has plummeted 20% since then. We also trimmed our position in Prysmian (MTA: PRY) after Taurient predicted it was likely to downgrade its long-term earnings guidance at its capital markets day at the end of March... which it did.

Understanding that confirmation bias can sway investors' decision-making has also aided in our ability to pick shorts. Analysts, we've come to realise, naturally seek information confirming their existing thesis – a tendency amplified in institutional settings where admitting error could carry career risks. This reluctance to accept negative signals creates opportunity.

We call this "Bad News Drift" - when investors are slow to react to negative news, initially waiting a few days or weeks to see if anything changes before making a decision. Sure, there may be an initial sell-off on negative news, but often it takes several trading days or even weeks to see its full impact.

With "Bad News Drift" in mind, we've shorted WiseTech Global (ASX: WTC), Tesla (NASDAQ: TSLA) and Bumble (NYSE: BMBL), all of which have been strong contributors to the Fund. 

In the case of Australia's most expensive tech company, the secret private life of WiseTech CEO and founder, Richard White, caused some initial pain in the company's share price. However, it was ultimately director resignations and an earnings downgrade that pushed the company off a cliff, with WiseTech's share price plummeting ~30% since mid-February. We have since closed out this short.

Meanwhile, if Elon Musk slinging a Nazi salute and a chainsaw on stage weren't a death knell for Tesla, its sales plummeting while Chinese producers go from strength to strength certainly has been. We've been shorting the stock since early February, and while it's been a volatile ride, its share price is down 23% since then.

Bumble, while certainly less dramatic, faces its own death spiral, and its turnaround efforts are unlikely to reverse fundamental declines in user growth, monetisation, and market share. We see no reason why this company will continue to exist in five years, particularly given competition from its rivals Tinder and Hinge. The stock was down 22% before we shorted it, but has fallen another 33% since.

Outlook