The Gamblers Fallacy in Stocks: Buy when Stocks go Up, Sell when Stocks go Down
Despite being aware of the “Gamblers Fallacy” I couldn’t really believe that it would hold true for stocks as well. It made sense to me that repeated events like rolling dice or flipping coins that are not subject to any or neglectable human influence should have the same outcome probabilities for each repetition, regardless of any prior roll or flip outcomes. There would remain a fifty percent chance for tails in my next flip, even if I had already flipped five tails in a row up to this point. However stocks are not completely independent of their prior developments, or at least that was what I thought. I assumed that a significant enough proportion of investors would still be more likely to invest in rising stocks, than in falling ones, not aware of the fallacy or being too attached to their falling stocks. With more and more platforms like Robinhood that have made trading accessible to a broader public, I assumed that there would be enough “trend-followers” to create such dynamics in the market. However, I was wrong…
To analyze if the outcome (Stock goes Up vs. Stock goes Down) is independent of the prior development, I devised a simple strategy idea that would capitalize on such developments. Whenever stocks had gone up two time intervals in a row, I would Buy and whenever they went down for just one time interval I would Sell. To simulate this trading strategy as easily as possible I used trading days as a time interval and the closing value of each trading day to evaluate if I would have profited or lost with my strategy.
The Strategy: I would Buy if the stock went up for two days in a row, I would Sell once the stock went down for one day.
When I simulated this strategy for the first few companies on the data from 2017–2020 I realized that just a little bit more than 50% of my total trades would have turned out profitable. This seemed already in itself not really promising for my underlying assumptions. Once I added a few more companies that had also decreased in total value over those years the estimates for my strategy went even below 50%, indicating that I would have lost more times than profited with the intended strategy if I had invested in those stocks. Across all sampled stocks I would have turned a moderate profit, however so had the whole market and just investing in an index and not looking at it again for three years would have probably been less effort and more profitable due to lower fees. Once I added the value increase (in %) over those three years as a controlling variable it became even clearer that the return on my strategy was highly dependent upon the general development of those stocks and by no means a magic bullet to increase returns (r = >.8).
Even though I didn’t find a profitable strategy, I gained new insights that I wouldn’t have assumed to be true if I hadn’t seen it for myself (or in existing research). If the stock goes Up or Down seems to be mostly independent of its prior development, it really does not make sense to base investment decisions solely upon prior trends. After all the Gambler’s Fallacy seems to hold true even in the face of the many investment decisions made by different individuals.
I’m anything but an expert in finance, I am just a curious mind who loves to learn, experiment, and investigate. Please feel invited to let me know if valuable additional information is available or if reasoning errors have been made. Thank you for letting me share this with you!
