The longest holding period on this site also has the most lopsided returns
Warren Buffett holds a disclosed position for a median of 3.125 years and his returns carry a skew of 4.78, the most lopsided of the three investors with a full per-trade record. Michael Burry holds for a median of 3 months and his skew is 0.47, close to symmetric. Li Lu sits between both on each figure.
Warren Buffett's median disclosed position is held for 3.125 years before it closes. Michael Burry's median position is held for 0.25 years, three months, a twelfth of Buffett's. Buffett's return skew, a measure of how lopsided a set of returns is, sits at 4.78. Burry's sits at 0.47, close to a symmetric distribution. Neither figure, median holding period or return skew, has appeared in a Tenachine guide before this one, and next to each other they describe two entirely different shapes of a trading record.
Skew above zero means a distribution has a long tail stretching toward large positive values, a few outsized results pulling the shape away from symmetric. Buffett's 4.78 is consistent with his Mastercard position, a single trade returning 1125.51 percent that this site has already covered as the largest raw gain among the three investors with a full disclosed record. A skew close to zero, Burry's 0.47, means no single trade is doing that kind of work. His results cluster closer together, without one runaway winner reshaping the whole distribution.
Why a longer hold tends to fatten the tail
A position held for three years has three years for a real winner to keep compounding past what a three-month hold could ever reach, and three years for a real loser to keep losing too. The asymmetry shows up because gains and losses are not symmetric over time: a position can only lose 100 percent of what was put in, but it can gain many times that, and the longer a winning position is allowed to run, the more room it has to reach those larger multiples. A short median hold caps how far any single position, winner or loser, can travel before it closes, which is one plausible reason Burry's distribution stays closer to symmetric while Buffett's stretches out.
Li Lu sits in the middle on both, and three points is not a pattern
Li Lu's median hold, 2.625 years, sits between Burry's and Buffett's, and so does his skew, 1.22. That ordering matching twice, across two unrelated investors' worth of data on top of Li Lu's, reads like a real relationship. It is also built from exactly three data points, one investor per position on the scale, which is nowhere near enough to call this a rule about how holding period relates to skew in general. Tenachine's guide on Li Lu's own Kelly criterion output already flags his 14-trade sample as thin enough to move a great deal with one more disclosed position. The same caution applies to reading three investors' worth of figures as a trend rather than three specific, individually interesting numbers that happen to line up this way in this window.
A 13F filing discloses long US equity positions as of one date, not a full trading history, so a position opened and closed between two quarterly filings never shows up in this data, and neither median holding period nor return skew can see it. Longer holding periods are not automatically better or worse than shorter ones, and a high skew is not a target to aim for, it is a description of what already happened to one investor's disclosed trades over one stretch of time. Past performance does not predict future results, for any of the three investors covered here.