When to Sell a Winning Stock: The Question Without a Good Answer

I still own a few shares of Nvidia that I bought years ago at a price that now looks absurd. The position lives in what I think of as my YOLO account, a small bucket where I put money into individual tech stocks for fun, decided to leave them alone, and got lucky. The Nvidia position 10x'd within a few years. I sold half to lock in gains. That felt prudent at the time. From there, Nvidia went up another 10x or more.

The remaining half is now worth far more than the entire original position would have been at the first 10x. In hindsight, selling half looks less like disciplined risk management and more like accidentally trimming a generational winner.

I do not take this as evidence that I am good at picking stocks. I take it as evidence that luck creates its own problem. Across a handful of old YOLO positions, I now hold roughly $50,000 of unrealized capital gains in old tech names. The percentage returns look enormous. I still have no clear answer to the most important question: when should I sell?

That uncertainty is one of the reasons I love index and systematic investing. Indexing does not eliminate market risk. It delegates the sell decision to a rule, a methodology, or a rebalancing schedule. I do not have to decide whether Nvidia, Apple, Microsoft, or some smaller winner is “done.” The methodology decides for me. That reduction in cognitive load is underrated.

Even a wildly successful stock pick can become a behavioral, tax, and portfolio-management headache.

Trading Requires Being Right Twice

Swing trading and day trading are hard for a structural reason that often gets glossed over: each round trip requires two correct decisions, not one. A good buy can be erased by a poor sell. A great buy can be made worse by a great sell that happens to come too early. Errors compound. If an investor is 51% likely to be right on the buy and 51% likely to be right on the sell, the joint probability of getting both right is only about 26%.

The asymmetry runs deeper than compound probabilities. The buy decision and the sell decision answer different questions. Buying asks whether a security is mispriced today, whether the investor's analysis is better than the market's, whether the expected return clears the cost of capital after spreads and taxes, and whether the position is the best use of the next investment dollar. Selling asks whether the original thesis has changed, whether the position is now too large relative to the portfolio, whether the future expected return is still attractive from today's price, what tax bill the sale would trigger, whether to sell all at once or gradually, whether to pair sales with realized losses, and whether the impulse to sell comes from the plan or from anchoring to cost basis. The sell list is longer, and most items on it did not exist when the buy was made.

The Evidence on Individual Investor Trading

Barber and Odean's 2000 study of 66,465 U.S. discount brokerage households found that active traders earned annual returns of about 11.4%, while the market returned about 17.9% over the sample period. The average household lagged the market and turned over roughly 75% of its portfolio per year. Trading is hazardous to your wealth, as the paper's title puts it bluntly.

A larger follow-up study using complete trading records from the Taiwan Stock Exchange found that individual investor trading produced systematic and economically large losses, with an estimated aggregate performance penalty of about 3.8 percentage points per year for the individual-investor population (Barber, Lee, Liu, and Odean). That penalty represented an annual transfer of about 2.2% of Taiwan's gross domestic product from individual traders to institutions. The evidence on average individual-investor trading is one of the more robust findings in empirical finance.

Selling Is Its Own Skill

The result that should give every active trader pause comes from the institutional side. Akepanidtaworn, Di Mascio, Imas, and Schmidt's 2018 study Selling Fast and Buying Slow analyzed 783 portfolios from experienced institutional managers with average portfolio size near $573 million. The managers showed evidence of skill on the buy side: their buy decisions outperformed a random-buying benchmark by a meaningful margin. Their sell decisions, in contrast, underperformed even random selling. A manager who closed her eyes and sold at random would have outperformed her actual sell decisions.

The authors trace the asymmetry to attention and process. Buy decisions get research, analyst meetings, and explicit deliberation. Sell decisions tend to use heuristics like “the position has run, lock in some gains,” “cut the laggard,” or “rebalance to make room for the next idea.” Those heuristics are fast, but they appear to destroy value relative to the buy-side process. If professionals with deep resources struggle on the sell side, a retail investor putting price alerts on Robinhood is unlikely to do better.

The Bessembinder Paradox

Hendrik Bessembinder's research on the U.S. stock market creates the paradox at the center of any honest discussion of individual stock picking. His 2018 paper Do Stocks Outperform Treasury Bills? found that 58% of U.S. common stocks in the CRSP database since 1926 had lifetime buy-and-hold returns below one-month Treasury bills. The best-performing 4% of listed companies accounted for the entire net gain of the U.S. stock market over the same period. The Summitward guide Most Stocks Lose to T-Bills walks through the full distribution.

Two things follow from that. First, individual stock picking is structurally hard because most stocks underperform. Second, the few stocks that drive the entire market's return are precisely the ones an investor most wants not to clip too early. A rule like “sell after a stock doubles” would have ended early positions in Amazon at $4, Microsoft before the cloud transition, Apple before the iPhone, and Nvidia before AI. The paradox is real: most individual stock positions disappoint, and the rare ones that do not become their own management problem.

Day Trading Is Even Harder

Day trading compresses the right-twice problem into hours or minutes. The SEC's investor education material defines day trading as rapidly buying, selling, and short-selling securities throughout the day to profit from short-term price movements (Investor.gov), and warns that day traders typically suffer severe financial losses in their first months of trading, that day trading commonly uses borrowed money, and that most individual investors lack the wealth, time, or temperament to make money and withstand the very substantial losses day trading can produce (SEC Day Trading: Your Dollars at Risk).

FINRA's 2026 update is worth noting. Effective June 4, 2026, Regulatory Notice 26-10 replaces the old pattern-day-trader framework, including the four-day-trades-in-five-business-days designation and the $25,000 minimum equity rule, with a new intraday margin requirement that applies to all margin accounts (FINRA Regulatory Notice 26-10). The change lowers friction for smaller accounts to day trade. Lower friction does not produce edge. A smaller minimum equity bar may make it easier for people to lose money faster, rather than easier for people to make money.

The Disposition Effect

Selling is hard partly because investors are systematically bad at it in a specific way. Terrance Odean's 1998 paper Are Investors Reluctant to Realize Their Losses? used 10,000 individual brokerage accounts and documented a strong preference for selling winners over losers, even after controlling for rebalancing, transaction costs, and subsequent portfolio performance. Investors realize winners about 50% more often than losers as a fraction of available positions. The behavior is called the disposition effect, and it is one of the most replicated findings in behavioral finance.

The effect is doubly costly. Realizing winners early creates tax bills that could have been deferred. Holding losers too long delays the loss-harvesting benefit. My instinct to “lock in” the first NVDA 10x was probably a textbook disposition-effect trade.

The Tax Trap of Being Right

A lucky winner in a taxable account creates a problem that did not exist when the position was bought: every share has a low cost basis, and selling any of it triggers tax. Federal long-term capital gains are taxed at 0%, 15%, or 20% depending on taxable income (IRS Topic No. 409). High-income households may also owe the 3.8% Net Investment Income Tax on the smaller of net investment income or modified AGI above $200,000 single or $250,000 married filing jointly (IRS NIIT page). State capital-gains rates pile on top, ranging from 0% in Texas, Florida, Washington, and Nevada up to 13.3% in California.

Three tax tools change the math meaningfully.

1. Pair sales with realized losses. Capital losses offset gains dollar-for-dollar; excess losses can offset up to $3,000 of ordinary income per year and carry forward indefinitely (IRS Topic 409). The Summitward guide on tax-loss harvesting covers the wash-sale window and replacement-security mechanics.
2. Sell gradually across tax years. Spreading a large gain across multiple years can keep marginal rates from spiking, avoid the NIIT, and preserve ACA premium subsidies or Medicare IRMAA brackets for the household.
3. Donate appreciated shares. IRS Publication 526 (irs.gov) treats long-term-held appreciated stock as capital-gain property. Donating shares directly to a qualified 501(c)(3) (including a donor-advised fund) lets the donor avoid the capital-gains tax on those shares and, if itemizing, deduct the fair market value subject to AGI limits. For a charitably inclined investor with a concentrated low-basis winner, this is the cheapest way to reduce the position.

The sell decision asks more than “do I still like this stock?” It also asks whether the position is worth its concentration risk, given that diversifying it would cost X in tax today, and that holding it could cost Y in a future drawdown.

Why Systematic Investing Reduces Cognitive Load

One overlooked benefit of index funds and systematic factor strategies is that they delegate the sell decision. A total-market index fund decides what to hold based on capitalization weights and a published rebalancing schedule. A systematic factor fund follows explicit rules for which securities enter and exit the portfolio. The investor does not have to evaluate whether Nvidia is “done”; the methodology handles inclusion and weighting automatically. The Summitward guide on systematic investing and the companion on what “passive” actually means cover the framework in depth.

That delegation has a behavioral payoff. Every minute spent wondering whether to sell a position is a minute not spent on higher-leverage decisions like savings rate, tax-advantaged account location, insurance gaps, or career investments. The investor who has put the equity sleeve on autopilot can think about whether to convert to a Roth this year, whether to increase the bond ladder, whether the umbrella policy is large enough. Those decisions move retirement timing more than any single-stock trade.

A Decision Framework for Old Winners

For an existing low-basis position in an old tech winner, a useful question to ask is:

If I had the after-tax value of this position in cash today, would I buy this same amount of this same stock as part of my financial plan?

For most investors holding old YOLO winners, the answer is no. The position exists because of historical luck, not because of a current allocation decision. That asymmetry is what makes the sell decision feel paralyzing: the cost basis is low, the tax is real, and there is no clear signal that says “sell now.”

A workable 7-step framework:

1. Set a maximum single-stock cap. A common rule is no individual security above 5% or 10% of liquid net worth without explicit documentation of why. The cap is a standing decision, made before any specific position triggers it.
2. Estimate the tax cost of diversifying now. Federal LTCG plus NIIT plus state, applied to the unrealized gain. The calculator below does the arithmetic for an old low-basis position.
3. Sell gradually if bracket effects matter. For an investor near a bracket boundary or the NIIT threshold, splitting the sale across two or three tax years can preserve thousands of dollars.
4. Pair sales with realized losses where available. A tax-loss-harvested position elsewhere in the portfolio offsets the gain dollar-for-dollar. See the TLH guide for wash-sale rules.
5. Donate appreciated shares if charitably inclined. A donor-advised fund accepts appreciated stock, sells it tax-free inside the fund, and lets the donor grant cash to charities over time.
6. Sell high-basis lots first. Specific-share identification at the broker reduces the immediate tax bill by selling the lots with the highest cost basis first, leaving the lowest-basis (highest-gain) lots for donation or step-up at death.
7. Do not let the tax tail wag the portfolio dog. A 15% or 23.8% capital-gains tax is real, but a 50% to 80% single-stock drawdown is also real. Concentration in a single name, even a great one, carries idiosyncratic risk that diversification removes for free. The Summitward guide on concentration risk works through the math of when paying the tax to diversify beats holding.

Try It: The Tax-Cost-of-Selling Calculator

The calculator runs the deterministic tax math for an old low-basis position. Enter the current value, the cost basis, the federal bracket, whether NIIT applies, the state rate, the total portfolio value, and a target single-stock cap. Toggle the donation option to see the tax saved by giving appreciated shares to a qualified charity rather than selling and donating cash. Default values are calibrated to the YOLO scenario in the opening anecdote (around $50k position, low basis, 15% federal bracket, 5% state).

When Individual Stock Picking Is Fine

Individual stock positions are not forbidden by the evidence; they are just rarely a wealth-building strategy. A reasonable structure is a clear separation between a core portfolio and a bounded satellite account. The core holds the diversified index, factor, or target-date funds that drive the financial plan. The satellite holds the YOLO positions, entertainment trades, or high-conviction individual ideas, treated as a hobby.

The structural rule that makes this work:

If the account is large enough that a 100% loss would meaningfully delay retirement, change college funding, or alter housing plans, it is not a YOLO account anymore. It is part of the financial plan and deserves financial-plan-level discipline.

Inside that structure, individual stock picking can be fine for:

• Investors trading small, explicitly bounded amounts as entertainment or education.
• Operators with a documented, repeatable edge and a written process that includes position limits, tax awareness, and honest benchmarking against a passive comparator.
• Concentrated single-stock holders dealing with restricted stock, founder equity, or RSU overhangs. The Summitward guide on selling RSUs at vest handles the default decision for new grants.
• Investors using stock picking as a way to learn how markets, companies, and accounting actually work, with a small enough stake that the tuition is affordable.

It is rarely fine for:

• High earners trading retirement money, tuition money, or a house down-payment fund.
• Anyone using margin, leveraged options, or concentrated positions without explicit understanding of the downside.
• Investors who do not track their after-tax, risk-adjusted performance against a passive benchmark.
• People who would describe their trading as a way to escape their job rather than a hobby they can afford to do badly.

Key Takeaways

• Trading needs two correct decisions. Buying and selling answer different questions. Errors compound, and the sell side is structurally harder.
• Selling is empirically harder than buying. Selling Fast and Buying Slow found that institutional managers with deep resources showed buy-side skill but underperformed even random selling.
• The Bessembinder paradox is real. Most stocks disappoint, and the rare ones that do not become their own management problem. Clipping a 10x winner too early can be the most expensive trade of a lifetime.
• Lucky winners create new problems. Concentration, tax lock-in, anchoring, and opportunity cost all show up after the win, not before.
• Systematic investing offloads the sell decision. An index or rules-based portfolio delegates inclusion, weighting, and rebalancing to a published methodology, which frees up cognitive bandwidth for higher-leverage planning decisions.
• For old winners, run the tax math first. Sell gradually, pair with losses, and donate appreciated shares where possible. Do not let the tax tail justify unlimited concentration.

Frequently Asked Questions

Isn't a 10x or 100x winner proof that I have stock-picking skill?

Sometimes, but skill is hard to distinguish from luck on a sample size of one or two positions. Bessembinder's data shows that a few stocks drive most of the market's return, so a randomly chosen portfolio of a few names will occasionally contain a massive winner by chance. The honest test is whether the rest of the portfolio also outperformed a passive benchmark on a risk-adjusted, after-tax basis over a long period. For most retail investors, that test fails even when the YOLO account contains a big winner.

Why not just hold forever and step up the basis at death?

Step-up in basis at death is a real benefit of holding low-basis positions. The trade-off is concentration risk. A position that is fine today at 10% of the portfolio could be 40% in five years if it continues to outperform, or 3% if it underperforms, either of which changes the financial plan's risk profile. A “hold forever” rule on a concentrated position is a bet that the investor will not need the money, the stock will not collapse, and the estate-tax rules will not change. That is a lot of bets riding on inaction.

What about a stop-loss order to protect the gain?

Stop-loss orders on long-term index or single-stock positions come with their own set of failure modes: execution risk, whipsaw, re-entry risk, and tax acceleration. The Summitward guide on stop-loss orders for long-term investors walks through why they usually do not solve the problem they appear to solve.

Does the FINRA pattern-day-trader rule change in June 2026 make day trading more attractive?

It makes day trading more accessible, not more profitable. Regulatory Notice 26-10 replaces the four-day-trades-in-five designation and the $25,000 minimum equity requirement with a broader intraday margin rule. Lower friction does not create edge. The Barber-Odean evidence on individual trader underperformance was collected under tighter regulatory rules than the new framework; a looser regime is unlikely to improve the underlying results.

How do I value an old position when the basis is unclear?

Brokers are required to track cost basis for securities acquired in 2011 or later (the relevant regulation is in IRC Section 6045(g)). For older positions, transaction confirmations, old 1099-B forms, or the broker's “average cost basis” can serve as an estimate. When the records are truly lost, the IRS expects the taxpayer to use a reasonable method (often $0 basis, which produces the worst-case tax bill). A CPA can help reconstruct basis if the position is large.

Should I use a robo-advisor or direct-indexing service to diversify out of a concentrated position?

Direct indexing pairs equity-market exposure with active loss-harvesting at the individual-security level, which can generate losses to offset the gains from selling a concentrated position. The Summitward guide on whether you need direct indexing covers when the fee and complexity are worth it. For most retail investors with $50k-$200k of unrealized gains, simple TLH inside a low-cost ETF portfolio captures most of the benefit at a fraction of the cost.

Related Guides

• Most Stocks Lose to T-Bills covers Bessembinder's skewness evidence in depth: why most individual stocks disappoint, and why diversification is the only reliable way to own the winners.
• Concentration Risk covers the portfolio-risk math (HHI, break-even hurdle, fan charts) for a single position that has become outsized.
• Sell Your RSUs at Vest handles the default decision for new equity grants, which prevents the YOLO-winner problem from forming in the first place.
• The Engineer's Guide to Systematic Investing covers the “delegate the decision to a rule” framework that index and factor funds operationalize.
• Tax-Loss Harvesting covers the wash-sale window and replacement-security mechanics for pairing realized losses with realized gains.
• Should Long-Term Index Investors Use Stop-Loss Orders? covers the related question of using exit triggers on long-term positions, and why they usually backfire.

Sources

1. Barber, B. M., and Odean, T. (2000). “Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors.” Journal of Finance, 55(2), 773-806. SSRN.
2. Barber, B. M., Lee, Y.-T., Liu, Y.-J., and Odean, T. (2009). “Just How Much Do Individual Investors Lose by Trading?” Review of Financial Studies, 22(2), 609-632. SSRN.
3. Akepanidtaworn, K., Di Mascio, R., Imas, A., and Schmidt, L. (2018). “Selling Fast and Buying Slow: Heuristics and Trading Performance of Institutional Investors.” Working paper. SSRN.
4. Bessembinder, H. (2018). “Do Stocks Outperform Treasury Bills?” Journal of Financial Economics, 129(3), 440-457. SSRN.
5. Odean, T. (1998). “Are Investors Reluctant to Realize Their Losses?” Journal of Finance, 53(5), 1775-1798. SSRN.
6. U.S. Securities and Exchange Commission. “Day Trading: Your Dollars at Risk.” Investor publications. sec.gov.
7. U.S. Securities and Exchange Commission, Office of Investor Education and Advocacy. “Day Trading” (glossary). investor.gov.
8. FINRA. “Regulatory Notice 26-10: Intraday Margin Requirements.” (Effective June 4, 2026.) finra.org.
9. Internal Revenue Service. “Topic No. 409, Capital Gains and Losses.” irs.gov.
10. Internal Revenue Service. “Net Investment Income Tax.” irs.gov.
11. Internal Revenue Service. “Publication 526, Charitable Contributions.” irs.gov.
12. FINRA. “Asset Allocation and Diversification.” finra.org.