When to Sell a Stock: Good Reasons, Bad Reasons and the Exit You Set First
The good reasons to sell, the bad ones, O'Neil's two sell rules from the source, why people sell winners and keep losers, the US tax lines, and how traders set the exit before they buy.
Most answers to this are written for long-term investors. Traders sell for different reasons, and both make the same mistake once the position is open. This covers both lists, plus the exit you write down before you buy.
Quick Answer
When to sell a stock comes down to four cases: the reason you bought it is gone, you need the money, the position has grown too big for your portfolio, or price hit the exit you set before buying. A drop alone or a rise alone isn't a reason. Deciding with the position open is the hard part, so decide at entry: where you're wrong, where you take profit, and what happens in between.
What Are the Good Reasons to Sell a Stock?
Nearly every investing guide leads with the same one: the reason you bought no longer holds. The full list is short, and each item has a catch.
| Reason | What you'd see | The catch |
|---|---|---|
| The reason you bought is gone | Slowing sales or earnings, shrinking margins, rising debt, a leadership change, a stronger rival, insider selling. Or you got the business wrong | One bad quarter isn't a broken thesis. Is the problem temporary or permanent? |
| Valuation ran past the business | A P/E far above its own history, its peers or the market | A weak timing tool. Expensive stocks can stay expensive for years |
| You need the money, or your plans changed | A house, tuition, a shorter horizon, less appetite for risk | None. The cleanest reason here |
| One position got too big | A winner became an outsized slice of the portfolio | Trim it back. You don't have to sell it all |
| A better use for the cash | A new idea with a clearly better outlook | High bar: tax on the gain, and you can be wrong twice |
| Harvesting a tax loss | A loser in a taxable account | Rebuy within 30 days before or after and the wash-sale rule disallows the loss for now. Taxes alone shouldn't decide it |
| Price hit the exit you set at entry | Your stop or target printed | The only one a price chart can show |
On size, Fidelity's guide warns that for investors, "having 5% or more of their portfolio in a single stock can increase their risk." That's Fidelity's line, not a law. The first six rows are mostly investor sells, and none of them shows up on a price chart. One test covers most of them: would you buy this stock today, at this price?
What Are Bad Reasons to Sell?
A price drop alone, a price rise alone, emotion, headlines and trying to time the top all feel like reasons to sell. On their own, they aren't.
- It dropped. A red number tells you the price moved. It doesn't tell you why. Ask whether the problem is temporary or permanent. Buying more to lower your average is its own trap, as the guide on averaging down a losing trade explains.
- It went up a lot. Winners can keep running, and selling them early is the costly half of the mistake in the next section.
- You're scared, or greedy. Fear, FOMO and regret pick the wrong click. Let a rule you wrote while calm decide. The trading psychology guide covers the biases.
- A headline, forecast or downgrade said so. If the news changes the business, that's a good reason. An opinion about the price, or what everyone else is selling today, isn't.
- You want to sell the top. Nobody sells at the high. William O'Neil wrote in a 1999 AAII Journal article: "The best time to sell and take your profit is when your stock is advancing and still looks strong and positive to most investors."
Why Is Selling So Hard? The Disposition Effect
Selling is hard because of the disposition effect, the habit of selling winners too early and riding losers too long. Hersh Shefrin and Meir Statman named it in 1985. Terrance Odean measured it in a 1998 Journal of Finance paper using 10,000 accounts at a large discount broker, 1987 to 1993.
Odean's numbers: the proportion of gains investors realized was 0.148, against 0.098 for losses, so a stock that was up was more than 50 percent more likely to be sold on a given day than one that was down. It cost them. The winners they sold beat the losers they kept by 3.4 percent over the next year, in excess return against the market.
Loss aversion is part of it. In Kahneman and Tversky's 1979 prospect theory, the value function is "generally steeper for losses than for gains," and selling a loser makes the loss real. The fix is boring: decide the sell before you buy and write it next to the entry. A trading plan template has a slot for it.
When to Sell a Stock for Profit
The best-known profit rule comes from William O'Neil, who founded Investor's Business Daily. In IBD's current wording: "take at least some profits when a stock gains 20% to 25% from the buy point." He crafted it in 1961 after noticing that stocks breaking out of well-formed bases tended to run up 20% to 25% before correcting. It's a rule of thumb. I couldn't find a test showing it beats holding.
Fair dissent: a target alone can cut a winner short. Sell it all at +25% on a stock that later triples and you kept a sliver of the move. Targets set from the chart work differently from analyst targets, and most traders sell only part at the first one. Trim, sell half, top-slice, scale out: same idea. The guide to taking partial profits at T1 and T2 covers the split.
One more from O'Neil's 1999 sell rules, a pure price cue: after months of advance, a stock "runs up faster than normal in just one or two weeks (usually 25% to 50%)." That's when people chase. It's a better time to take some off.
When to Sell a Stock at a Loss
O'Neil's loss rule, from the same AAII Journal piece: "If a stock goes 7% or 8% below your purchase price, sell it." He called it "your small insurance policy against potentially devastating losses." It's 7% to 8%, not 10%, and it's measured from what you paid: "Once you have a profit in the stock, other rules come into play." So it isn't a trailing stop. The CANSLIM breakdown covers the rest of his system.
Why cut small and early? Arithmetic.
| Loss | $10,000 becomes | Gain needed to get back to $10,000 |
|---|---|---|
| 7% | $9,300 | 7.5% |
| 8% | $9,200 | 8.7% |
| 10% | $9,000 | 11.1% |
| 20% | $8,000 | 25% |
| 33.3% (one third) | $6,667 | 50% |
| 50% | $5,000 | 100% |
| 75% | $2,500 | 300% |
A stop-loss order makes the rule mechanical, with caveats. Per Investor.gov, a triggered stop becomes a market order, and "the stop price is not the guaranteed execution price." A short intraday dip can set it off, and a stop-limit may not fill at all. A trailing stop follows price up by a set percentage or dollar amount and stays fixed when price falls. The primer on how stop orders work compares the types, and the trailing stop walkthrough covers picking the distance.
How Do Traders Decide When to Sell Before They Buy?
Investors sell when something off the chart changes. Traders sell on levels picked before the entry. Same principle, applied to price.
| Question | Investor | Trader |
|---|---|---|
| What ends the position | The business changed, or your life did | Price broke the level the idea needed |
| Where the stop goes | Often nowhere, or a percent like O'Neil's 7% to 8% | Beyond the swing low or support that breaks the setup |
| Taking profit | Valuation, position size, a better use for the cash | Some at T1, the next resistance |
| Letting a winner run | Hold while the thesis holds | Trail the stop under each new higher low |
| A near-vertical run | Mostly ignored | Take some off. Don't add |
For a long, the stop goes beyond the level that breaks the setup, usually the last swing low. T1 is the next resistance, so finding support and resistance is most of the job. Once it pays, you can move the stop to breakeven or trail it. A near-vertical run far from support is an overextended chart: take some off, don't add.
Plenty of investing guides say chart sell signals don't work consistently. As forecasts, fair. A stop below the swing low forecasts nothing. It's the price where you agreed in advance that the idea failed, and the recovery table is why you want it close. Four questions before you buy:
- Where's the level that breaks the idea? Your stop goes beyond it.
- Where's the next opposing level? That's T1, your first take-profit.
- Does T1 pay at least twice what you're risking? Below 2:1, you need to win more than a third of your trades just to break even.
- Is price already stretched far above support? Then wait for a pullback to a level, or pass.
Set the exit before the entry, on the chart you're about to trade.
Upload the screenshot and SnapPChart grades the setup from A+ to F, then says whether to enter now, wait for a pullback to a named level, or pass. It grades a setup before you're in it. It doesn't watch a position or tell you when to sell one you already hold. Your first grade is free: a summarized analysis of your chart and a graded second opinion on whether to take it, wait or pass. Entry, stop, targets and the trade plan come with a paid plan.
Grade this setupOn a paid plan, the grade comes with the exit planned before the entry: a stop beyond the nearest structural level with a line on why it sits there, T1 at the next opposing level, T2 when the chart shows one, and trailing-stop guidance when it applies. If T1 doesn't pay at least twice the risk (2.5 times on forex), the plan is thrown out and the grade comes back as a pass. It reads only the chart (the AI trading guide lists what a chart can't show), so the investor sells above stay yours.
Do Taxes Change When You Should Sell?
In the US they change what a sale costs. From IRS Topic 409 (updated September 24, 2026):
- Held more than one year, the gain or loss is long-term. One year or less is short-term.
- Net short-term gains are taxed as ordinary income.
- For 2025, the rate on most net capital gain is no higher than 15% for most individuals.
- Losses offset gains. Excess losses deduct up to $3,000 a year against other income ($1,500 if married filing separately), and the rest carries forward.
Selling a loser to bank the loss is tax-loss harvesting. Under the wash-sale rule in IRS Publication 550, buying substantially identical stock within 30 days before or after the sale disallows the loss for now and adds it to the new shares' cost. Let taxes choose between sales you already wanted. Don't let them start one. These are US rules, and the day trading taxes guide covers how short-term profits get taxed.
Frequently Asked Questions
How do you know when to sell a stock?
Mostly, you decided earlier. Before buying, write down why you own it and what would prove you wrong: a business change for an investor, a price level for a trader. When one of those happens, you sell. When neither has, a scary day or a big gain is just noise.
Should I sell a stock that has doubled?
Not just because it doubled. Check two things: does the reason you bought still hold, and is it now too big a slice of your portfolio? Some investors sell half on a double, which takes the original stake off the table and lets the rest run. In a taxable account, check the holding period first.
Should I sell a stock at a loss?
Yes, if the reason you bought is gone or price hit the stop you set at entry. What you paid doesn't matter to the stock. A useful test: would you buy it today, at this price, with fresh money? If the answer is no and nothing says it's temporary, holding is just hoping.
What is a trailing stop?
A sell stop that moves up with price and stays put when price falls. Investor.gov's example: you buy at $20, it reaches $22 and you set a $1 trailing stop. It peaks at $24, so the stop sits at $23, and you're sold when it trades back to $23, though the fill can differ.
Does a stop loss guarantee my price?
No. A stop becomes a market order once triggered and fills at whatever the market pays. If a stock gaps below your stop at the open, you sell near wherever it opens, which can be well below your stop. A stop-limit caps how low you'll sell, but if price drops through the limit it may not fill at all, and you're still in.
This article is for educational and informational purposes only and does not constitute financial or tax advice. William O'Neil's 7% to 8% and 20% to 25% rules are quoted as one investor's published rules of thumb, not tested findings. Research findings are summarized from the linked papers, and the recovery table is arithmetic. Tax figures are US federal rules from IRS pages read in October 2026 and change over time, so check with a tax professional. The exit diagram is a schematic, not real prices. Trading carries a substantial risk of loss. SnapPChart grades a setup to enter from one chart screenshot. It does not monitor positions, send alerts, tell you when to sell a stock you hold, read fundamentals, valuation, news or taxes, or guarantee outcomes. Always do your own research and never trade with money you cannot afford to lose.
Writes about AI-assisted day trading, technical analysis, and the systems traders actually use to stay disciplined.
Decide the exit before you're in the trade.
SnapPChart grades a setup before you're in it. You upload one chart screenshot and it reads the trend, the swing structure and the support and resistance it can see, then grades the setup from A+ to F and says whether to enter now, wait for a pullback to a named level, or pass. It's a one-time read, not a monitor, so it won't tell you when to sell a stock you already hold. Your first grade is free: a summarized analysis of your chart and a graded second opinion on whether to take it, wait or pass. Entry, stop, targets and the trade plan come with a paid plan.