No honest page can hand you a list that fits everyone. This one gives you the method instead: five screening steps, the historical evidence for each, dated examples that went well and badly, and the mistakes that cost investors most.
No single list of “best stocks to buy now” fits everyone; the shortlists that held up best historically combined a fair benchmark, a profitable business at a sane price, agreement among analysts with strong records, and a position size the owner could hold through a crash. Here is the evidence for each step and how to build your own.
A list is the wrong starting point for three reasons. It goes stale within weeks as prices and ratings move. It ignores what you already own, so the same stock can diversify one portfolio and double the main risk of another. And it hides the base rate: most stocks do worse than a plain index fund. In Exent’s data, 155 of 422 of today’s largest US-listed companies beat the S&P 500 index over the 10 years to Sep 30, 2026; the median one gained 239.9% against 312.9% for the index fund.
Professionals face the same odds. According to S&P Dow Jones Indices’ SPIVA U.S. Year-End 2025 scorecard, 79% of active large-cap US funds trailed the S&P 500 in 2025. And over the long run, Hendrik Bessembinder’s 2018 study found that 1,092 companies, 4.3% of nearly 26,000 US stocks, accounted for all the market’s wealth creation above Treasury bills from 1926 to 2016. Picking stocks is a hunt for a small minority, so the method matters more than any single name.
Run every idea through the same five steps, in order. Each row shows what to check and what the data says about that criterion on its own. Figures marked “our data” are as of Sep 30, 2026 unless dated otherwise.
| Step | What to check | What the data shows | Source |
|---|---|---|---|
| 1. Benchmark first | Would an S&P 500 index fund do the same job with less risk? | 36.7% of 422 large US stocks beat the index over 10 years (survivors only); 79% of active large-cap funds trailed it in 2025 | Our data; SPIVA 2025 |
| 2. Business quality | Several years of revenue growth, steady or rising margins, free cash flow behind the earnings, debt it can carry | 180 of 794 companies scored in the Altman Z-score distress zone; 10x winners of 2009–2024 skewed toward profitable firms with a high free cash flow yield | Our data; Yartseva (2025) |
| 3. Price and trend | Valuation against its own history and peers; how far it has already run | Cheapness or trend alone did not beat the index: a value fund returned 11.4% a year and a momentum fund 16.2%, vs 15.3% for the S&P 500 (Jul 2016 to Sep 2026) | Our data (fund returns) |
| 4. Who agrees | Which analysts rate it, and their past accuracy, not the average rating | 70.8% of ratings were Buys; upgrades by 5-star analysts beat the S&P 500 over 3 months 60.0% of the time vs 49.8% for unranked analysts | Our analyst data |
| 5. Fit and size | Correlation with your largest holdings, the weight you start at, its worst past fall | Nvidia fell 85% in 2007–09 and 62% in 2022; the median uploaded portfolio held 77% in its top five positions | Our data |
The pattern in the third column is the point: no single criterion was enough. A quality fund returned 14.2% a year over the same period, also below the index, and the momentum fund’s worst fall was 30.2% against 23.9% for the S&P 500. The steps work as filters that remove weak ideas, not as a formula that finds winners. A step-by-step checklist for one idea is in how to evaluate a stock idea.
Mostly profitable, growing and not obviously expensive, though plenty of companies with the same traits went nowhere. A 2025 working paper by Anna Yartseva, “The Alchemy of Multibagger Stocks”, studied 464 US stocks that rose at least tenfold between 2009 and 2024. Small, profitable companies at modest valuations with a high free cash flow yield did best; price momentum helped but reversed quickly. It is a working paper, not yet peer reviewed.
The balance sheet is the quickest filter. In our data, 180 of 794 US companies with the needed figures scored in the Altman Z-score “distress” zone, a classic warning of financial strain. A stock can still rise from there, but it is a different bet from a cash-rich business. The full research summary is in what the biggest stock winners had in common, and the odds of finding the next one in how to find the next Nvidia.
Analysts say “buy” to most stocks, so the rating alone tells you little; their track record tells you more. In Exent analyst data covering 921,573 calls from 1,121 ranked analysts over 7 years (statistics as of Oct 20, 2026), 70.8% of ratings in the last 12 months were Buys and 2.9% were Sells.
| Analyst group | Analysts | Price targets reached (median) | Calls higher after 12 months (median) | Calls not higher after 12 months |
|---|---|---|---|---|
| 5 stars | 142 | 60.5% | 68.6% | 31.4% |
| 4 stars | 966 | 54.4% | 63.2% | 36.8% |
| All calls pooled | 1,121 | 53.6% | 64.6% | 35.4% |
| Analyst group | 5 stars | 4 stars | All calls pooled |
|---|---|---|---|
| Analysts | 142 | 966 | 1,121 |
| Price targets reached (median) | 60.5% | 54.4% | 53.6% |
| Calls higher after 12 months (median) | 68.6% | 63.2% | 64.6% |
| Calls not higher after 12 months | 31.4% | 36.8% | 35.4% |
So even the best-rated analysts were wrong about direction roughly three times in ten, and after the average call the stock fell 22.5% below the rating-day price at its lowest point in the following year. Star ratings come from those same past calls, so the gap between groups describes history, not a forecast, and the window was mostly a rising market.
Fresh moves by strong analysts carried a little more information. Across 11,514 upgrades from 2019 to Sep 15, 2026, stocks upgraded by 5-star analysts beat the S&P 500 over the next three months 60.0% of the time, against 49.8% for analysts outside the ranking. Details in did analysts see the biggest winners coming and on the analyst rankings; the method behind “where top analysts agree” is on what top analysts are recommending.
Both, often in the same stock. The table mixes outcomes on purpose. Calendar-year total returns are from our data as of Sep 30, 2026.
| Example | When | What happened | Lesson |
|---|---|---|---|
| NVIDIA (NVDA) | 10 years to Sep 30, 2026 | +13575% in total, yet −50.3% in 2022 alone | Winners fell hard on the way; size decided who held on |
| ARK Innovation ETF (ARKK) | 2020 to 2022 | +152.7% in 2020, then −23.4% and −67.0% | Buying after the hot year was the costly part |
| Zoom (ZM) | 2020 to 2022 | +395.8% in 2020, then −45.5% and −63.2% | A real business boom does not protect the price |
| PayPal (PYPL) | 2020 to 2022 | +116.5% in 2020, then −19.5% and −62.2% | A well-known leader can still halve |
| Intel (INTC) | 2023 to 2024 | +94.5% in 2023, then −59.6% in 2024 | A rebound year is not a trend |
| BioMarin Pharmaceutical (BMRN) | Targets set from 2019 | Only 23.0% of 586 analyst price targets were reached within 12 months (median stock: 68.6%) | Persistent analyst optimism can be wrong for years |
| 12 past Gems shown on /radar | Flagged Jan 8, 2026 to Apr 1, 2026 | +47% to +226% from flag date to the exit date shown | Selected winners only; says nothing about the misses |
Read the last row with care. It is the public sample of 12 past Gems on the Stocks Radar page: Exent’s recorded figures, not independently audited, chosen to show stocks that worked. A selected set of winners cannot tell you a hit rate. Judge any picks service, ours included, on a complete time-stamped record with losses, as explained in has following top analysts’ consensus picks worked.
Avoid the habits that turn a reasonable idea into a large loss. Several popular signals also looked weaker in the data than their reputation suggests:
| Popular signal | What happened next, in our data |
|---|---|
| Insider buying | 17,865 open-market insider buys since Feb 2012: 51.4% beat the S&P 500 over 12 months; median +1.0% vs the index |
| Joining the S&P 500 | 277 additions since 2013: median −6.4% vs the index over the next 12 months; 39.8% beat it |
| Beating earnings estimates | Across 12,015 quarterly reports, the stock fell after 43.3% of earnings beats |
More on the first two in how to stop chasing hot stocks and is it too late to buy a stock after it doubled.
No list fits everyone, and most stocks lag the index: 36.7% of today’s largest US stocks beat the S&P 500 over the 10 years to Sep 30, 2026. A sounder approach is to screen for a profitable business at a reasonable price, check which analysts with strong records agree, and size each position so you can hold it through a crash.
Most of them: 70.8% of analyst ratings in the past 12 months were Buys and 2.9% were Sells. More useful is where analysts with strong past records agree; even so, the median 5-star analyst’s calls were not higher a year later 31.4% of the time.
Nobody can know. Nvidia returned about 13,575% over the 10 years to Sep 30, 2026 and +22.4% over the last year, but it also fell 85% in 2007–09 and 62% in 2022. 47 of 49 covering analysts rated it a Buy. The better question is how large a position you could hold through another fall of that size.
A classic 1987 study by Meir Statman found a well-diversified portfolio of randomly chosen stocks needs at least 30 stocks, or 40 for some investors. Stocks that move together count as one bet, so correlation matters as much as the count.
Any current list would be a guess dressed as a fact. Past AI winners such as Nvidia also had falls of more than half, and buying several AI stocks often means owning one theme several times. The same five steps apply: business, price, who agrees, fit and size.
Only weakly. In Exent’s data, 51.4% of insider buys since 2012 beat the S&P 500 over the next 12 months, and stocks fell after 43.3% of earnings beats.
Exent tracks 4,000+ analysts, keeps only the elite few with proven records, and surfaces the Gems where they agree.
Educational content, not investment advice. Examples are illustrative; past performance does not guarantee future results.