# How to tell if a finance influencer actually knows what they're talking about > Anyone can sound confident about a stock. Here's how to separate a real track record from luck, survivorship bias, and selective memory, and how Quantral scores credibility. By Maya Koeva · 2026-06-19  Open finance X or a busy investing subreddit and you'll find no shortage of people telling you what to buy. Some are genuinely sharp. Most are not. And almost all of them sound equally certain. The hard part was never finding opinions; it's working out whose opinions have actually earned your attention. This is a guide to doing that: what a track record really is, the traps that make weak voices look strong, and how Quantral turns all of it into a single credibility measure. ## A track record is more than a few good calls A track record is the full, graded history of someone's public calls: what they said, when they said it, and what happened afterward. The important word is **full**. One viral "I called NVIDIA at $40" screenshot is not a track record. It's a single data point, and usually the one they chose to show you. To mean anything, a track record needs three things: - **Enough calls to rule out luck.** A coin can land heads five times in a row. So can a stock picker. - **Every call counted**, not just the ones that worked. - **A consistent rule** for what counts as right and wrong, applied the same way every time. Without those, you're not looking at evidence. You're looking at a highlight reel. ## Hit rate, and why it isn't the whole story The most obvious measure is **hit rate**: the share of someone's calls that worked out. It's useful, but on its own it's easy to misread. **Sample size matters.** A 70% hit rate across three calls tells you almost nothing. The same 70% across two hundred calls tells you a lot. The more calls behind a number, the harder it is for luck to explain it. **Base rates matter.** In a rising market, "buy" calls look right by default. Being correct 60% of the time when almost everything went up is not impressive. What matters is how someone did relative to how easy the period was. **Magnitude matters.** Ten small correct calls can be wiped out by one catastrophic wrong one. A picker who is "usually right" but occasionally disastrous may be worse than one who is right less often but never blows up. ## The traps that make weak voices look strong Most of the work in judging a voice is spotting the ways a mediocre one can look brilliant. **Survivorship bias.** The loudest accounts are often just the ones whose risky bets happened to pay off. The equally loud accounts that made the opposite bet and blew up went quiet, or deleted their posts. You're seeing the winners of a lottery and mistaking them for the skilled. **Cherry-picking and the deleted-post problem.** It costs nothing to broadcast a win and quietly remove a loss. A track record you can trust is the one a person can't quietly edit after the fact. **Confidence is not accuracy.** Conviction, clean charts, and good writing all *feel* like skill. They are not the same thing. Some of the most confident voices online are also the least calibrated, and the certainty is exactly what makes them persuasive. **Recency.** One great call last week can erase the memory of a year of mediocre ones, for the author and for you. Judge the body of work, not the latest post. ## What to actually look for When you're sizing up a voice, a short checklist goes a long way: - A **real sample**: many calls over a meaningful stretch of time, not a highlight reel. - **Timestamps** you can verify, so the call came before the move, not after. - **Losses acknowledged**, not hidden. Honesty about being wrong is a green flag, not a weakness. - **Calibrated language** ("I think, and I've sized it small") over absolute certainty ("100%, all in"). - **Reasoning**, not just a ticker and a rocket emoji. You want to understand *why*, so you can judge whether the logic holds. - **Consistency across conditions**, not just one lucky run in one kind of market. ## How Quantral measures credibility Checking all of that by hand, for every account, on every stock, is exactly the work that doesn't scale. It's the work Quantral does for you. We track a curated set of authors on finance X and Reddit and grade their calls over time. For an author's history to count toward our **trusted voice** bar, we require at least five graded calls, with more than half of them correct. That accuracy then feeds the [signal score](/learn/what-is-a-stock-signal): a mention from a proven voice moves the needle more than the same mention from an anonymous account making its first call. This is why two stocks with nearly identical mention counts can earn very different scores. If most of one stock's attention comes from voices with a track record and most of another's comes from the anonymous crowd, Quantral treats them differently, because you should too. You can see exactly that effect in our [monthly most-mentioned ranking](/blog/most-mentioned-stocks-june-2026), where the "trusted mentions" share often tells a different story than the raw volume. ## The bottom line A confident post is not a track record. Before you give anyone's opinion weight, ask the boring questions: How many calls? Counted how? Over what kind of market? And were the losses left in? The voices actually worth following are usually the ones being honest about the times they got it wrong. --- *Quantral surfaces signals and context from public sources to support your own research. Nothing here is financial advice or a recommendation to buy or sell.*