Reddit, StockTwits trending and the WallStreetBets daily thread, scraped every morning before the US open, Mar 07 - Oct 13, 2026. Returns from that day's open, compared with SPY. Not investment advice.
Stock mentions studied
5,901
Reddit bullish posts: 20 days vs SPY
-1.6%
StockTwits trending after a 60%+ week: 20 days vs SPY
-35.0%
Multi-bagger picks that had social buzz
4%
The short answer
- Bullish vs bearish chatter does not matter. Reddit's bullish and bearish posts were followed by the same returns.
- YOLO posts and call-buying talk predict nothing.
- Hyped cheap stocks fall. The cheaper the stock people talk about, the worse it does afterwards.
- StockTwits trending is a 'late to the party' signal: stocks that trend after already running 30-60%+ in a week usually give it back - and the bigger the run, the bigger the fall.
- Our multi-bagger model almost never picks talked-about stocks (4% of picks) - it buys what nobody is discussing.
1. All sources
| Source | Mentions | 5 days vs SPY | 20 days vs SPY | Beat SPY (20d) | Lost 30%+ |
|---|
| Reddit | 3,435 | -0.5% | -1.2% | 46% | 4% |
| StockTwits trending | 1,618 | -0.9% | -2.7% | 43% | 6% |
| WSB daily thread | 848 | +0.1% | +0.0% | 50% | 3% |
2. Sentiment (AI-rated from the posts)
| Source: sentiment | Mentions | 5 days vs SPY | 20 days vs SPY | Beat SPY (20d) | Lost 30%+ |
|---|
| Reddit: bearish | 692 | -0.4% | -1.5% | 44% | 4% |
| Reddit: bullish | 1,692 | -0.4% | -1.6% | 45% | 4% |
| Reddit: neutral/mixed | 1,051 | -0.7% | -0.3% | 49% | 3% |
| WSB daily thread: bearish | 219 | +0.2% | +0.0% | 50% | 4% |
| WSB daily thread: bullish | 454 | +0.4% | +0.2% | 51% | 3% |
| WSB daily thread: neutral/mixed | 175 | -0.7% | -0.8% | 48% | 2% |
3. YOLO, options and gain/loss posts (Reddit + WSB thread)
| Posts mentioning | Mentions | 5 days vs SPY | 20 days vs SPY | Beat SPY (20d) | Lost 30%+ |
|---|
| YOLO | 124 | -2.2% | +1.4% | 53% | 5% |
| Calls | 271 | -0.8% | +0.0% | 50% | 3% |
| Puts | 128 | -0.7% | -1.1% | 43% | 6% |
| Bragging about gains | 47 | +1.4% | +7.7% | 62% | 10% |
| Loss posts | 8 | +0.9% | +1.7% | 50% | 0% |
| None of these | 3,773 | -0.3% | -1.1% | 46% | 3% |
Gain-bragging posts look positive (+7.7%) but cover only 47 mentions of 28 stocks - a hint at most.
4. Price of the stock being talked about
| Share price | Mentions | 5 days vs SPY | 20 days vs SPY | Beat SPY (20d) | Lost 30%+ |
|---|
| Under $5 | 170 | -2.2% | -8.3% | 35% | 15% |
| $5-20 | 1,125 | -1.4% | -4.8% | 40% | 6% |
| Over $20 | 4,606 | -0.3% | -0.8% | 47% | 4% |
5. StockTwits trending, in detail
StockTwits' own sentiment was not captured by the scraper (always "neutral"), so this section uses price behaviour around the trending day.
| Move in the 5 days before trending | Entries | Trending day open-close | 20 days vs SPY | Beat SPY | Lost 30%+ |
|---|
| Fell 10%+ | 154 | +0.2% | -0.5% | 49% | 9% |
| Flat | 1,145 | -0.2% | -2.4% | 44% | 3% |
| Up 10-30% | 234 | +1.7% | -6.0% | 42% | 6% |
| Up 30-60% | 42 | +4.8% | -18.8% | 40% | 30% |
| Up 60%+ | 25 | +14.4% | -35.0% | 17% | 61% |
The bigger the run before a stock trends, the harder it falls afterwards: stocks up 60%+ in the prior week often still rise on the trending day itself (the last burst of buying), then lose most of it - 61% fell another 30%+ within a month.
| Gap at the open on the trending day | Entries | Trending day open-close | 20 days vs SPY | Beat SPY | Lost 30%+ |
|---|
| Gap down 5%+ | 88 | -1.7% | -6.5% | 38% | 12% |
| No big gap | 1,354 | -0.0% | -2.5% | 44% | 5% |
| Gap up 5-20% | 141 | +3.4% | -7.4% | 44% | 13% |
| Gap up 20%+ | 30 | +8.7% | -8.7% | 35% | 31% |
Most frequent trending names: DJT (22 days), SNDK (15 days), ASTS (14 days), SLS (14 days), GME (13 days), ORCL (13 days), MU (13 days), SOFI (13 days), MSFT (13 days), AAPL (13 days), NBIS (12 days), NVDA (11 days).
6. Our multi-bagger model and social buzz
Of the model's 310 weekly picks since March, only 13 (4%) were mentioned on Reddit, StockTwits trending or the WSB thread in the week before the buy - 2 on StockTwits, none of them the 'trending after a big run' kind. The model picks under-followed stocks, which fits an earlier finding that its picks with no recent news did best. An 'avoid hyped stocks' filter would almost never trigger for it.
| Pick | Signal week | Mentioned on | Result so far |
|---|
| WTI | 2026-04-10 | WSB thread | -2.7% |
| SABR | 2026-04-17 | Reddit | +3.3% |
| XRX | 2026-05-08 | Reddit, WSB thread | -8.3% |
| SG | 2026-05-22 | Reddit | -25.5% |
| MRVI | 2026-06-12 | WSB thread | +40.9% |
| HIVE | 2026-06-18 | WSB thread | -20.2% |
| MRVI | 2026-06-18 | WSB thread | +40.9% |
| HIMS | 2026-06-18 | WSB thread | -15.9% |
| HIMS | 2026-07-02 | Reddit | -15.9% |
| XRX | 2026-07-31 | WSB thread | -8.3% |
| ABCL | 2026-08-07 | StockTwits | +43.2% |
| AMC | 2026-09-11 | Reddit | +21.9% |
| CAPR | 2026-10-02 | StockTwits | -10.5% |
How much to trust this
- Only 7 months of data in one market; same-day mentions move together, so the real sample is much smaller than the counts.
- About 40 slices were tested, so one or two will look strong by chance. The two patterns above (cheap hyped stocks and StockTwits after a big run) are the consistent ones - they get stronger step by step and match published research on retail attention and short-term reversals.
- Use buzz as a warning flag, not a buy signal. The study will be rerun as more data comes in.
Discussion
No comments yet.