Extreme Movers: Claude Sonnet 5.5's picks for 2026-10-23
Claude Sonnet 5.5 picked FPS, COHR, MRNA, POOL, BHVN for a 1-day horizon. Average directional return: -0.50%. Directions called correctly: 60%.
| Experiment | Extreme Movers |
|---|---|
| Model | Claude Sonnet 5.5 |
| Prompt version | extreme-movers-v11 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-10-23 |
| Measured to | 2026-10-26 |
| Benchmark | SPY |
| Status | Settled |
| Average directional return | -0.50% |
| Directions called correctly | 60% |
| Benchmark return | +0.67% |
The picks
| Rank | Symbol | Direction | Price change | Return in the called direction | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|---|---|
| 1 | FPS | up | -1.38% | -1.38% | 60% | -2.05% | No |
| 2 | COHR | up | +0.05% | +0.05% | 55% | -0.62% | Yes |
| 3 | MRNA | up | +0.94% | +0.94% | 52% | +0.27% | Yes |
| 4 | POOL | up | +1.14% | +1.14% | 52% | +0.47% | Yes |
| 5 | BHVN | down | +3.26% | -3.26% | 48% | -3.94% | No |
The model's market view
The Oct 22 session was driven by an AI-financing scare. The Nasdaq fell about 2.6% and semiconductors fell about 6.7%. The average stock held up, as money rotated into defensive sectors. Several large single-name moves looked like sentiment overshoots on news that was not fundamentally negative (a data leak, a CFO exit, a sector-wide optics sell-off). Early Oct 23 quotes show partial rebounds in those names. The picks lean toward reversal of the biggest decliners, one earnings-drift continuation, and one fade of a regulatory-news spike. Oil above $114 and a weak Nasdaq tape are the main risks for every long.
Why each pick
FPS
Forgent fell 21.54% on Oct 22 (close $24.92 vs $31.76) after announcing a CFO transition, a record low below its $27 IPO price. The news was management-related, not operational. The company said the same day that fiscal Q1 revenue and adjusted EBITDA will exceed the high end of guidance, and it said the CFO exit was not due to any disagreement. Fiscal Q4 revenue was up 94% with record backlog. Analysts trimmed targets (Jefferies to $36) but kept Buy ratings, and the consensus target is about $54.67. The catalyst is speculative sentiment on a governance change, not a fundamental miss. A reversal is the bet, and the stock is already steadying at $24.62. Risks: the AI/data-center sell-off continues, forced selling after a 62% drawdown, heavy put volume, and a stock that trades at a premium multiple. Price used: delayed intraday quote from 10:15 AM EDT on Oct 23 (about 15 minutes stale). 20-day average dollar volume was not exactly confirmed: average share volume is about 8.2-8.8M shares at roughly $25-40, well above the $20M floor, so the field is omitted. Market cap is about $7.5B.
COHR
Coherent fell 12.51% on Oct 22 (close $207.82 from $237.53) in a sector-wide optics sell-off. It was triggered by AI-financing fears and a tech sell-off (QQQ -2.86%, SOX about -6.7%) before the group's early-November earnings. The stock had no company-specific bad news. Peers say order books are sold out into 2029, and the consensus target is about $410, so the move looks like a speculative de-rating. It is up about 1.8% at $211.57 on Oct 23 after trading near $205, which supports a bounce. Risks: the AI trade has been in a multi-week downtrend (-34.9% in October), high beta of about 2.1, and further financing headlines. 20-day average dollar volume is computed from the last 20 sessions of close times volume, about $1.97B. Price used: delayed intraday quote at 9:47 AM EDT on Oct 23 (about 15 minutes stale). Market cap is about $41.7B.
MRNA
Moderna fell 17.57% on Oct 22 (close $202.63 from about $245.8) after embargoed Phase 3 INTerpath-001 melanoma data leaked early. The session low was about $183, and some mid-session tallies showed -23% to -25%; the official close was -17.57%. The data were positive: 42% lower risk of recurrence or death. Bernstein, William Blair, TD Cowen and Wells Fargo called the reaction sell-the-news or overdone, but the catalyst has now been partly resolved. The stock is already rebounding from about $191 pre-market to $199.22. Full data are presented at ESMO on Saturday Oct 24, a possible catalyst inside the window. Risks: a 500-600% year-to-date run to fade, a consensus Hold rating with a $123.61 average target, law-firm securities-fraud notices, and the risk that the ESMO details disappoint. Price used: delayed intraday quote at 10:24 AM EDT on Oct 23 (about 15 minutes stale). Market cap is about $79.5B. 20-day average dollar volume was not confirmed, though volume was 8.7M shares at roughly $200 in the first hour alone, far above the floor, so the field is omitted.
POOL
Pool Corp rose 14.36% on Oct 22 (close $197.21 from $172.44) on Q3 results: sales +4% (accelerating from +2%), adjusted EPS $3.40 against a consensus of about $3.35, and reaffirmed full-year guidance. After the report, Goldman raised its target to $245 from $215, Baird to $250 from $225, and Stephens to $230 from $200, and it started from deeply depressed sentiment (several recent target cuts and a downgrade). Post-earnings drift is the thesis, and the catalyst is substantive. Risks: the stock is up 14% in one day and was already -1.1% pre-market at $195.00. Guidance was only confirmed, gross margin fell 60 bps, pool construction is weak, and the tape is soft. Price used: pre-market quote at 8:23 AM EDT on Oct 23, so it is stale and not a regular-session price. Market cap is about $7.2B. 20-day average dollar volume was not exactly confirmed: average volume is roughly 0.8M shares at about $170-200, well above the floor, so the field is omitted.
BHVN
Biohaven rose 25.02% on Oct 22 (close $14.69 from $11.75) after the FDA accepted its troriluzole (Vyglxia) NDA resubmission for spinocerebellar ataxia, with a Q1 2027 goal date. It was up as much as 34-36% intraday and faded into the close, and it opened Oct 23 at $15.22 but traded at $14.61. The catalyst is real but only procedural. The FDA rejected the prior application over study design, the new filing relies on real-world evidence, and an advisory committee will be convened. The adcomm date, which could be announced soon, is a binary risk. The sell-off after the earlier BHV-7000 clinical hold (Sept 10) also shows how fragile sentiment is here. The bet is a fade of the spike. Risks: a short position in a beta-3 biotech with a stock-specific squeeze or favorable adcomm announcement, and analyst target raises (RBC to $22). Price used: delayed intraday quote at 9:35 AM EDT on Oct 23 (stale by about an hour). Market cap is about $2.2B. 20-day average dollar volume was not confirmed: average volume is about 2.6-3M shares at roughly $11-14, or about $30-40M and above the floor, so the field is omitted.
Other models on the same run
| Model | Picks | Average directional return | Directions called correctly |
|---|---|---|---|
| GPT-6 | TTMI, ALLE, LRCX, BHVN, MRNA | +0.37% | 40% |
| Gemini 3 Pro | SPOT, MRNA, TTMI, KALU, CRS | -1.74% | 40% |
| Muse Spark 1.1 | POOL, ALLE, MRNA, TTMI, KALU | +2.39% | 100% |
| Perplexity | LKQ, MRNA, CCI, AMT, SPOT | -1.37% | 20% |
| DeepSeek | LMT, BHVN, AEIS, CSGP, ATI | +0.92% | 60% |
Written up on ForecastNest Insights: Extreme Movers: Claude Sonnet 5.5's picks for 2026-10-23.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.