Oct 26, 2026 - 6 AI models call the biggest movers in the morning session. 3 models finished positive

Asked Oct 26, 10:05 AM EDT (morning session). Muse Spark 1.1, Gemini 2.5 Pro, Perplexity, Grok 4.7, Claude Sonnet 5.5 and DeepSeek answered the same standard prompt independently. Window: from Oct 26, 10:05 AM EDT to Oct 27, 10:05 AM EDT, the same time one session later. Benchmark: SPY.

Five direction-adjusted close-to-close returns; a down call reverses the sign. The score is the mean signed return, without subtracting the benchmark. In each chart SPY is on the left and the picks on the right, every candle drawn from zero to that asset's move; ✓ marks a pick that moved the way it was called.

Model	Correct	Hit rate	Return	SPY	Directional return
Muse Spark 1.1	4/5	80.00%	3.25%	0.36%	3.25%
Gemini 2.5 Pro	1/5	20.00%	-2.89%	0.36%	-2.89%
Perplexity	4/5	80.00%	1.78%	0.36%	1.78%
Grok 4.7	2/5	40.00%	-2.26%	0.36%	-2.26%
Claude Sonnet 5.5	3/5	60.00%	1.95%	0.36%	1.95%
DeepSeek	2/5	40.00%	-1.41%	0.36%	-1.41%

Muse Spark 1.1

View full picks and reasoning

S&P 500 closed Oct 23 at 7623.68 down 0.52% on rising Treasury yields and Brent crude spiking to $116, with 3262 advancers vs 6348 decliners. Energy led at +2.55% on oil supply fears in Strait of Hormuz, Real Estate +0.36% on tower deal hopes, Technology +0.34% on cybersecurity, while Consumer Staples lagged -1.82% on 【entity-University of Michigan¦canonical_name=University of Michigan】 inflation expectations rising to 4.7%. Top themes: 【entity-SpaceX¦canonical_name=SpaceX】 negotiating access to 80,000 Crown Castle sites to build a fourth national wireless network pressuring 【entity-AT&T¦canonical_name=AT&T】, 【entity-Verizon¦canonical_name=Verizon】, 【entity-T-Mobile¦canonical_name=T-Mobile】, while lifting CCI, AMT, SBAC; cybersecurity rally on two critical PHP vulnerabilities and Deutsche Bank naming CRWD and PANW as AI security winners; Bloom Energy sold despite doubling revenue and raising outlook. For Oct 27, oil volatility and Fed inflation test keep market choppy, favoring continuation in cyber and tower momentum and potential reversal bounce in oversold earnings-beat losers.

1. CCI up — confidence 68%. Ranked #1 S&P 500 gainer at Oct 23 close, closed at $85.97 up 9.06% vs prior day from $78.83 on volume 15.6M shares, market value $36.6B. Catalyst is report 【entity-SpaceX¦canonical_name=SpaceX】 is in talks to lease approximately 80,000 Crown Castle tower sites to build a nationwide wireless network, potentially worth billions and providing a major new revenue source, making CCI top-performing S&P 500 stock Friday per TipRanks. Catalyst is substantive if executed but currently speculative talks, with sympathy gains in AMT and SBAC. Passes filters: price >=5, market cap $36.76B >=1B, avg dollar volume $305.85M >=20M. Reference price $88.225 is latest intraday Oct 26 quote from Finnhub, about 2.6% above Oct 23 close, fresh within session. Risk: no binding contract, prior quarter revenue down 7.2%, rate pressure. Thesis is continuation up on deal optimism and momentum.
   Return: -1.45%; alpha: -1.80%.
2. CRWD up — confidence 65%. Ranked #2 S&P 500 gainer Oct 23, closed at $257.57 up 8.25% per TickerBrief, with high $257.91 low $234.99. Catalyst is cybersecurity rally after two critical PHP vulnerabilities raised cyberattack concerns and Deutsche Bank named CRWD and PANW as winners from rising AI token usage, driving demand for AI-driven security tools. Catalyst is substantive: real vulnerability-driven demand, not just technical squeeze. Passes filters: price $257.57 >=5, market cap $262.47B >=1B, avg daily traded value $1.78B >=20M. Reference price $255.6 is current intraday Oct 26 quote from Finnhub, slightly below Oct 23 close, fresh delayed quote. Risk: stock up 13% weekly and near 52-week high $286.99, valuation stretched at 48x sales. Thesis continuation up as cyber bid persists.
   Return: 3.54%; alpha: 3.18%.
3. BE up — confidence 62%. Worst S&P 500 performer Oct 23, closed at $193.30 down 11.24% from $217.77 on 36M shares traded, despite Q3 revenue more than doubling to $1.24B and raising 2026 outlook to $4.4-4.6B for third time. Catalyst is cash-flow timing fear: Bloom counts sale when control passes but billing follows contract milestones, so cash arrives later quarter, plus peer FuelCell down only 4.1% and Plug flat, so move is idiosyncratic. Trefis notes drop was far larger than peers. Catalyst is substantive earnings beat but speculative cash worry, creating oversold setup about 44% below record close $346 on June 22. Passes filters: price $193.30 >=5, market cap $57.15B >=1B, avg daily traded value $2.63B >=20M. Reference price $197.34 is latest Oct 26 intraday from Finnhub, up ~2% from close, fresh. Risk: beta 3.7, continued cash concerns. Thesis reversal up as market re-prices beat and outlook raise.
   Return: 11.20%; alpha: 10.84%.
4. T down — confidence 60%. Major S&P 500 loser Oct 23, closed at $20.93 down 7.96% on report SpaceX negotiating access to at least 80,000 cell sites across US to build fourth nationwide wireless network complementing Starlink, stoking competition fears for AT&T, Verizon, T-Mobile. Oppenheimer downgraded and Bernstein cut to Market Perform citing Starlink competition, softer pricing and Q4 headwinds, despite AT&T beating estimates and raising outlook Oct 21. Catalyst is substantive competitive threat but speculative long-term impact on $1.6T broadband market. Passes filters: price $20.93 >=5, market cap $141.03B >=1B, avg daily traded value $1.19B >=20M. Reference price $21.075 is current Oct 26 intraday from Finnhub, slightly above close, fresh. Risk: oversold RSI 31, dividend yield 5.3% may attract buyers. Thesis continuation down as SpaceX overhang persists into next session.
   Return: -0.64%; alpha: 0.28%.
5. PANW up — confidence 58%. Top 5 S&P 500 gainer Oct 23, closed at $379.14 up 6.82% vs prior $354.92, or $380.55 up 7.22% per TradingKey, on cybersecurity bid alongside CRWD. Catalyst is two critical PHP vulnerabilities plus platformization strategy, next-generation security ARR growth and Chronosphere and CyberArk acquisitions driving 106.6% YTD gain. Source GlobalBanking notes PANW 379.14 +6.82% last traded Oct 23. Catalyst is substantive: real security demand shift toward AI security, not pure momentum. Passes filters: price $380.55 >=5, market cap $311.41B >=1B, avg daily traded value $1.83B >=20M. Reference price $383.36 is latest Oct 26 intraday from Finnhub, slightly above close, fresh. Risk: valuation 947x P/E, 52-week high $432.33, pullback risk if tech rotates. Thesis continuation up as cyber remains defensive bid with oil spike.
   Return: 2.32%; alpha: 1.96%.

Full prediction and audit record

Gemini 2.5 Pro

View full picks and reasoning

US equity markets demonstrate selective sector rotation driven by headline-driven selloffs in telecom and healthcare alongside anticipation of Q3 earnings reports in industrial and utility sectors. Large-cap movers that suffered sharp speculative reactions are primed for mean-reversion, while high-quality earnings beat candidates retain upward momentum.

1. T up — confidence 75%. AT&T dropped 7.96% following speculative news reports that SpaceX plans an aggressive cellular satellite expansion. The market reaction appears overdone relative to near-term fundamental cash flows, creating a strong mean-reversion reversal opportunity.
   Return: -0.64%; alpha: -1.00%.
2. FICO up — confidence 72%. Fair Isaac fell 7.24% in recent trading amid pre-earnings repositioning ahead of its early November reports. Given solid core software margins and dominant market position, the sharp pullback presents a short-term bullish reversal opportunity.
   Return: -6.91%; alpha: -7.26%.
3. MRNA up — confidence 70%. Moderna fell 7.00% amid sector-wide profit taking following recent cancer vaccine developments. Strong institutional support and underlying pipeline valuation favor a technical bounce and direction reversal over the next trading session.
   Return: -8.46%; alpha: -8.82%.
4. AWK up — confidence 68%. American Water Works surged 4.87% on strong institutional buying and steady guidance ahead of utility sector earnings. Defensively positioned macro demand supports trend continuation to the upside.
   Return: -1.75%; alpha: -2.11%.
5. NUE up — confidence 65%. Nucor rose 3.55% ahead of its Q3 earnings release on positive industrial demand commentary and strong historical post-earnings performance, supporting continued upside momentum.
   Return: 3.33%; alpha: 2.97%.

Full prediction and audit record

Perplexity

View full picks and reasoning

The source session was Friday, October 23, 2026. Large-cap gainers were led by tower, cybersecurity, and energy-services names, while LKQ collapsed on a substantive operational and leadership shock. I favor selective continuation in earnings-backed SLB and the cybersecurity complex, while treating CCI's SpaceX-related move as more speculative. LKQ's negative catalyst is fundamental and likely to produce continued pressure, though its oversized decline creates bounce risk.

1. SLB up — confidence 62%. SLB rose 7.08% on October 23 after reporting Q3 adjusted EPS of $0.64 versus a $0.62 consensus, with revenue of $9.45 billion above the $9.27 billion estimate; the catalyst is substantive. I expect continuation because the earnings beat, free cash flow disclosure, and Middle East recovery commentary give the move a fundamental basis. Risk: energy-sector beta and profit taking after a large one-day rally. Reference price is the October 23 after-hours quote of $53.45; I could not refresh an intraday October 26 quote. The 20-day average dollar volume was not independently verified and is omitted, although the observed October 23 session dollar volume of roughly $1.01 billion far exceeds the $20 million floor.
   Return: 2.87%; alpha: 2.52%.
2. LKQ down — confidence 58%. LKQ fell 22.47% on October 23 after preliminary Q3 results missed expectations, the CEO was replaced, and the company retained Alvarez & Marsal to review its European operations; this is a substantive negative catalyst, not mere sentiment. I expect continuation lower as analysts cut estimates and investors reassess European execution, though the 22% drop and low valuation make a sharp relief bounce the main risk. Reference price is the October 23 after-hours quote of $17.50; I could not refresh an intraday October 26 quote. The 20-day average dollar volume was not independently verified and is omitted, although the October 23 session dollar volume of roughly $408 million far exceeds the $20 million floor.
   Return: -1.62%; alpha: 1.26%.
3. CCI up — confidence 54%. Crown Castle gained 9.06% on October 23 after Bloomberg reported advanced discussions with SpaceX concerning leases for tens of thousands of U.S. cell sites. The catalyst is speculative because no definitive agreement was announced, but the potential contract scale is large enough that momentum and headline follow-through could continue. I expect continuation higher, with reversal risk if talks are denied, delayed, or valued as less incremental than the market assumes. Reference price is the October 26 pre-market quote of $85.76, the freshest verified price found. The 20-day average dollar volume was not independently verified and is omitted, although the October 23 session dollar volume of roughly $1.49 billion far exceeds the $20 million floor.
   Return: -1.45%; alpha: -1.80%.
4. CRWD up — confidence 52%. CrowdStrike rose 8.25% on October 23 as cybersecurity stocks rallied on warnings about critical PHP flaws and expectations of increased remediation demand. The catalyst is partly substantive, since real vulnerabilities can drive near-term security spending, but the move is also thematic and momentum-driven. I expect continuation, supported by the sector rally and positive analyst-target revisions, while recognizing valuation and the possibility that the security-flaw narrative fades quickly. Reference price is the October 23 after-hours quote of $258.40; I could not refresh an intraday October 26 quote. The 20-day average dollar volume was not independently verified and is omitted, although the October 23 session dollar volume of roughly $3.45 billion far exceeds the $20 million floor.
   Return: 3.54%; alpha: 3.18%.
5. PANW up — confidence 50%. Palo Alto Networks gained 7.22% on October 23 in the same cybersecurity rally triggered by critical PHP vulnerabilities; the catalyst is a mix of a real security event and speculative sector positioning. I expect modest continuation as investors rotate toward security names, but the move is less fundamentally anchored than SLB's earnings beat and the stock already trades near the upper end of its recent range. Risk includes a broad technology pullback and profit taking. Reference price is the October 23 after-hours quote of $382.10; I could not refresh an intraday October 26 quote. The 20-day average dollar volume was not independently verified and is omitted, although the October 23 session dollar volume of roughly $2.51 billion far exceeds the $20 million floor.
   Return: 2.32%; alpha: 1.96%.

Full prediction and audit record

Grok 4.7

View full picks and reasoning

Friday's session showed sharp divergence with telecom pressure from SpaceX competition reports and strength in energy and cybersecurity. Extreme movers driven by company-specific catalysts (earnings misses, competitive trial data, leadership changes) are likely to see mixed follow-through, with some continuation on substantive negative news and potential fade on speculative gains.

1. LKQ down — confidence 65%. LKQ plunged 22.47% on Friday after announcing a CEO departure, preliminary Q3 results showing a loss and Europe organic revenue decline of 9.8%, and withdrawal of full-year guidance due to ERP disruptions. This is a substantive operational and leadership catalyst. Hypothesis is continuation of the decline as investors digest the guidance pull and pending Oct 29 full results. Reference price is the latest pre-market quote near 17.40 on Oct 26 (stale by a few hours from open). Market cap ~4.41B and 20-day average dollar volume well above 20M based on ~3.7M share average volume at ~20 price. Source confirms the close and move.
   Return: -1.62%; alpha: 1.26%.
2. CORT down — confidence 60%. Corcept fell 20.13% after rival Genmab reported strong Phase 3 data for Rina-S in platinum-resistant ovarian cancer, directly challenging Corcept's Lifyorli growth driver. This is a substantive competitive catalyst with no company news of its own. Hypothesis is continuation lower as the market reassesses peak sales potential. Reference price uses the Friday close of 85.86 as the most confirmed recent trade (pre-market quotes around 86 were thin and earlier). Market cap approximately 9.3B; 20-day average dollar volume exceeds 20M given ~1.2M share average volume at ~90-100 prices.
   Return: -5.06%; alpha: 4.70%.
3. RHI up — confidence 55%. Robert Half surged 21.74% on Friday after Q3 earnings that met or beat expectations on sales and EPS. The move is earnings-driven and substantive for a staffing name, though the size of the gain may invite some profit-taking. Hypothesis is continuation higher into the next session on the positive print and volume confirmation. Reference price is the latest pre-market quote of 45.82 on Oct 26. Market cap ~4.7B; 20-day average dollar volume comfortably above 20M based on historical volumes around 1.5-2M shares at ~35-40 prices.
   Return: -5.34%; alpha: -5.70%.
4. BE down — confidence 58%. Bloom Energy dropped 11.24% despite reporting record Q3 results and raising guidance, as analysts focused on cash outflows, expansion costs, and cut some price targets. The catalyst mixes substantive earnings detail with sentiment-driven target reductions. Hypothesis is continuation of the decline as valuation and cash concerns linger. Reference price is the pre-market quote near 192.70 on Oct 26. Market cap ~57B; 20-day average dollar volume is very large (~16M shares average at ~200 prices).
   Return: 11.20%; alpha: -11.56%.
5. CCI up — confidence 50%. Crown Castle rose about 9% on reports that SpaceX is in talks to lease space on its cell towers for a nationwide wireless network. The catalyst is speculative (talks, not a signed deal) but sector-relevant given prior telecom weakness on the same theme. Hypothesis is continuation of the gain as the narrative supports tower demand. Reference price is the latest pre-market quote near 86.29 on Oct 26. Market cap ~36.6B; 20-day average dollar volume well above 20M (~4M shares average at ~80-85 prices).
   Return: -1.45%; alpha: -1.80%.

Full prediction and audit record

Claude Sonnet 5.5

View full picks and reasoning

Source session was Friday 2026-10-23: the S&P 500 fell 0.52%, the Dow fell 1.23%, energy led and telecom, consumer and AI-linked names lagged. Monday pre-market, index futures were higher into a heavy week of megacap earnings and a Fed decision. Verizon beat and raised guidance, Meta announced a large AI funding plan, and neocloud names fell after a Musk comment. Oil eased. I could not find live Monday quotes for any pick, so prices are Friday closes where I could confirm them. Picks lean toward fading telecom and fundamentals-driven AI-infrastructure selloffs, one post-earnings continuation, and one fade of a speculative tower-REIT rally. Fields I could not confirm are omitted.

1. RHI up — confidence 58%. Robert Half rose 21.74% on 2026-10-23 to a $45.87 regular-session close, a new 52-week high, on volume of about 6.9M versus roughly 2.0M average. The catalyst is substantive: Q3 results reported 2026-10-22 beat on EPS ($0.49 vs $0.47) with revenue in line, and commentary pointed to Talent Solutions returning to growth and higher Q4 guidance. I bet on continuation: post-earnings drift after a large re-rating, plus high short interest (reported near 23% of float in August, not refreshed). Risks are that the stock already sits above the average analyst target (about $39.56) and that a 22% gap often fades. Price used is the Friday 2026-10-23 close, one session stale; I found no Monday quote. Market cap is the Yahoo figure at the close. Screens: price above $5 and market cap above $1B confirmed; liquidity is well above $20M a day (about $80M average over the last 13 sessions, computed from published daily volumes and closes), but I could not confirm the exact 20-day figure, so that field is omitted.
   Return: -5.34%; alpha: -5.70%.
2. T up — confidence 54%. AT&T fell 7.96% on 2026-10-23 after a Bloomberg report that SpaceX is in private talks to lease about 80,000 cell towers. The catalyst is speculative: it is a report of talks, not a signed lease or a network launch. AT&T had also dropped about 5% on 2026-10-21 despite a Q3 beat and a raised profit outlook, so the stock was already oversold on the SpaceX narrative. Verizon rose about 2% pre-market on 2026-10-26 after beating and raising guidance, which gives a sympathy tailwind. Because the stock fell in the source session, a rebound is a reversal with direction up. Risks are further SpaceX headlines and FCC or spectrum news. I could not confirm the Friday close, a current quote, or the exact market cap or dollar volume, so those fields are omitted. Market cap was roughly $150B-160B in recent weeks and average volume is tens of millions of shares, so the screens are comfortably met in size, but the exact figures are unverified.
   Return: -0.64%; alpha: -1.00%.
3. BE up — confidence 52%. Bloom Energy fell 11.24% on 2026-10-23 to a $193.30 close, about 44% below its June record close, on volume of 34.7M versus a 15.9M three-month average. The catalyst is largely substantive but the reaction is a derating: Q3 revenue rose about 138% to a record $1.24B, adjusted EPS was $0.94 versus $0.69 expected, and full-year revenue guidance was raised for the third time to $4.4B-$4.6B. Selling centered on negative free cash flow of about $181M and a broad slide in AI-power stocks, and the CFO said the cash use is not the start of a trend. I bet on a reversal up as a beaten-down growth name bounces, helped by Monday's higher futures and Meta's AI funding news. Risks are high beta (above 3), the Musk neocloud comment weighing on AI infrastructure, and further profit-taking. Price is the Friday close, one session stale. Market cap is not directly confirmed (about $55B from the roughly 284M share count times the close) and 20-day dollar volume is clearly in the billions from the share volume, so both fields are omitted.
   Return: 11.20%; alpha: 10.84%.
4. FIX up — confidence 50%. Comfort Systems USA fell 9.26% on 2026-10-23, the largest decline among S&P 500 names that day, after reporting Q3 results on 2026-10-22. The catalyst for the drop looks like multiple compression rather than a fundamental problem, so the rebound case is substantive: revenue was $3.77B (up about 50%), diluted EPS $13.91 versus $8.25 a year earlier, same-store backlog $17.39B (up $3.8B sequentially), operating cash flow $1.01B in the quarter, and the quarterly dividend was raised 11%. The stock had already fallen about 4.6% on 2026-10-20 and sat roughly a third below its 52-week high before the report. I bet on a bounce, which is a reversal with direction up. Risks are continued AI and data-center capex derating and a premium valuation. I could not confirm the Friday close, a current quote or the exact market cap, so those fields are omitted. Market cap was about $51.6B at $1,464 on 2026-10-19, so it is well above the $1B floor, and average volume is roughly 0.5M shares at above $1,000 a share, far above the $20M dollar-volume floor.
   Return: 3.11%; alpha: 2.75%.
5. CCI down — confidence 42%. Crown Castle rose 9.06% on 2026-10-23 on the same SpaceX report, which suggested SpaceX may lease up to 80,000 U.S. cell sites. The catalyst is speculative: Bernstein and others noted that buying spectrum or discussing leases does not commit SpaceX to a ground network rollout, no lease has been signed, and sell-side estimates do not include SpaceX revenue. I bet on a partial giveback of a headline-driven pop, which is a reversal with direction down, and it also offsets my long positions in the carriers. Risks are strong momentum in tower names since the 2026-10-09 spectrum news, possible confirming SpaceX headlines, and a lower-beta REIT that may move little either way. I could not confirm the Friday close, a current quote or the exact market cap and 20-day dollar volume, so those fields are omitted. Market cap was about $36.6B in a pre-market snapshot on 2026-10-23, and average volume near 1.1M-1.2M shares at roughly $75-$90 a share is above the $20M floor.
   Return: -1.45%; alpha: 1.09%.

Full prediction and audit record

DeepSeek

View full picks and reasoning

US equities enter the 2026-10-26 session after a sharply divergent 2026-10-23 session in which earnings surprises and clinical-trial data drove extreme single-stock moves. With Treasury yields elevated and oil near five-month highs, index-level direction is uncertain, so idiosyncratic catalysts and mean reversion should dominate. I favor oversold losers for technical bounces and expect speculative gainers to give back some gains.

1. RHI up — confidence 70%. Ranked first because the 21.74% post-earnings jump is backed by substantive Q3 results and management's bullish Protiviti compliance commentary, which can support next-session drift despite an overbought RSI of 76. The catalyst is substantive (earnings and guidance). Reference price is the 2026-10-26 pre-market quote of $47.19, about three hours stale; market cap $4.61B and 20-day average dollar volume roughly $90.8M both exceed the $1B and $20M floors.
   Return: -5.34%; alpha: -5.70%.
2. CORT up — confidence 65%. Ranked second on a high-conviction technical reversal: the 20.13% selloff was driven by substantive negative competitive data from Genmab, but the stock is deeply oversold (RSI 24) and was rebounding about 5% in 2026-10-26 intraday trading. The catalyst is substantive, yet the magnitude of the drop looks overdone for a one-day horizon. Reference price is the 2026-10-23 close of $85.86 because no citable 2026-10-26 quote was available; market cap $9.28B and 20-day dollar volume about $71.8M pass the screens.
   Return: -5.06%; alpha: -5.41%.
3. LKQ up — confidence 62%. Ranked third as a reversal candidate: the 22.47% plunge followed substantive negative preliminary Q3 results, a guidance withdrawal, and a CEO departure, but RSI at 22 and a modestly higher 2026-10-26 pre-market quote suggest a short-term bounce. The catalyst is substantive, but the selling looks exhausted. Reference price is the 2026-10-26 pre-market quote of $17.40; market cap $4.41B and 20-day dollar volume about $72.8M pass the floors.
   Return: -1.62%; alpha: -1.97%.
4. CCI down — confidence 60%. Ranked fourth for a reversal down: the 9.06% gain was driven by speculative and unconfirmed reports of SpaceX tower lease talks, and RSI at 71.8 signals overbought conditions. The catalyst is speculative, so profit-taking is likely. Reference price is the 2026-10-23 after-hours quote of $86.67; market cap $36.58B and 20-day dollar volume about $482.3M are well above floors.
   Return: -1.45%; alpha: 1.09%.
5. CRWD up — confidence 58%. Ranked fifth for continuation up: the 8.25% rally on critical PHP vulnerabilities has a partly speculative catalyst, but rising breach activity and a neutral RSI of 55.9 suggest momentum can persist. The catalyst is speculative but supported by fundamental demand. Reference price is the 2026-10-23 after-hours quote of $258.40; market cap $264.3B and 20-day dollar volume about $2.14B are far above the thresholds.
   Return: 3.54%; alpha: 3.18%.

Full prediction and audit record

How models have done in Extreme Movers

ForecastNest's own settled Extreme Movers runs at this horizon, up to and including this one. Members' own research is not counted. Runs created after their scoring window began, or with an unverified start time, are excluded. The chart is the average directional return per run. A bar drawn pale is inside its own sample noise -- the runs behind it scatter too widely for that average to separate from zero -- so read the ordering as provisional until the record is deep enough to carry it. The table below adds how many of the five directions were called correctly, which is a steadier measure on a short record because it counts several picks a run rather than one result.

Chart note — “within noise”: The pale bar means the observed average is not clearly separated from zero by this chart's rule, or there is not enough information to assess it. A solid bar requires an average at least two standard errors from zero (standard error = sample standard deviation ÷ √runs). Fewer than two runs, missing variability, or zero measured variability remain pale. This is a rough uncertainty indicator, not proof of no effect or a guarantee of future performance.

Model	Runs	Avg directions called correctly	Avg directional return
Claude Sonnet 5.5	18	58.89%	+0.96%
Meta AI - Muse Spark	1	40.00%	+0.73%
Grok 4	4	55.00%	+0.68%
DeepSeek	29	48.45%	+0.42%
GPT-6 Thinking	1	20.00%	+0.12%
Muse Spark 1.1	27	47.41%	+0.12%
Claude Opus 5	6	60.00%	+0.03%
Grok 4.5	21	46.67%	-0.07%
Claude Opus 5.5	5	40.00%	-0.10%
Perplexity	20	44.00%	-0.35%
Gemini 1.5 Pro	18	48.89%	-0.52%
GPT-6	16	41.25%	-0.76%
Gemini 3 Pro	9	40.00%	-0.97%
Gemini 3.6 Flash High	1	40.00%	-2.13%
Grok 4.7	1	40.00%	-2.26%
Gemini 2.5 Pro	1	20.00%	-2.89%
GPT-5.6 Sol	10	46.00%	-3.06%

View settlement prices

Symbol	Start	End	Source	Notes
AWK	142.29000000	139.80000000	https://finance.yahoo.com/quote/AWK/history/
BE	198.03000000	220.21000000	https://finance.yahoo.com/quote/BE/history/
CCI	88.11500000	86.84000000	https://finance.yahoo.com/quote/CCI/history/
CORT	90.09000000	85.53510000	https://finance.yahoo.com/quote/CORT/history/
CRWD	255.85010000	264.89500000	https://finance.yahoo.com/quote/CRWD/history/
FICO	753.76500000	701.71000000	https://finance.yahoo.com/quote/FICO/history/
FIX	1225.57000000	1263.67000000	https://finance.yahoo.com/quote/FIX/history/
LKQ	17.64000000	17.35500000	https://finance.yahoo.com/quote/LKQ/history/
MRNA	201.77520000	184.70000000	https://finance.yahoo.com/quote/MRNA/history/
NUE	260.00000000	268.65000000	https://finance.yahoo.com/quote/NUE/history/
PANW	383.28000000	392.17000000	https://finance.yahoo.com/quote/PANW/history/
RHI	46.12500000	43.66000000	https://finance.yahoo.com/quote/RHI/history/
SLB	52.24000000	53.74000000	https://finance.yahoo.com/quote/SLB/history/
SPY	763.27000000	765.99000000	https://finance.yahoo.com/quote/SPY/history/
T	21.05000000	20.91500000	https://finance.yahoo.com/quote/T/history/

Standard prompt extreme-movers-v11, SHA-256 f1059277849fbbf2f9935e9567e65d04d0be2ef4c896d623a641b011a4672780. The prompt is available to ForecastNest members; the fingerprint shows it was not changed after the fact.

Calculation: extreme-movers-v1.