Oct 29, 2026 - 6 AI models call the biggest movers in the morning session. 3 models finished positive

Asked Oct 29, 9:52 AM EDT (morning session). Grok 4.7, Perplexity, Muse Spark 1.1, Claude Sonnet 5.5, DeepSeek and GPT-6 answered the same standard prompt independently. Window: from Oct 29, 9:52 AM EDT to Oct 30, 9:52 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
Grok 4.7	2/5	40.00%	-0.26%	-1.03%	-0.26%
Perplexity	1/5	20.00%	-2.07%	-1.03%	-2.07%
Muse Spark 1.1	2/5	40.00%	0.31%	-1.03%	0.31%
Claude Sonnet 5.5	3/5	60.00%	1.26%	-1.03%	1.26%
DeepSeek	3/5	60.00%	3.69%	-1.03%	3.69%
GPT-6	2/5	40.00%	-0.82%	-1.03%	-0.82%

Grok 4.7

View full picks and reasoning

Markets closed mixed on October 28 with modest gains in major indices amid earnings reports. Extreme movers were driven by earnings reactions in healthcare, industrials, and tech, with several large-cap names showing high-volume moves that meet liquidity filters. Outlook for the next session leans toward selective continuation of earnings-driven extremes rather than broad mean reversion, given the fundamental nature of most catalysts and elevated volume.

1. AXGN up — confidence 68%. AXGN ranked among top percentage gainers on the October 28 session after reporting strong Q3 earnings before the open, with revenue and profit surprising positively. The move is substantive as it reflects updated business expectations for peripheral nerve repair products rather than pure sentiment. Price path showed a gap higher and strong close near the high on massive volume (about 5x average). Latest price 49.34 (regular-session, as of approximately 9:39 AM EDT on 2026-10-29, delayed/real-time mix from MarketWatch); prior close was 49.25. Partial session so far shows modest further gains. Catalyst appears only partially reflected; further upside supported by volume confirmation and beat. Strongest counter-evidence is the already extended 5-day run and RSI near overbought. Invalidation would be a close back below 45 or a material guidance cut. Avg dollar volume estimate exceeds 70M based on ~1.9M share average and recent prices.
   Return: -1.86%; alpha: -0.83%.
2. VRT up — confidence 72%. VRT was a top S&P 500 gainer on October 28 after strong Q3 earnings and raised full-year guidance tied to data-center and AI power demand. The catalyst is substantive and alters business expectations. Price action featured a large gap and close near session highs on roughly 3x average volume. Latest price approximately 238.90 (regular-session as of ~9:36 AM EDT 2026-10-29 from Yahoo Finance); prior close 242.94. Partial session shows modest pullback from the prior close. Much of the initial reaction is priced, but order strength and guidance raise support further upside. Strongest counter is the intraday fade so far today and prior volatility. Invalidation would be a break below 228 or a sector-wide AI hardware pullback. Market cap and volume easily clear filters.
   Return: -1.85%; alpha: -0.82%.
3. FFIV down — confidence 65%. FFIV was the largest S&P 500 loser on October 28 after fiscal Q4 results that beat but came with a weaker growth outlook and hardware demand concerns. The catalyst is substantive and revises near-term expectations downward. Price path showed a sharp gap down and close near the lows on elevated volume. Latest price around 362 (pre-market/early regular as of ~9:00 AM EDT 2026-10-29 from multiple sources); prior close 365.53. Partial session shows continued pressure. Reaction appears incomplete given the magnitude of the outlook cut. Strongest counter-evidence is the prior strong YTD performance and any potential oversold bounce. Invalidation would be a reclaim of 400 or positive peer guidance. Market cap and liquidity clear thresholds.
   Return: -3.29%; alpha: 4.32%.
4. ENPH down — confidence 63%. ENPH ranked among larger losers on October 28 after Q3 results that missed and included multiple price cuts plus cautious guidance amid weak residential solar demand. Catalyst is substantive and revises business expectations lower. Price action closed near the lows of a high-volume day that set a fresh 52-week low. Latest price approximately 24.50 (early regular session as of ~9:35 AM EDT 2026-10-29 from stockanalysis.com); prior close 24.99. Partial session shows further softness. Downside appears not fully priced given the guidance tone. Strongest counter is extreme oversold readings and high short interest that could fuel a squeeze. Invalidation would be a reclaim of 28 or positive policy/demand news. Liquidity and size filters are met.
   Return: 0.91%; alpha: 0.11%.
5. GMAB up — confidence 60%. GMAB was a top percentage gainer on October 28 amid positive pipeline and commercial updates for its antibody portfolio (including Darzalex franchise strength). The move reflects both sentiment and tangible progress in oncology assets. Price closed well off the lows after a wide range on very high volume. Latest price approximately 57.51 (early regular as of ~9:45 AM EDT 2026-10-29 from TradeSmith); prior close 57.29. Partial session is roughly flat to slightly higher. Catalyst appears partially absorbed; further upside possible on continued analyst upgrades. Strongest counter is the already large one-day move and valuation concerns. Invalidation would be a close below 50 or negative pipeline news. Size and liquidity clear filters.
   Return: 0.05%; alpha: 1.08%.

Full prediction and audit record

Perplexity

View full picks and reasoning

The source session (2026-10-28) favored AI-infrastructure and memory names after Vertiv's earnings and Oracle's investor-day targets, while earnings disappointments hit F5, Enphase, LPL Financial and Boeing. For the one-day horizon, I favor continuation in the fundamentally supported AI-infrastructure gainers and continuation in the earnings-driven decliners with deteriorating operating disclosures, while treating F5's extreme one-day loss as the best reversal candidate because its guidance-driven selloff appears largely priced in and several analysts raised targets.

1. VRT up — confidence 62%. Vertiv closed the source session up 16.48% at 242.94 after reporting Q3 EPS of 2.21 versus a 1.82 estimate and guiding to roughly 38% growth; this is a substantive catalyst because orders and AI data-center demand expectations changed, not merely sentiment. The latest quote found was 235.21, down 3.18% in extended hours at 5:23:24 PM EDT on 2026-10-28 (extended-hours quote), so part of the earnings reaction has already faded; I still expect continuation higher into the next regular close because the fundamental guidance is new and the stock closed near the upper part of its 228.82-245.31 range on 19.47 million shares versus a reported 5.72 million average. Counter-evidence: the after-hours decline and the stock's proximity to its 52-week-high zone imply profit-taking risk. Invalidation: a next-session close below the source-session low of 228.82 would undermine the continuation thesis. Reference price is the latest extended-hours quote; the 20-day average dollar volume was not directly published, so the reported figure is the reported average share volume multiplied by the latest price and should be treated as an approximation.
   Return: -1.85%; alpha: -0.82%.
2. ORCL up — confidence 55%. Oracle rose 10.73% to 144.54 in the source session after investor-day disclosures including a fiscal-2030 revenue target of 225 billion dollars and cash-flow goals; this is a substantive strategic catalyst, although its execution and financing requirements remain uncertain. The latest quote found was 141.75, down 1.93% in overnight Blue Ocean trading at 1:09:02 AM EDT on 2026-10-29 (extended-hours quote), suggesting partial giveback but not a full reversal; I expect continuation higher over the remaining window as investors digest the longer-term targets. The session closed well above its 128.42 low and near the upper half of its 128.42-148.47 range on 81.5 million shares versus a reported 30.36 million average. Counter-evidence: the stock remains below its 52-week high of 281.62 and has substantial negative levered free cash flow, so skepticism about AI capex could reassert itself. Invalidation: a next-session close below the source-session low of 128.42 would undermine continuation. The 20-day average dollar volume was not directly published; the reported value is the reported average share volume multiplied by the latest price.
   Return: -4.30%; alpha: -3.27%.
3. FFIV up — confidence 54%. F5 fell 20.43% to 365.53 in the source session after Q4 results and guidance tied to a costly AI investment cycle; the latest quote found was 363.35, down 0.60% in overnight trading at 8:36:13 PM EDT on 2026-10-28 (extended-hours quote). Because the decline was driven by guidance and hardware-replacement concerns rather than a fraud or liquidity event, and because several analysts raised price targets after the print, I expect a partial reversal higher. The stock closed near the low of its 361.78-429.46 range, which supports a bounce attempt, while the huge volume relative to its reported 575,207 average confirms capitulation-style activity. Counter-evidence: a Piper Sandler downgrade and the close near the session low show genuine fundamental concern, so the rebound may fail. Invalidation: a next-session close below 361.78 would invalidate the reversal thesis. The 20-day average dollar volume was not directly published; the reported value is the reported average share volume multiplied by the latest price.
   Return: -3.29%; alpha: -2.26%.
4. ENPH down — confidence 58%. Enphase fell about 16% in the source session after a Q3 miss and disappointing Q4 outlook, with reporting citing price cuts and weak residential-solar demand; this is a substantive negative catalyst because it lowers near-term revenue and earnings expectations. The latest quote found was 24.76, down 16.80% as of 3:35:19 PM EDT with the market open (intraday quote; timestamp indicates the quote was captured later than the nominal answering time), and the stock closed near its 23.83 session low. I expect continuation lower because the guidance deterioration is fundamental and the stock made a new 52-week low; a dead-cat bounce is possible, but the earnings call gave no clear positive offset. Counter-evidence: the extreme decline and a market capitalization near 3.26 billion dollars could attract value buyers, and the average price target of 49.13 implies substantial perceived upside. Invalidation: a next-session close above the source-session open of 23.89 would undermine the continuation thesis. The 20-day average dollar volume was not directly published; the reported value is the reported average share volume multiplied by the latest price.
   Return: 0.91%; alpha: 0.11%.
5. LPLA down — confidence 52%. LPL Financial fell 13.58% in the source session after disclosing that Commonwealth asset retention slipped to 83% and recruited assets dropped 26%; this is a substantive negative catalyst because it directly reduces expected advisory-asset growth. The latest quote found was 271.37, down 14.55% as of 1:38:52 PM EDT with the market open (intraday quote; timestamp indicates the quote was captured later than the nominal answering time), and the stock closed near its 270.14 session low after opening at 294.17. I expect continuation lower because the retention and recruiting deterioration is operational rather than purely sentimental, and multiple brokers cut price targets. Counter-evidence: Q3 EPS of 5.97 beat the 5.69 estimate, and the low beta of 0.54 plus a 21.37 billion dollar market cap suggest the selloff could stabilize quickly. Invalidation: a next-session close above the source-session open of 294.17 would undermine the continuation thesis. The 20-day average dollar volume was not directly published; the reported value is the reported average share volume multiplied by the latest price.
   Return: -0.01%; alpha: 1.04%.

Full prediction and audit record

Muse Spark 1.1

View full picks and reasoning

S&P 500 closed +0.28% on Oct 28 after Fed held 3.75-4.00%, but single-stock dispersion was extreme on earnings and investor days. AI infrastructure names like Vertiv and 【entity-Oracle¦canonical_name=Oracle】 surged on beats and long-term targets, while F5 and CoStar collapsed on guidance cuts and spend concerns. Breadth was 249 up vs 254 down in S&P 500, so stock-specific catalysts dominated index. Expect continuation of substantive guidance changes and potential mean-reversion in oversold quality names.

1. VRT up — confidence 70%. Rank 1 gainer Oct 28 in S&P 500, +16.48% to $242.94 close, top of S&P movers list. Latest quote $242.46 at 2026-10-29T09:30:00-04:00 from finnhub.io, delayed regular-session, 0.2% below prior close, indicating intraday pause but still near highs. Reference price is intraday delayed Oct 29, not yesterday close, freshness ~4 hours before prediction window end. Catalyst substantive: Q3 revenue $4.31B vs $3.76B forecast, EPS $2.21 vs $1.82, operating cash +64%, free cash +50%, raised 2026 revenue midpoint to $14.65B and EPS to $7.15, AI factory demand. Price closed near high $245.31 with volume 19.54M vs avg 6.46M, supporting continuation. Counter-evidence: +14.5% in 5 days, trading -36% below 52-week high $379.94, RSI extended. Invalidation would be close below Oct 28 low $228.81 or below prior close $208.56.
   Return: -1.85%; alpha: -0.82%.
2. FFIV down — confidence 72%. Rank 1 loser Oct 28 in S&P 500, -20.43% to $365.53, worst S&P performer. Latest quote $362.49 at 2026-10-29T09:45:00-04:00 from finnhub.io, delayed regular-session, down -0.83% vs prior close, showing no bounce yet. Reference price is live-delayed Oct 29 intraday, freshness under 1 hour. Catalyst substantive: fiscal 2027 revenue growth guide 8-9% vs 11% referenced, despite Q4 EPS $4.42 vs $4.26 consensus and $1.5B buyback announced Oct 27. Guidance cut changes valuation multiple. Price gapped down from $459.36, range $361.78-$429.46, close near low, volume 3.56M vs average ~0.6M, supports continuation down. Counter-evidence: buyback authorization and oversold bounce potential after -20% one-day. Invalidation: close above intraday high $429.46 Oct 28.
   Return: -3.29%; alpha: 4.32%.
3. ORCL up — confidence 65%. Rank 3 gainer Oct 28 S&P, +10.73% to $144.54 close, per heygotrade. Latest quote $139.55 at 2026-10-29T09:30:00-04:00 from finnhub.io, delayed regular-session, down -3.45% vs prior close, showing partial fade but still above $130 prior level. Reference price is delayed Oct 29 regular session, freshness ~4 hours. Catalyst substantive: investor day set $225B fiscal 2030 revenue goal, $30B+ free cash flow target FY29, return to positive FCF H2 FY28, cap new raises to $20B ($10B stock + $10B debt) through FY30, excluding refinancing. Changes funding risk and long-term expectations. Price broke above 50-day SMA $142.08, volume 81.9M vs 20d avg ~26M, 2.7x ATR. Counter: stock down -28% YTD despite bounce, still below $287 high. Invalidation: close below $130.53 Oct 27 close.
   Return: -4.30%; alpha: -3.27%.
4. FLEX up — confidence 60%. Rank 2 S&P gainer Oct 28, +11.61% to $93.03 close per stockanalysis.com. Latest quote $89.905 at 2026-10-29T09:35:00-04:00 from finnhub.io, delayed regular-session, down -3.36% vs prior close, showing gap fill but still above breakout. Reference price delayed Oct 29. Catalyst substantive: Q3 earnings before open Oct 28, revenue up 20.6% YoY Jun 2026, data center and AI infrastructure exposure, supply chain leverage. Price gapped from $83.35 to $93.03 on 11.5M shares vs avg 5.19M, closing near high. Counter: 1-month return -21.02% before bounce, indicates prior downtrend. Invalidation: close below $83.35 Oct 27 close.
   Return: -1.44%; alpha: -0.42%.
5. CSGP up — confidence 55%. Rank 2 loser Oct 28 S&P, -13.96% to $27.54 close, per marketbeat. Latest quote $27.785 at 2026-10-29T09:40:00-04:00 from finnhub.io, delayed regular-session, up +0.89% vs prior close, early sign of reversal. Reference price delayed Oct 29, freshness ~4 hours. Catalyst mixed but substantive: Q3 earnings Oct 27 after close EPS $0.38 vs $0.34 consensus beat, but revenue miss and Vexcel acquisition $1.2B and Homes.com spend raised concerns, leading to 14% drop near 52-week low $25.89. Previous change -13.96% was on 18.3M shares vs avg 4.9M, indicating capitulation. Hypothesis reversal up expects oversold bounce after beat. Counter: revenue miss and acquisition dilution may continue to pressure. Invalidation: close below $26.86 Oct 28 low.
   Return: 5.83%; alpha: 6.86%.

Full prediction and audit record

Claude Sonnet 5.5

View full picks and reasoning

Source session was Wednesday 2026-10-28. The Fed held rates at 3.75% to 4% with two dissents for a hike, and indexes finished slightly higher. The big moves were single-stock earnings reactions. After the close, Alphabet (-10% to -11%), ServiceNow (about -18%), Fortinet (about -26%) and Sprouts (about -23%) sold off sharply in premarket trading, which pressures software and security names and leaves a risk-off tone for growth. I could not retrieve post-open quotes for any pick, so every price below is a prior-session or intraday-prior-session figure with its date stated. Average dollar volume and 20-session OHLC statistics could not be verified for any pick, so those fields are omitted or null. Confidence is modest across the board.

1. VRT up — confidence 55%. Source-session previous_change_percent is +16.48% (Cappnotes close; StockTitan showed +17.3% at 3:01 pm ET), so the call is continuation with direction up. Catalyst (facts): Q3 sales +61%, organic orders more than doubled, full-year sales forecast raised by $650 million, and customers prepaid $4.1 billion ahead of delivery. This changes business expectations and is not purely sentiment. Inference: post-earnings drift and Alphabet's raised 2027 capex language are supportive for data-center power and cooling suppliers. Price path: close was about $208.56 on Oct 27 and the stock traded at $244.56 at 3:01 pm ET on Oct 28 (StockTitan); the daily history and 20-session range could not be retrieved, so I cannot say how extended it is. Latest_price 244.56, quote_time 2026-10-28T19:01:00-04:00 reported as 3:01 pm ET (2026-10-28T19:01:00Z equivalent is not claimed; timezone-exact time unverified), source StockTitan, regular-session intraday quote and stale by about one day. No post-open Oct 29 quote was found, so partial-session figures are null. Event-to-quote reaction is about +17.3% versus the Oct 27 close of about $208.56. Much of the news appears already priced in, and the expected move after the current quote is a modest further drift. Strongest counter-evidence: a 16% to 17% one-day jump in a name that was around $223 on Oct 14 (down 6% that day), plus a risk-off premarket in Nasdaq futures after Alphabet. Invalidation: a close back below the Oct 27 close of about $208.56, or an AI-capex sentiment break. Market cap is derived (about 385 million shares times $244.56), not a quoted figure. Average dollar volume could not be verified; the stock is a large cap with heavy trading, but that is not a verified number.
   Return: -1.85%; alpha: -0.82%.
2. AXGN up — confidence 50%. Source-session previous_change_percent is +31.47% (regular-session gain, StockAnalysis and Cappnotes), so the call is continuation with direction up. Catalyst (facts): Q3 revenue grew 26.2% and the full-year revenue forecast was raised to at least $290 million. That is substantive, not speculative. Inference: post-earnings drift in a small-cap medtech with a clean beat-and-raise. Price: last verified price is $49.25 on 2026-10-28 (StockAnalysis gainers table, labelled as updated Oct 28; regular-session close assumed, exact timestamp unverified). Market cap $2.89 billion is verified from the same table, and volume that day was about 9.28 million shares. The pre-event price path was around $36 on Oct 20 (ChartRow) and about $36 in October monthly data, so the stock gained roughly 35% from the Oct 20 level into this print; that is a stale, non-event-specific reference and I report the exact event-to-quote reaction as null. No Oct 29 quote was found; latest_price 49.25 is a prior-session regular close, stale by about one day. Strongest counter-evidence: a 31% one-day gain with a high short-volume share historically and a recent offering history (a $124 million offering in January 2026), which can invite supply into strength. Invalidation: a close below the pre-earnings level near $37.5 (the Oct 28 close divided by 1.3147), which would signal the gap fully failing. Average dollar volume was not verified; the 30-day average volume of about 1.1 million shares seen in August would imply roughly $40 million to $50 million, but that is a different window, so the field is omitted.
   Return: -1.86%; alpha: -0.83%.
3. FFIV down — confidence 45%. Source-session previous_change_percent is about -20.4% (StockTitan regular-session close; Chartmill showed 365.28, -94.08, -20.48%), so the call is continuation with direction down. Catalyst (facts): fiscal 2027 revenue growth guided to 8% to 9% after 11% in fiscal 2026, and Piper Sandler downgraded to Neutral. Inference: this is a growth-expectation reset, and the Oct 29 premarket sell-off in Fortinet (about -26%) with Palo Alto, CrowdStrike and Zscaler down more than 5% to 6% pressures the whole security group, which argues for further weakness. Price path: the stock was about $466.65 on Oct 8 with a 52-week high of $472.25, then closed about $365.28 on Oct 28, so it is still roughly 55% above its November 2025 low of $223.76 and far from a washed-out level. Latest_price 365.28, prior regular-session close for 2026-10-28, stale by about one day; no live Oct 29 quote was found, so freshness of any premarket move cannot be verified. Event-to-quote reaction versus the Oct 27 close is -20.4%. Strongest counter-evidence: StockTitan noted the release gives no explicit reason for a lower share price, the revenue guide is still positive growth, and a violent one-day reset often bounces. Invalidation: a reclaim of the Oct 27 close of about $459 would undermine the thesis; a more practical signal is an intraday recovery above $365.28 on rising volume. Market cap of about $20.7 billion is derived (26.43 billion at $466.65 scaled to $365.28), not quoted. Average dollar volume could not be verified; Robinhood showed average volume of about 508,000 shares, which at these prices is far above $20 million but is a different window.
   Return: -3.29%; alpha: 4.32%.
4. ENPH up — confidence 40%. Source-session previous_change_percent is about -17.1% (StockTitan regular-session figure at 3:01 pm ET; final close figure not verified), so the call is reversal with direction up. Catalyst (facts): Q3 revenue of $293.3 million, Q4 revenue guide of $225 million to $255 million, and the company cited soft residential solar demand, higher interest rates and the expiry of the Section 25D tax credit. This is a substantive fundamental downgrade, not mere sentiment, which is the main reason confidence is low. Inference: a short-covering or oversold bounce is plausible because short interest was about 17.5% of float with about 4.6 days to cover (equibles, August 2026 data, stale), but an oversold decline alone does not establish a rebound. Price path: October 2026 traded roughly $29 to $34 (about $29.04 on Oct 14 after a 9.35% drop), and I could not verify the Oct 28 close or an Oct 29 quote, so reference_price and source quote are omitted and latest_price is null with freshness unverifiable. Pre-event price, event-to-quote reaction and partial-session data are null. Market cap was not verified; from 131.8 million shares outstanding it would be well above $1 billion at any plausible price, but I report no figure. Average dollar volume was not verified; the 30-day average volume of about 4.2 million shares seen in August suggests it clears $20 million, but that is not a verified 20-day dollar figure. Strongest counter-evidence: guidance below prior-quarter revenue, a stated policy and rate headwind, and a risk-off tape. Invalidation: a break to new lows on heavy volume after the open, or a further guide-down headline.
   Return: 0.91%; alpha: 1.94%.
5. CSGP up — confidence 40%. Source-session previous_change_percent is about -15.4% (StockTitan regular-session figure at 3:01 pm ET; final close figure not verified), so the call is reversal with direction up. Catalyst (facts): Q3 revenue rose 14% to $952 million and full-year guidance was raised, yet the shares fell, and StockTitan notes the release gives no reason for a lower price. Inference: the sell-off likely reflects margin or Homes.com investment concerns, which have driven repeated declines in 2026, so this is a sentiment-and-profitability debate rather than a collapse in demand. An oversold decline alone does not establish a rebound, and the stock has been in a long downtrend (roughly $68 in October 2025 to the high $20s to low $30s in mid-2026). I could not verify the Oct 28 close, an Oct 29 quote, market cap or average dollar volume, so reference_price, market_cap_usd and avg_dollar_volume_20d_usd are omitted; for scale, market cap was about $12 billion to $13 billion at roughly $29 to $32 per share in mid-2026 data, which would likely still clear the $1 billion floor, but the $20 million dollar-volume floor and the $5 price floor are only assumed, not confirmed. Latest_price and quote_time are null and freshness cannot be verified. Strongest counter-evidence: a persistent downtrend, activist and Homes.com spending controversy, and a risk-off premarket tape in software and information services. Invalidation: a break to new 52-week lows on heavy volume after the open.
   Return: 5.83%; alpha: 6.86%.

Full prediction and audit record

DeepSeek

View full picks and reasoning

US equities enter the 2026-10-29 session with a clear AI-infrastructure risk-off tone. On 2026-10-28, AI and memory names rallied, but after Microsoft and Alphabet reported capital spending below consensus, neoclouds and AI power names reversed lower. At the same time, F5 and Enphase suffered severe negative earnings and guidance reactions, while Old Dominion fell despite a headline beat on cost and margin concerns. My selections are all downside calls: two reversal shorts on source-session AI gainers and three continuation shorts on source-session earnings losers. The main market-wide risk is a broad rebound if investors decide the capex shortfall reflects timing rather than weaker demand.

1. CRWV down — confidence 76%. CoreWeave was one of the largest source-session gainers, closing at $63.93 on 2026-10-28, up 7.32%, on 61.24M shares versus a 20-day average of 57.59M. The move was a rebound from a 52-week low of $51.24 and was not company-specific. After the close, Microsoft reported quarterly capex and finance leases of $45.2B versus the $50.7B consensus, and Alphabet also missed capex estimates. Microsoft is a major CoreWeave customer, and the shortfall raises questions about the timing of third-party data-center capacity deliveries (https://www.tipranks.com/news/why-are-coreweave-nebius-and-iren-stocks-tumbling-in-pre-market-today-oct-29). This is a substantive negative read-through for neocloud revenue timing, not just sentiment. The observed next-session reaction is already negative: the latest regular-session quote is $61.57 at 2026-10-29T09:59:00-04:00, down 3.69% from the prior close, according to Yahoo Finance live intraday data while the session remains open. The source-session close was near the day's high, so the reversal gap is meaningful. I expect the negative catalyst to drive a lower close. The main risk is a fast squeeze because the stock had just bounced from a 52-week low. Invalidation would be a close above $63.93.
   Return: -9.41%; alpha: 10.44%.
2. NBIS down — confidence 73%. Nebius closed at $182.34 on 2026-10-28, up 6.96%, on 24.75M shares. It participated in the same AI-infrastructure rebound. After the close, Microsoft's $45.2B capex versus $50.7B consensus and Alphabet's capex miss hit neoclouds. Nebius has a Microsoft computing-capacity agreement worth nearly $20B through the 2030s, so delayed third-party capacity is a direct revenue-timing risk (https://www.tipranks.com/news/why-are-coreweave-nebius-and-iren-stocks-tumbling-in-pre-market-today-oct-29). The latest quote is $172.28 at 2026-10-29T09:58:47-04:00, down 5.52% from the prior close, based on Yahoo Finance live intraday data while the session is open. The stock closed the source session at its high, making the after-hours and premarket reversal a sharp sentiment shift. I expect further weakness as the market reprices neocloud capex timing. The main counter-risk is that Microsoft still expects capex above $50B next quarter and second-half acceleration, which could produce a bounce. Invalidation would be a close above $182.34.
   Return: -7.30%; alpha: 8.33%.
3. FFIV down — confidence 69%. F5 was the clearest negative earnings reaction in the source session, closing at $365.53 on 2026-10-28, down 20.43%, on 3.56M shares versus a 20-day average of 742,550. The company beat fiscal Q4 EPS and revenue, but guided fiscal Q1 2027 EPS midpoint to $4.31 versus $4.35 consensus and revenue midpoint to $885M versus $886.4M. Fiscal 2027 revenue growth is guided to 8-9% from 11% in fiscal 2026, higher component costs pressure gross margin, and Piper Sandler downgraded the stock to Neutral (https://markets.businessinsider.com/news/stocks/f5-stock-ffiv-dives-nearly-20-on-weak-guidance-and-an-analyst-downgrade-1036680049). Those are substantive guidance issues. The stock closed near its low and below the 20-day range. The latest quote is $358.72 at 2026-10-29T09:47:36-04:00, down 1.86% from the prior close, according to Yahoo Finance live intraday data. Most of the initial drop is already reflected, but post-earnings drift and the downgrade favor further weakness. Counter-evidence includes the fiscal Q4 beat and a fiscal 2027 EPS midpoint slightly above consensus. Invalidation would be a close above $365.53.
   Return: -3.29%; alpha: 4.32%.
4. ENPH down — confidence 65%. Enphase closed at $24.99 on 2026-10-28, down 16.03%, on 26.61M shares versus a 20-day average of 6.35M. Q3 revenue was $293.3M versus $303.85M expected, adjusted EPS was $0.41 versus $0.51, and Q4 revenue guidance of $225M-$255M was far below the roughly $300.5M consensus. U.S. product sell-through fell 35% year over year (https://sg.finance.yahoo.com/news/enphase-shares-tumble-quarterly-earnings-164736885.html). This is a substantive demand and guidance deterioration, not just sentiment. The stock closed near its low and at a new 20-day low. The latest quote is $24.82 at 2026-10-29T09:59:17-04:00, down 0.68% from the prior close, according to Yahoo Finance live intraday data. The modest further decline suggests the market is still absorbing the guide-down. Counter-evidence is that the stock is already near its 52-week low, the Q4 guide could prove conservative, and AI data-center power is optionality but not near-term revenue. Invalidation would be a close above $24.99.
   Return: 0.91%; alpha: 0.11%.
5. ODFL down — confidence 60%. Old Dominion closed at $173.32 on 2026-10-28, down 7.33%, on 4.86M shares versus a 1.87M 20-day average, one of the largest NASDAQ declines. Q3 EPS of $1.65 and revenue of $1.59B beat consensus, but the stock fell as Q4 operating-ratio guidance of about 73.5% and higher cost inflation per shipment raised margin concerns, while tons per day were still down 0.1% (https://www.trefis.com/stock/odfl/articles/621364/what-is-behind-the-move-in-old-dominion-stock/2026-10-29). The reaction looks more valuation and cost-driven than a clean earnings miss, so the catalyst is mixed rather than purely substantive. The stock closed near its low and below its 20-day range. The latest quote is $168.98 at 2026-10-29T10:00:35-04:00, down 2.51% from the prior close, according to Yahoo Finance live intraday data. Continued selling supports a lower close, but the EPS and revenue beat and above-seasonal Q4 revenue guide are the main counter-evidence. Invalidation would be a close above $173.32.
   Return: 0.63%; alpha: 0.40%.

Full prediction and audit record

GPT-6

View full picks and reasoning

I favor selective continuation supported by earnings changes and completed price action, with modest confidence because substantial repricing has already occurred. October 28 is the latest completed source session. The explicit October 29–30 timestamp window governs evaluation despite the conflicting generic regular-close wording. Forecasts below concern additional movement after each current quote, not gains or losses already realized.

1. FFIV down — confidence 61%. Ranked first for the clearest combination of adverse forward expectations and a completed breakdown. Before earnings, the last five sessions showed recovery followed by a nearly flat October 27. The October 28 reaction then gapped down and closed near its low on 4.81 times average completed-session volume. F5's substantive release projects fiscal 2027 revenue growth of 8–9%, versus 13% in the latest quarter: https://www.f5.com/company/news/press-releases/earnings-q4-fy26. Much of that disappointment is already priced in; from the current quote I expect additional selling, not another repetition of the initial collapse. Strong quarterly growth and the additional $1.5 billion repurchase authorization are the strongest fundamental counter-evidence. A future close above $429.46 would undermine continuation. latest_price=362.00 USD; quote_time=2026-10-29T13:55:00Z; regular-session snapshot labeled real-time by the provider, approximately 20 minutes old at research cutoff; source=https://stockanalysis.com/stocks/ffiv/history/. Its displayed intraday high was inconsistent with the quote, reducing confidence.
   Return: -3.29%; alpha: 4.32%.
2. FLEX up — confidence 60%. Ranked second because a substantial earnings upgrade accompanies improving completed-session demand. Before the release, FLEX had recovered from October 23 weakness, including a strong October 27 advance. Its October 28 gap initially suffered a deep pullback but finished near the high on 2.31 times average completed-session volume. The company raised fiscal 2027 adjusted EPS guidance from $4.42–4.74 to $4.84–5.16: https://investors.flex.com/news/news-details/2026/FLEX-REPORTS-SECOND-QUARTER-FISCAL-2027-RESULTS/default.aspx. That is substantive, although considerable optimism is already reflected in the baseline-to-quote gain. From the current quote I expect renewed buying over the remaining window. Today's opening gap down and the price remaining below its earlier 20-session high are counter-evidence. A future close below $83.76 would undermine the thesis. latest_price=91.25 USD; quote_time=2026-10-29T14:03:06Z; regular-session snapshot, provider latency unverified, approximately 12 minutes old at research cutoff; source=https://www.google.com/finance/quote/FLEX:NASDAQ.
   Return: -1.44%; alpha: -0.42%.
3. LH up — confidence 59%. Ranked third for improving earnings supported by a rising pre-event price path, tempered by extension. Before the release, LH consolidated near $329–331 after an October 22 advance. The October 28 reaction gapped higher and established a new 20-session high on 2.58 times average completed-session volume, but left a substantial upper wick. Labcorp reported 6.9% revenue growth, 14.5% adjusted EPS growth and higher annual guidance: https://labcorp.gcs-web.com/news-releases/news-release-details/labcorp-reports-strong-2026-third-quarter-results-raises-full. This is substantive, and much of the benefit is already reflected in the price. Today's observed recovery from its early low supports a modest additional rise after the current quote. The rejected $371 high and initial opening decline are the strongest price counter-evidence. A future close below $341.40 would undermine continuation. latest_price=361.67 USD; quote_time=2026-10-29T14:06:00Z; regular-session snapshot labeled real-time, approximately nine minutes old at research cutoff; source=https://stockanalysis.com/stocks/lh/.
   Return: 1.71%; alpha: 2.74%.
4. CSGP down — confidence 57%. Ranked fourth because deteriorating bookings and completed selling pressure support continuation, while improving profits and today's rebound argue against it. Before the release, successive sessions already showed falling closes. The October 28 reaction gapped down, made a new 20-session low and closed near the bottom on 3.02 times average completed-session volume. Verified net new bookings were $37 million versus $84 million a year earlier, despite record adjusted EBITDA and increased annual EBITDA guidance. Sources: https://www.costargroup.com/press-room/2026/costar-group-reports-record-third-quarter-revenue-952-million-profitability-surges and https://investors.costargroup.com/news-releases/news-release-details/costar-group-q3-revenue-increases-20-year-over-year-net-new. Bookings weakness is substantive; my inference is that growth concerns will outweigh the profitability improvement after the current quote. Much of the adverse repricing has already occurred, and today's rebound is meaningful counter-evidence rather than future opportunity. A future close above $30.68 would undermine the bearish call. latest_price=27.93 USD; quote_time=2026-10-29T14:03:55Z; regular-session snapshot, provider latency unverified, approximately 11 minutes old at research cutoff; source=https://www.google.com/finance/quote/CSGP:NASDAQ.
   Return: 5.83%; alpha: -4.80%.
5. VRT up — confidence 56%. Ranked fifth because strong fundamental evidence competes with substantial repricing already completed. Before earnings, VRT had recovered for two sessions after October 23 weakness. The October 28 reaction opened far above the prior close, suffered a deep lower wick and recovered near its high on 2.96 times average completed-session volume. Vertiv reported 61% sales growth and raised annual guidance, a substantive improvement rather than sentiment alone: https://www.vertiv.com/en-us/about/news-and-events/home/. From the current quote I expect a modest additional advance as investors reassess earnings capacity. The stock remains below its 20-session high, and the large event-spanning gain leaves limited room for disappointment. A future close below $228.81 would undermine continuation. latest_price=243.90 USD; quote_time=2026-10-29T14:06:00Z; regular-session snapshot labeled real-time, approximately nine minutes old at research cutoff; source=https://stockanalysis.com/stocks/vrt/history/. The latest page's partial high was inconsistent with its quote, so I retain an earlier consistent partial snapshot and reduce confidence.
   Return: -1.85%; alpha: -0.82%.

Full prediction and audit record

How models have done in Extreme Movers + OHLC

ForecastNest's own settled Extreme Movers + OHLC 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
Grok	1	80.00%	+1.42%
Gemini 1.5 Pro	3	71.67%	+0.96%
Muse Spark 1.1	13	60.00%	+0.88%
GPT-6	13	56.92%	+0.84%
DeepSeek	12	48.33%	+0.13%
Perplexity	11	50.91%	+0.02%
Grok 4.7	5	48.00%	-0.41%
Gemini 3.6 Flash High	1	60.00%	-0.43%
Grok 4.5	4	35.00%	-0.54%
Claude Opus 5.5	3	60.00%	-0.69%
Gemini 3 Pro	4	45.00%	-0.72%
Claude Sonnet 5.5	5	36.00%	-0.88%
Grok 4	1	40.00%	-1.54%

View settlement prices

Symbol	Start	End	Source	Notes
AXGN	49.95000000	49.01990000	https://finance.yahoo.com/quote/AXGN/history/
CRWV	61.14500000	55.38990000	https://finance.yahoo.com/quote/CRWV/history/
CSGP	27.71510000	29.33000000	https://finance.yahoo.com/quote/CSGP/history/
ENPH	25.17000000	25.40000000	https://finance.yahoo.com/quote/ENPH/history/
FFIV	360.05500000	348.20000000	https://finance.yahoo.com/quote/FFIV/history/
FLEX	89.31000000	88.02000000	https://finance.yahoo.com/quote/FLEX/history/
GMAB	57.77000000	57.80000000	https://finance.yahoo.com/quote/GMAB/history/
LH	362.34000000	368.53500000	https://finance.yahoo.com/quote/LH/history/
LPLA	276.27000000	276.23000000	https://finance.yahoo.com/quote/LPLA/history/
NBIS	171.16500000	158.67000000	https://finance.yahoo.com/quote/NBIS/history/
ODFL	171.68000000	172.76000000	https://finance.yahoo.com/quote/ODFL/history/
ORCL	140.71500000	134.66500000	https://finance.yahoo.com/quote/ORCL/history/
SPY	768.14010000	760.24000000	https://finance.yahoo.com/quote/SPY/history/
VRT	241.43000000	236.97000000	https://finance.yahoo.com/quote/VRT/history/

Standard prompt extreme-movers-v11+ohlc3, SHA-256 8c7f67c19a4897cc85da742f2e7a48feefd535b792e9d02d07ecc4a798ed7c89. The prompt is ForecastNest's own research and is not published; the fingerprint shows it was not changed after the fact.

Calculation: extreme-movers-ohlc-v1.