Event Gaps: Claude Sonnet 5.5's picks for 2026-11-03
Claude Sonnet 5.5 picked OKTA, IDXX, AMD, ANET, DDOG for a 1-day horizon. Intraday alpha vs SPY: -0.61%. Picks with positive alpha: 60%.
| Experiment | Event Gaps |
|---|---|
| Model | Claude Sonnet 5.5 |
| Prompt version | event-gap-v13 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-11-03 |
| Measured to | 2026-11-04 |
| Benchmark | SPY |
| Status | Settled |
| Intraday alpha vs SPY | -0.61% |
| Picks with positive alpha | 60% |
| Benchmark return | +1.58% |
The picks
| Rank | Symbol | Gap direction | Price change | Return in the called direction | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|---|---|
| 1 | OKTA | +1.78% | +1.78% | 60% | +0.20% | Yes | |
| 2 | IDXX | +3.25% | +3.25% | 52% | +1.67% | Yes | |
| 3 | AMD | +0.72% | +0.72% | 50% | -0.86% | No | |
| 4 | ANET | +9.68% | +9.68% | 48% | +8.10% | Yes | |
| 5 | DDOG | -10.55% | -10.55% | 45% | -12.14% | No |
The model's market view
The window spans a high-variance stretch: US midterm results overnight, AMD, Arista and Super Micro reports after the close tonight, and Datadog before the open on Nov 4, with VIX above 21 and negative dealer gamma noted by one desk note. AI hardware and software were sold hard last week and Monday (Intel -29% on Oct 30, AMD -10.9% then -7.5%, Nasdaq -1.07% on Monday), and this morning shows a partial rebound. I lean long on three names where the bar was reset lower into a scheduled print (AMD, Arista, Datadog), on one mechanical flow event (Okta joining the S&P 500), and on one post-earnings drift candidate (IDEXX). All five calls are long, so the book carries market beta that SPY-relative scoring will partly offset. Earnings outcomes are close to coin flips with large implied moves, so confidences are modest. Estimates of expected move are below the options-implied moves because I only claim a small directional edge.
Why each pick
OKTA
Direction up, expected about 3% over the remaining window. Event (scheduled, official): S&P Dow Jones Indices announced on Oct 30 that Okta replaces Kenvue in the S&P 500 effective prior to the open on Wednesday Nov 4 (S&P press release); index funds that track the S&P 500 normally buy at the Nov 3 close, which falls inside this window (that timing is my inference of standard practice). It adds demand but does not change business expectations (flow/sentiment, not fundamentals). Price path before: closed Oct 30 at 208.24, then fell 11.22% on Monday to 184.87 on 5.9M shares (1.7x the 20-day average) amid a cybersecurity sell-off (Check Point -6.98% on a revenue miss, Fortinet -7.25%, Zscaler -5.90% per Cappnotes); attributing Okta's drop to sympathy selling is my inference. Observed reaction: the stock is down 9.7% from the Oct 30 pre-event close at the current quote, so the announcement has not produced a lasting gain; the announcement reaction itself is null because its timing was not verified. Latest quote: latest_price 188.02, quote_time 2026-11-03T10:28:37-05:00, live regular-session Nasdaq real-time via https://finance.yahoo.com/quote/OKTA/ (+1.70% vs Monday close). PARTIAL SESSION as of 10:28 ET: open 183.75, high_so_far 190.43, low_so_far 180.00, cumulative_volume 890,079 (relative intraday volume cannot be established). The quote is roughly at the window start, so the elapsed move is about zero. Expected move after the current quote: modest drift higher into the close from index demand, with a risk of fading after the effective date (the window ends 15:32Z Nov 4, about 1 hour after the effective open). Completed daily bars (US Eastern regular sessions, split-adjusted, source https://finance.yahoo.com/quote/OKTA/history/):
2026-10-27 O 219.23 H 228.00 L 219.00 C 226.07 V 2505900
2026-10-28 O 225.57 H 228.72 L 221.27 C 222.66 V 2312500
2026-10-29 O 209.44 H 222.35 L 204.11 C 221.86 V 3314500
2026-10-30 O 218.99 H 222.64 L 202.82 C 208.24 V 5258800
2026-11-02 O 205.00 H 206.00 L 183.12 C 184.87 V 5893600
return_5d_pct -15.97; high_20d 235.00; low_20d 183.12; avg_volume_20d 3,400,100. Price action: Monday closed near its low (a 20-day low) after a five-day decline, which can mark capitulation but does not by itself establish a rebound; the call rests on the index flow. Counter-evidence: S&P MidCap 400 funds must sell Okta, cyber/software sentiment is weak, and the stock fell 11% on Monday despite the already-announced inclusion, so some flow may be priced in. Invalidation: a trade below Monday's low of 183.12, or a flat/weak close on Nov 3 despite index buying. Market cap 32.87B (Yahoo, intraday); 20-day average dollar volume 709,563,925 is my computation (mean of close times volume over the 20 sessions ending 2026-11-02).
IDXX
Direction up, expected about 1.5%. Event (occurred): IDEXX reported Q3 results before the open on Nov 2: revenue up 10% to 1.21 billion dollars and a raised 2026 revenue and EPS outlook (StockTitan; Cappnotes gives the new EPS range as 14.96 to 15.16), with multiple analyst target increases (Yahoo). That changes earnings expectations, not just sentiment, which favors post-earnings drift. No new scheduled event falls inside the window, so this is a drift call, not an event-gap call, and I size it small. Price path: the stock was in a 500-590 range through October (Oct 30 close 551.21), gapped on Nov 2 to open 625.00 and closed 671.64 (+21.85%) near the day high of 673.19 on 1.98M shares (3.1x the 20-day average), a breakout to a 20-day high. Observed reaction: from the pre-event close of 551.21 to the current quote is +21.2%, so most of the repricing is done and I forecast only a modest further drift. Latest quote: latest_price 668.07, quote_time 2026-11-03T10:29:18-05:00, live regular-session real-time via https://finance.yahoo.com/quote/IDXX/ (about -0.5% vs Monday close). PARTIAL SESSION as of 10:28 ET: open 662.26, high_so_far 672.70, low_so_far 655.52, cumulative_volume 176,221 (relative volume cannot be established). A pre-market indication near -1.8% at 8:13 ET (Yahoo) recovered to about flat. The quote is about at the window start (the elapsed move is about zero). Completed daily bars (US Eastern regular sessions, split-adjusted, source https://finance.yahoo.com/quote/IDXX/history/):
2026-10-27 O 563.23 H 567.55 L 559.03 C 561.25 V 489000
2026-10-28 O 556.78 H 567.77 L 555.21 C 562.91 V 351800
2026-10-29 O 570.17 H 576.17 L 568.01 C 575.98 V 478000
2026-10-30 O 583.75 H 586.00 L 540.61 C 551.21 V 1173500
2026-11-02 O 625.00 H 673.19 L 620.20 C 671.64 V 1976100
return_5d_pct 17.79; high_20d 673.19; low_20d 501.17; avg_volume_20d 631,360. Price action supports continuation (gap-and-go, close at about 97% of the day's range, volume about 3x average), though today's flat tape shows no extension yet. Counter-evidence: a one-day gain of about 22% and about 30% over a month (Trefis) leave valuation stretched, and early-session volume is light. Invalidation: a move back below Monday's open of 625.00 (gap fill). Market cap 52.631B (Yahoo, intraday); 20-day average dollar volume 355,542,371 is my computation (mean of close times volume).
AMD
Direction up, expected about 4% over the window (the options-implied move was about 8.3%, per earnings-watcher as of Nov 2). Event (scheduled, official date): AMD reports Q3 after the close on Nov 3 with a call around 4:15 pm ET (Yahoo and a desk note); consensus is about 1.96 EPS and 13.20 billion revenue (Tickzen). Expectations facts: the stock fell 10.9% on Oct 30 (504.80 vs 566.42, my computation) after Intel's weak Q4 guide and 7.5% on Monday to 466.84, so the bar has been cut sharply; one desk note says it may need only a solid outlook to stabilize, and cites a recently announced supply commitment to Anthropic (not independently verified). Anticipation: the stock has already rebounded 4.66% today into the print. Observed reaction: the current quote is down 13.7% from the Oct 29 close of 566.42, which is a sector de-rating, not an event reaction (event-to-quote reaction is null since the report has not happened). Latest quote: latest_price 488.61, quote_time 2026-11-03T10:34:12-05:00, live regular-session real-time via https://finance.yahoo.com/quote/AMD/. PARTIAL SESSION as of 10:34 ET: open 480.03, high_so_far 497.35, low_so_far 473.54, cumulative_volume 8,441,121 (relative intraday volume cannot be established). The window started at 10:32 ET, so the elapsed move is negligible. Completed daily bars (US Eastern regular sessions, split-adjusted, source https://finance.yahoo.com/quote/AMD/history/):
2026-10-27 O 539.49 H 563.92 L 538.50 C 560.26 V 20599900
2026-10-28 O 573.08 H 579.18 L 560.52 C 578.44 V 18848600
2026-10-29 O 562.27 H 575.10 L 560.57 C 566.42 V 16873300
2026-10-30 O 508.53 H 520.40 L 492.21 C 504.80 V 39697500
2026-11-02 O 510.71 H 512.82 L 466.07 C 466.84 V 31990800
return_5d_pct -12.13; high_20d 658.52; low_20d 466.07; avg_volume_20d 21,422,410. Price action: two heavy-volume down days (1.5x to 1.9x the average) took the stock from 566 to 467 and to a 20-day low, so risk is already partly priced; oversold alone does not establish a rebound. Counter-evidence: Intel's collapse on weak guidance shows the market punishing soft Q4 outlooks, AI-hardware sentiment is fragile, and AMD fell more than 5% after its last two reports per a headline in my search results (not verified in detail). Invalidation: a post-report move that takes the stock below Monday's low of 466.07, or a Q4 guide below consensus. Market cap 797.644B (Yahoo, intraday); 20-day average dollar volume 11,922,850,600 is my computation (mean of close times volume).
ANET
Direction up, expected about 4% (options-implied about 10.5% per earnings-watcher as of Nov 2). Event (scheduled, officially confirmed): Arista reports Q3 after the US close on Nov 3 (company announcement, per therighttrader and StockAnalysis); the exact release time was not verified, so I use the 4:00 pm ET close as the event_time. Consensus is about 0.99 EPS and 3.3 billion revenue (therighttrader), and the company has beaten EPS in eight straight quarters (same source). The last reaction was +3.6% (Aug 5, ChartRow), but the November 2025 report produced -8.55% despite a beat (Barchart), so the guide matters more than the beat. Price path: the stock fell from a 217.41 October high to 164.14 on Monday (five-day return -7.39%, near the 20-day low of 163.25) with volume 1.2x the 20-day average; post-earnings drift since August is already negative (-7.8%, EarningsWhispers). Hence expectations look lower than in August; the claim is my inference. Latest quote: latest_price 169.69, quote_time 2026-11-03T10:35:25-05:00, live regular-session real-time via https://finance.yahoo.com/quote/ANET/ (+3.4% vs Monday). PARTIAL SESSION as of 10:34 ET: open 165.60, high_so_far 171.80, low_so_far 165.21, cumulative_volume 2,814,411 (relative volume cannot be established). The elapsed move since the window start is negligible. Event reaction: not available (report pending). Completed daily bars (US Eastern regular sessions, split-adjusted, source https://finance.yahoo.com/quote/ANET/history/):
2026-10-27 O 180.60 H 189.27 L 180.00 C 187.00 V 5653600
2026-10-28 O 190.17 H 191.11 L 184.77 C 186.72 V 5004800
2026-10-29 O 183.50 H 187.38 L 182.65 C 184.22 V 5367100
2026-10-30 O 173.57 H 174.99 L 166.86 C 168.86 V 8341100
2026-11-02 O 170.60 H 171.00 L 163.25 C 164.14 V 7386000
return_5d_pct -7.39; high_20d 217.41; low_20d 163.25; avg_volume_20d 6,204,290. Counter-evidence: the stock already bounced 3.4% into the print, AI-networking names are in a de-rating phase, and a strong beat with a cautious guide produced -8.55% last November. Invalidation: a post-report drop below Monday's low of 163.25. Market cap 214.017B (Yahoo, intraday); 20-day average dollar volume 1,165,842,202 is my computation (mean of close times volume).
DDOG
Direction up, expected about 5% (options-implied about 14.6% per earnings-watcher as of Nov 2). Event (scheduled, official): Datadog reports Q3 before the US open on Wednesday Nov 4, with a call at 8:00 am ET (Datadog announcement via BriefGlance; Yahoo lists the earnings date as Nov 4), which is inside the window. Anticipation: the stock fell 22.9% in five sessions (from 280.24 on Oct 26 to 216.10 on Monday, on 1.5x average volume) amid AI-spend and software worries, so the bar has been cut; this is my inference. Read-through risk: Dynatrace fell about 12% to 15% today after a beat because its profit and free-cash-flow guidance was trimmed (Arize acquisition dilution, a company-specific factor per Investing.com and Seeking Alpha), and Shopify and Gartner also fell on results; these can pressure software multiples into Datadog's print. Datadog's history: strong beats and large upside reactions (a 23.13% day-zero jump in November 2025, per Barchart), but softer usage from a major AI customer was flagged in Q2 (nemo.money summary), so concentration is a risk. Latest quote: latest_price 219.67, quote_time 2026-11-03T10:34:59-05:00, live regular-session real-time via https://finance.yahoo.com/quote/DDOG/ (+1.65% vs Monday close). PARTIAL SESSION as of 10:34 ET: open 213.00, high_so_far 222.43, low_so_far 211.74, cumulative_volume 1,336,854 (relative volume cannot be established). The elapsed move since the window start is negligible. Event reaction: not available (report pending). Completed daily bars (US Eastern regular sessions, split-adjusted, source https://finance.yahoo.com/quote/DDOG/history/):
2026-10-27 O 278.97 H 282.80 L 274.00 C 280.82 V 2734100
2026-10-28 O 279.84 H 285.00 L 275.49 C 278.41 V 2511700
2026-10-29 O 269.55 H 274.66 L 260.56 C 265.84 V 3730200
2026-10-30 O 253.62 H 253.87 L 228.57 C 235.27 V 5864200
2026-11-02 O 235.51 H 236.39 L 215.73 C 216.10 V 4647000
return_5d_pct -22.89; high_20d 300.68; low_20d 215.73; avg_volume_20d 3,023,875. Price action: a steep slide with high volume at 20-day lows can mean either capitulation or continued distribution, so I do not treat the oversold state alone as a signal. Counter-evidence: software sentiment is poor (Dynatrace, Shopify, Gartner reactions today), AI customer concentration, and the possibility that a beat is again met with selling. Invalidation: a new low below today's partial-session low of 211.74 before the report, or a post-report drop that takes it below Monday's low of 215.73. Market cap 78.876B (Yahoo, intraday); 20-day average dollar volume 796,495,117 is my computation (mean of close times volume).
Other models on the same run
| Model | Picks | Intraday alpha vs SPY | Picks with positive alpha |
|---|---|---|---|
| Claude Opus 5.5 | FN, FISV, PLTR, SHOP, AMD | -2.88% | 20% |
| GPT-6 | PLTR, AMD, SHOP, ANET, AAPL | -2.75% | 20% |
| GPT-6.1 Sol | FN, AMD, SHOP, PLTR, IT | -3.76% | 20% |
| Perplexity | AMD, ANET, SMCI, QCOM, UBER | +1.00% | 40% |
| Grok 4.7 | AMD, SHOP, SMCI, ALAB, ANET | -2.47% | 20% |
| DeepSeek | LSCC, IBTA, FN, OKTA, PLTR | -5.38% | 20% |
Written up on ForecastNest Insights: Event Gaps: Claude Sonnet 5.5's picks for 2026-11-03.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.