Asked Oct 29, 1:31 PM EDT (midday). GPT-6 answered the same standard prompt independently. Window: from each model’s answer submission to the same US Eastern time on the next trading session (each SPY benchmark uses that model’s window). Benchmark: SPY.
Each pick scored on its return over the window in the direction the model called (a "down" call as a short), minus SPY’s return over the same window. 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 beat SPY over the window.
| Model | Correct | Hit rate | Return | SPY | Alpha vs SPY |
|---|---|---|---|---|---|
| GPT-6 | 5/5 | 100.00% | 1.29% | -1.67% | 2.96% |
GPT-6
Scoring window: Oct 29, 1:43 PM EDT → Oct 30, 1:43 PM EDT.
View full picks and reasoning
I favor further weakness in stocks whose earnings outlooks and analyst revisions reinforce failed price recoveries, with Porch providing a positive earnings-revision exception. Four down calls create shared exposure to a technology rebound. Confidence measures evidence strength, not directional probability. Expected moves are prospective estimates for the actual ForecastNest save-to-next-session window, ending at the same Eastern wall-clock time on October 30, subject to the specified closing-time rule. The save timestamp is not yet known; prediction_time remains the batch identifier. Quotes are the latest timestamped observations found across the sources checked and may lag submission. Historical calculations use only the 20 completed regular sessions from October 1 through October 28. Average dollar volume is the arithmetic mean of daily split-adjusted Close multiplied by share volume; all five also exceed $20 million using the conservative lower bound of low_20d multiplied by avg_volume_20d. Market capitalizations are rounded provider figures. Earlier earnings and analyst-related moves are already observed context and are excluded from expected_move_pct.
- NOW — confidence 64%. Verified action: TD Cowen cut its target from $165 to $160 while retaining Buy, citing softer cRPO outperformance and fourth-quarter cRPO guidance approximately one percentage point below Street expectations. The publication date is verified, but the displayed intraday timezone was not established; published_at therefore uses the required date-only midnight placeholder, not an actual issuance time. Fundamental counterweight: ServiceNow reported subscription revenue growth and raised its full-year outlook: https://investor.servicenow.com/news/news-details/2026/ServiceNow-Reports-Third-Quarter-2026-Financial-Results/default.aspx. Further estimate pressure is my inference, not company guidance.
Before earnings and today's analyst action, the October 23 rally gave way to three lower closes. October 28's rebound toward $155.15 failed, leaving its close only 27.56% above the day's low on 2.32 times average completed-session volume. Today's gap below the completed 20-session low and trading near the partial-session low support continuation. The strongest counter-evidence is the enormous decline already absorbed despite continued business growth and TD Cowen's retained Buy rating; a sharp rebound is plausible. I expect an additional 3.5% decline over the scored window, not a repeat of today's loss. A subsequent regular-session close above the verified $121.50 partial-session high would invalidate the continuation thesis.
latest_price=110.46 USD; quote_time=2026-10-29T13:27-04:00; quote_type=regular-session snapshot, vendor labeled real-time; source=https://stockanalysis.com/stocks/now/history/. Pre-earnings price=150.13 at the October 28 regular close, 2026-10-28T16:00-04:00, from the same history source; pre_earnings_to_quote_return_pct=-26.423766. This combines earnings, revisions and market effects. pre_action_price=null; action_to_quote_return_pct=null because no immediately preceding analyst-note quote was verified.
PARTIAL SESSION as_of=2026-10-29T13:27-04:00: open=120.91, high_so_far=121.50, low_so_far=110.34, cumulative_volume=49254804 shares; source=https://stockanalysis.com/stocks/now/history/. A same-time volume baseline was unavailable, so relative intraday volume cannot be established. Rounded market-cap source=https://stockanalysis.com/stocks/now/.
Completed history uses US Eastern regular sessions and split-adjusted OHLC Close, not dividend-adjusted Adj Close. Source=https://stockanalysis.com/stocks/now/history/.
daily_5d:
session_date=2026-10-22 open=154.25 high=157.60 low=153.57 close=157.00 volume=17406898
session_date=2026-10-23 open=156.38 high=159.50 low=153.72 close=158.96 volume=11100759
session_date=2026-10-26 open=158.55 high=159.30 low=155.00 close=158.18 volume=14739149
session_date=2026-10-27 open=156.10 high=157.13 low=152.88 close=154.78 volume=13481100
session_date=2026-10-28 open=151.60 high=155.15 low=148.22 close=150.13 volume=28786305
Sixth close: 2026-10-21=153.12; return_5d_pct=-1.952717; high_20d=159.50; low_20d=133.47; avg_volume_20d=12414211.4 shares. The 20-session statistics exclude October 29.
TD Cowen cut its target to Buy, target $160 (from $165) on 2026-10-29 — source
Return: -5.57%; alpha: 7.25%. - GOOGL — confidence 62%. Verified action: TD Cowen lowered Alphabet's target from $485 to $460 and maintained Buy. It attributed the revision to Search and YouTube revenue misses and trimmed longer-term estimates; it also reported Cloud revenue above consensus, providing substantial counter-evidence. These are attributed analyst assessments, not independently interchangeable consensus measures. The publication date is verified, but its displayed intraday timezone was not established; published_at is the required date-only midnight placeholder rather than an actual issuance time.
Before earnings and the revision, October 22-26 showed persistent weakness: the October 23 and 26 closes were near their lows, followed by a two-session recovery. October 28 closed 59.47% up its range on 2.47 times average completed-session volume, showing meaningful buying before the report. Today's gap below the completed $316.60 range floor overturns that recovery. Trading below today's open after a bounce from $280.36 supports my inference of continuing selling pressure. I expect a further 2.3% decline over the scored window. The strongest counter-evidence is that the large loss already discounts considerable disappointment while Cloud remains strong; the intraday bounce could develop into a reversal. A subsequent regular-session close above the verified $288.33 partial-session high would invalidate this continuation call.
latest_price=283.60 USD; quote_time=2026-10-29T13:21-04:00; quote_type=regular-session snapshot, vendor labeled real-time; source=https://stockanalysis.com/stocks/googl/history/. Pre-earnings price=324.89 at the October 28 regular close, 2026-10-28T16:00-04:00, from the same history source; pre_earnings_to_quote_return_pct=-12.708917. This is the combined earnings, analyst and market-period change, not an isolated analyst effect. pre_action_price=null; action_to_quote_return_pct=null because an immediately preceding analyst-note quote was unavailable.
PARTIAL SESSION as_of=2026-10-29T13:21-04:00: open=286.88, high_so_far=288.33, low_so_far=280.36, cumulative_volume=88688096 shares; source=https://stockanalysis.com/stocks/googl/history/. Relative intraday volume cannot be established without a verified same-time baseline. Rounded issuer market-cap source=https://stockanalysis.com/stocks/googl/.
Completed history uses US Eastern regular sessions and split-adjusted OHLC Close, not dividend-adjusted Adj Close. Source=https://stockanalysis.com/stocks/googl/history/.
daily_5d:
session_date=2026-10-22 open=330.46 high=331.43 low=322.00 close=324.22 volume=27748122
session_date=2026-10-23 open=321.15 high=324.53 low=320.18 close=320.21 volume=17569640
session_date=2026-10-26 open=320.89 high=321.48 low=316.60 close=316.80 volume=23991837
session_date=2026-10-27 open=317.00 high=323.88 low=316.86 close=320.58 volume=28633058
session_date=2026-10-28 open=323.77 high=326.41 low=322.66 close=324.89 volume=63109029
Sixth close: 2026-10-21=334.76; return_5d_pct=-2.948381; high_20d=359.40; low_20d=316.60; avg_volume_20d=25585349.15 shares. The 20-session statistics exclude October 29.
TD Cowen cut its target to Buy, target $460 (from $485) on 2026-10-29 — source
Return: -1.48%; alpha: 3.15%. - PRCH — confidence 61%. Verified action: StoneX raised its target from $23 to $30 while maintaining Buy following quarterly results. The publication date is verified, but its displayed intraday timezone was not established; published_at uses the required date-only midnight placeholder, not an actual issuance time. The fundamental improvement is more than sentiment: Porch reported owner-business revenue excluding the reciprocal of $140.8 million, up 22%, and adjusted EBITDA excluding the reciprocal of $44.5 million, up 116%, and raised annual adjusted EBITDA guidance to $135-140 million from $119-125 million. These measures must not be confused with consolidated revenue. Primary source=https://ir.porchgroup.com/investors/news/news-details/2026/Porch-Group-Reports-Third-Quarter-2026-Results/default.aspx.
Before earnings and the revision, the stock declined sharply: October 22-26 closes weakened, and October 28's attempted rise to $18.12 reversed to a close only 8.08% above its low on 1.97 times average completed-session volume. Today's gap and continued trading above the former $18.75 completed-range ceiling indicate that the new information has reversed that earlier distribution. My inference is that the earnings reset can attract additional buying, supporting a further 3% gain over the scored window. The strongest counter-evidence is the roughly 25% advance already recorded and retreat from $21.30; these show that substantial good news is priced in and profit-taking is active. A subsequent regular-session close below the verified former breakout level of $18.75 would invalidate the bullish thesis.
latest_price=20.42 USD; quote_time=2026-10-29T13:19:39-04:00; quote_type=regular-session timestamped web snapshot; source=https://www.google.com/finance/quote/PRCH:NASDAQ. Pre-earnings price=16.30 at the October 28 regular close, 2026-10-28T16:00-04:00; source=https://stockanalysis.com/stocks/prch/history/; pre_earnings_to_quote_return_pct=25.276074. This combines earnings, analyst and market effects. pre_action_price=null; action_to_quote_return_pct=null because no immediately preceding StoneX-note quote was verified.
PARTIAL SESSION as_of=2026-10-29T13:15-04:00: open=19.46, high_so_far=21.30, low_so_far=18.74, cumulative_volume=4467229 shares; source=https://stockanalysis.com/stocks/prch/. This partial snapshot precedes the latest quote. Relative intraday volume cannot be established without a verified same-time baseline. The quote source reports rounded market capitalization of $2.32 billion.
Completed history uses US Eastern regular sessions and split-adjusted OHLC Close, not dividend-adjusted Adj Close. Source=https://stockanalysis.com/stocks/prch/history/.
daily_5d:
session_date=2026-10-22 open=18.43 high=18.69 low=17.90 close=17.94 volume=1684630
session_date=2026-10-23 open=17.89 high=18.29 low=17.65 close=17.73 volume=1216999
session_date=2026-10-26 open=17.86 high=17.96 low=17.27 close=17.37 volume=977103
session_date=2026-10-27 open=17.49 high=17.72 low=17.12 close=17.66 volume=1153012
session_date=2026-10-28 open=17.70 high=18.12 low=16.14 close=16.30 volume=3233929
Sixth close: 2026-10-21=18.59; return_5d_pct=-12.318451; high_20d=18.75; low_20d=16.08; avg_volume_20d=1637457.8 shares. The 20-session statistics exclude October 29. Even low_20d multiplied by avg_volume_20d is $26330321.424, above the liquidity floor.
StoneX raised its target to Buy, target $30 (from $23) on 2026-10-29 — source
Return: -1.60%; alpha: 0.08%. - FTNT — confidence 59%. Verified action: BofA Securities downgraded Fortinet from Buy to Neutral and reduced its target from $220 to $140. The rating change takes precedence over the accompanying target cut, so action=downgrade. published_at is the timestamp of the cited MT Newswires report, not a verified original research-note release time. The business concern is growth composition: product revenue grew 37%, slower than the prior quarter's 52%, and full-year services guidance fell to $5.15-5.17 billion from $5.18-5.22 billion. Strong counter-evidence is that total revenue and billings guidance rose, with quarterly revenue of $2.08 billion and non-GAAP EPS of $0.91. Primary source=https://www.globenewswire.com/news-release/2026/10/28/3390743/0/en/fortinet-reports-strong-third-quarter-2026-financial-results.html.
Before the earnings release and downgrade, October 23-27 brought three rising closes, followed by an October 28 reversal that closed only 8.75% above its low on 2.77 times average completed-session volume. Today's gap below the completed $166.22 range floor represents a severe repricing. The rebound from $127.40 is the strongest price-based counter-evidence; however, the quote remains below the attempted recovery to $139.44. My inference is that recurring-services uncertainty can sustain selling after that rebound, supporting a further 2.5% decline over the scored window. The already observed 22% loss is not the forecast opportunity. A subsequent regular-session close above the verified $139.44 partial-session high would invalidate the bearish continuation thesis.
latest_price=135.18 USD; quote_time=2026-10-29T13:29:59-04:00; quote_type=regular-session timestamped web snapshot; source=https://www.google.com/finance/quote/FTNT:NASDAQ. Pre-earnings price=174.08 at the October 28 regular close, 2026-10-28T16:00-04:00; source=https://stockanalysis.com/stocks/ftnt/history/; pre_earnings_to_quote_return_pct=-22.346048. This combines earnings, analyst and market effects rather than isolating the downgrade. pre_action_price=null; action_to_quote_return_pct=null because no immediately preceding BofA-note quote was verified.
PARTIAL SESSION as_of=2026-10-29T13:24-04:00: open=127.96, high_so_far=139.44, low_so_far=127.40, cumulative_volume=11628375 shares; source=https://stockanalysis.com/stocks/ftnt/history/. This partial snapshot precedes the latest quote. Relative intraday volume cannot be established without a verified same-time baseline. The quote source reports rounded market capitalization of $99.34 billion.
Completed history uses US Eastern regular sessions and split-adjusted OHLC Close, not dividend-adjusted Adj Close. Source=https://stockanalysis.com/stocks/ftnt/history/.
daily_5d:
session_date=2026-10-22 open=174.22 high=177.71 low=171.69 close=172.98 volume=3961836
session_date=2026-10-23 open=172.88 high=180.17 low=171.19 close=179.99 volume=3225722
session_date=2026-10-26 open=181.34 high=184.76 low=178.30 close=183.33 volume=3228339
session_date=2026-10-27 open=181.41 high=186.32 low=180.82 close=185.00 volume=3585303
session_date=2026-10-28 open=183.76 high=185.13 low=173.02 close=174.08 volume=11099352
Sixth close: 2026-10-21=174.05; return_5d_pct=0.017236; high_20d=197.13; low_20d=166.22; avg_volume_20d=4010051.5 shares. The 20-session statistics exclude October 29.
BofA Securities downgraded to Neutral, target $140 (from $220) on 2026-10-29 — source
Return: 0.04%; alpha: 1.63%. - MSFT — confidence 56%. Verified action: BofA Securities analyst Tal Liani raised Microsoft's target from $600 to $630 while retaining Buy. StreetInsider supplies the report timestamp; it is not independently verified as the original note-release time. The unchanged Buy rating, targets and analyst name are also reported at https://ca.finance.yahoo.com/news/wall-streets-top-analysts-are-loving-microsofts-q1-results-161837180.html. The fundamental case is positive: Microsoft reported quarterly revenue of $92.3 billion, up 19%, non-GAAP EPS of $4.89, and Azure revenue growth of 48%: https://www.microsoft.com/en-us/investor/earnings/fy-2027-q1/press-release-webcast. These results and the target increase are the strongest counter-evidence to my down call.
Before earnings and the revision, October 22's near-low close was followed by a recovery on October 23 and 26. The next two closes stalled, and October 28 finished near the middle of its range on 1.42 times average completed-session volume, offering mixed rather than decisive continuation evidence. Today's opening gap above the prior close failed: the stock traded from $540.22 down to $519.56 and remains below both the opening price and prior close. My inference is that this rejection of positive news can produce another 1.4% decline over the scored window despite the longer-term target raise. This is a price-response hypothesis, not a claim of deteriorating reported fundamentals, and receives the lowest evidence confidence. A subsequent regular-session close above the verified $540.22 opening level would invalidate the failed-gap thesis.
latest_price=524.09 USD; quote_time=2026-10-29T13:19-04:00; quote_type=regular-session snapshot, vendor labeled real-time; source=https://stockanalysis.com/stocks/msft/history/. Pre-earnings price=531.01 at the October 28 regular close, 2026-10-28T16:00-04:00, from the same history source; pre_earnings_to_quote_return_pct=-1.303177. This combines earnings, revisions and market effects. pre_action_price=null; action_to_quote_return_pct=null because an immediately preceding BofA-note quote was unavailable.
PARTIAL SESSION as_of=2026-10-29T13:19-04:00: open=540.22, high_so_far=541.29, low_so_far=519.56, cumulative_volume=23232638 shares; source=https://stockanalysis.com/stocks/msft/history/. Relative intraday volume cannot be established without a verified same-time baseline. Rounded market-cap source=https://stockanalysis.com/stocks/msft/.
Completed history uses US Eastern regular sessions and split-adjusted OHLC Close, not dividend-adjusted Adj Close. Source=https://stockanalysis.com/stocks/msft/history/.
daily_5d:
session_date=2026-10-22 open=533.14 high=533.60 low=520.30 close=521.08 volume=21266610
session_date=2026-10-23 open=517.75 high=525.77 low=515.70 close=524.51 volume=13681067
session_date=2026-10-26 open=527.40 high=533.25 low=521.69 close=532.60 volume=19021036
session_date=2026-10-27 open=532.35 high=534.70 low=528.33 close=531.33 volume=19532321
session_date=2026-10-28 open=531.82 high=535.00 low=526.14 close=531.01 volume=28594163
Sixth close: 2026-10-21=532.95; return_5d_pct=-0.364012; high_20d=551.15; low_20d=512.17; avg_volume_20d=20144011.9 shares. The 20-session statistics exclude October 29.
BofA Securities (Tal Liani) raised its target to Buy, target $630 (from $600) on 2026-10-29 — source
Return: -1.02%; alpha: 2.69%.
How models have done in Analyst Revisions
ForecastNest's own settled Analyst Revisions runs at this horizon, up to and including this one. Members' own research is not counted. Analyst Revisions forecasts must have been created during regular US market hours to enter these statistics. Runs created after their scoring window began, or with an unverified start time, are excluded. The chart is the average alpha vs SPY 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 picks beat SPY, 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 picks with positive alpha | Avg alpha vs SPY |
|---|---|---|---|
| Gemini Pro | 1 | 100.00% | +0.74% |
| Grok 4.5 | 8 | 55.00% | +0.57% |
| Gemini 3.6 Flash High | 2 | 60.00% | +0.54% |
| GPT-6 | 15 | 52.00% | +0.35% |
| DeepSeek | 15 | 53.33% | +0.14% |
| Perplexity | 12 | 46.67% | -0.00% |
| Muse Spark 1.1 | 14 | 52.86% | -0.12% |
| Gemini 3 Pro | 7 | 42.86% | -0.20% |
| Gemini 2.5 Flash | 2 | 40.00% | -0.21% |
| Claude Sonnet 4.6 | 1 | 80.00% | -0.32% |
| Claude Opus 5.5 | 3 | 40.00% | -0.71% |
| Grok 4 | 2 | 20.00% | -0.72% |
| Grok 4.7 | 3 | 40.00% | -1.14% |
| Gemini 1.5 Pro | 2 | 50.00% | -1.26% |
| Claude Sonnet 5.5 | 5 | 44.00% | -1.30% |
View settlement prices
| Model | Symbol | Start | End | Source | Notes |
|---|---|---|---|---|---|
| GPT-6 | SPY | 768 | 755.1401 | https://finance.yahoo.com/quote/SPY/history/ | |
| GPT-6 | NOW | 110.665 | 104.5 | https://finance.yahoo.com/quote/NOW/history/ | |
| GPT-6 | GOOGL | 283.15 | 278.965 | https://finance.yahoo.com/quote/GOOGL/history/ | |
| GPT-6 | PRCH | 20.365 | 20.04 | https://finance.yahoo.com/quote/PRCH/history/ | |
| GPT-6 | FTNT | 135.23 | 135.29 | https://finance.yahoo.com/quote/FTNT/history/ | |
| GPT-6 | MSFT | 525.95 | 520.585 | https://finance.yahoo.com/quote/MSFT/history/ |
Standard prompt analyst-revision-v3, SHA-256 14f9c25d63c218b7aff4622ca97540ad5b877cb1dc2eabdcd8549b8afc9c6390. The prompt is ForecastNest's own research and is not published; the fingerprint shows it was not changed after the fact.
Calculation: analyst-revision-submission-v1.