2026-10-21

Desk take

What we think of this print

Close print for 2026-10-21. VIX 17.31 (low) and SPY 773.17 set the tape. Hybrid desk: the scanner names the setup, the model grades that same print, and published conviction is the blend. The model cannot add a name that did not fire.

Amazon.com, Inc. (AMZN)
Scanner what happened: AMZN traded cheaper than its industry peers on trailing P/E at 86/100 (setup 100).
Why it printed:
- Trailing P/E 21.1 is below the internet retail / big-box peer median 28.6 (26% cheaper).
- Market cap still sits above the peer median ($2.83T vs $350.4B).
- Cap is more than 2× the peer median — quality/size filter.
- VIX regime is low (17.31).
- 21d relative strength vs SPY is 2.9%.
- Relative volume 0.83x.
Invalidation: Multiple re-rates up to the peer median P/E, or a daily close back below SMA200; stop ~249.78.
Continuation of the 2026-10-20 alert, not a new one — this is print 6 of one run: the same call has gone out 5 times since 2026-10-14 across 7 calendar days, most recently on 2026-10-20. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-14, not to today.
Same-setup analogs on 21 names (n=265, n_eff=86.4, 95% CI 47–68%): 58% 21-session net hit rate against a 59% base rate, average +3.0%, vs SPY +1.2%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 86/100: Amazon presents a solid cheap_vs_industry setup with a trailing P/E well below retail peer medians. The massive market cap acts as a quality filter, and positive 21-day relative strength against SPY supports continuation. Low VIX regime further facilitates accumulation in large-cap growth.
Risk: A multiple re-rate or a daily close back below the SMA200 invalidates the thesis.

Taiwan Semiconductor (TSM)
Scanner what happened: TSM traded cheaper than its industry peers on trailing P/E at 80/100 (setup 98).
Why it printed:
- Trailing P/E 27.8 is below the semiconductors peer median 49.5 (44% cheaper).
- Market cap still sits above the peer median ($2.24T vs $902.3B).
- Cap is more than 2× the peer median — quality/size filter.
- VIX regime is low (17.31).
- Relative volume 0.88x.
Invalidation: Multiple re-rates up to the peer median P/E, or a daily close back below SMA200; stop ~407.09.
Continuation of the 2026-10-19 alert, not a new one — this is print 10 of one run: the same call has gone out 9 times since 2026-10-07 across 14 calendar days, most recently on 2026-10-19. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-07, not to today.
Same-setup analogs on 21 names (n=265, n_eff=86.4, 95% CI 47–68%): 58% 21-session net hit rate against a 59% base rate, average +3.0%, vs SPY +1.2%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 80/100: Taiwan Semiconductor trades at a notable discount relative to semiconductor peers while maintaining dominant scale. Despite negative short-term relative strength, the valuation gap and low VIX backdrop offer an attractive entry. The setup captures a quality name on a relative value basis.
Risk: Failure to hold above the SMA200 or an unexpected industry multiple contraction invalidates the trade.

NVIDIA Corporation (NVDA)
Scanner what happened: NVDA traded cheaper than its industry peers on trailing P/E at 77/100 (setup 94).
Why it printed:
- Trailing P/E 31.7 is below the semiconductors peer median 37.5 (16% cheaper).
- Market cap still sits above the peer median ($5.36T vs $902.3B).
- Cap is more than 2× the peer median — quality/size filter.
- VIX regime is low (17.31).
- Relative volume 0.80x.
Invalidation: Multiple re-rates up to the peer median P/E, or a daily close back below SMA200; stop ~210.16.
Continuation of the 2026-10-19 alert, not a new one — this is print 13 of one run: the same call has gone out 12 times since 2026-10-02 across 19 calendar days, most recently on 2026-10-19. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-02, not to today.
Same-setup analogs on 21 names (n=265, n_eff=86.4, 95% CI 47–68%): 58% 21-session net hit rate against a 59% base rate, average +3.0%, vs SPY +1.2%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 77/100: NVIDIA is flagged under cheap_vs_industry with a P/E modestly below its peer median alongside an exceptionally large market cap. The low VIX regime provides a stable macro climate for mega-cap tech digestion. RSI sits near neutral, avoiding chasing behaviour.
Risk: A breakdown below the SMA200 or sudden margin compression in AI hardware invalidates the thesis.

Texas Instruments (TXN)
Scanner what happened: TXN pulled back to its 20-day average inside an uptrend at 74/100 (setup 90).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=55.8 is a reset, not a chase.
- Close is 0.6% from SMA20 (pullback/tag).
- 21d relative strength vs SPY is 4.9%.
- Relative volume 1.62x.
Invalidation: Stop ~266.54 (2x ATR) or daily close back below SMA50.
Continuation of the 2026-10-14 alert, not a new one — this is print 2 of one run: the same call went out on 2026-10-14. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-14, not to today.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 74/100: Texas Instruments exhibits a clean pullback_in_uptrend with price holding above rising moving averages. Robust relative strength and above-average volume signal active institutional sponsorship on the dip. The RSI reset confirms a healthy consolidation rather than distribution.
Risk: Loss of the SMA50 or a drop below the 2x ATR stop level invalidates the setup.

Analog Devices (ADI)
Scanner what happened: ADI pulled back to its 20-day average inside an uptrend at 74/100 (setup 90).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=56.9 is a reset, not a chase.
- Close is 0.6% from SMA20 (pullback/tag).
- 21d relative strength vs SPY is 4.0%.
- Relative volume 0.97x.
Invalidation: Stop ~380.81 (2x ATR) or daily close back below SMA50.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 74/100: Analog Devices shows a constructive pullback in a well-established bullish trend structure. Price tagging the short-term moving average with positive 21-day relative strength underlines underlying strength. RSI near the midpoint represents a standard consolidation phase.
Risk: A daily close below the SMA50 invalidates the bullish continuation pattern.

KLA (KLAC)
Scanner what happened: KLAC sat in a deep drawdown during VIX stress at 73/100 (setup 88).
Why it printed:
- Close is -46.8% from the 52-week high — a deep drawdown.
- Drawdown reaches 35% from the high.
- Drawdown is at least the mean of this name's largest completed pullbacks (37.9%).
- RSI_14=31.4 is washed out.
- Trailing P/E is 3% cheaper than the industry-peer median.
- TTM EPS is still growing through the drawdown (12.9% YoY).
Invalidation: Stop ~148.17 (2x ATR) or TTM earnings turning negative.
Continuation of the 2026-10-13 alert, not a new one — this is print 2 of one run: the same call went out on 2026-10-13. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-13, not to today.
Same-setup analogs on 21 names (n=221, n_eff=58.2, 95% CI 50–74%): 63% 21-session net hit rate against a 59% base rate, average +3.6%, vs SPY +1.4%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 73/100: KLA fits the deep_drawdown_value criteria with a severe pullback from highs accompanied by an oversold RSI reading. Crucially, TTM EPS continues to grow year-over-year despite the price compression, offering fundamental backing to the washout buy. Historical analogues suggest favourable forward odds.
Risk: Deteriorating underlying fundamentals or a drop below the 2x ATR stop invalidates the trade.

Micron Technology (MU)
Scanner what happened: MU traded cheaper than its industry peers on trailing P/E at 70/100 (setup 85).
Why it printed:
- Trailing P/E 12.1 is below the semiconductors peer median 64.3 (81% cheaper).
- Market cap still sits above the peer median ($1.03T vs $696.8B).
- VIX regime is low (17.31).
Invalidation: Multiple re-rates up to the peer median P/E, or a daily close back below SMA200; stop ~818.46.
Continuation of the 2026-10-15 alert, not a new one — this is print 11 of one run: the same call has gone out 10 times since 2026-10-02 across 19 calendar days, most recently on 2026-10-15. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-02, not to today.
Same-setup analogs on 21 names (n=265, n_eff=86.4, 95% CI 47–68%): 58% 21-session net hit rate against a 59% base rate, average +3.0%, vs SPY +1.2%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 70/100: Micron Technology trades at a steep discount to semiconductor peers on a trailing P/E basis. Although relative strength is weak over the short term, the extreme valuation gap qualifies it as a valid value candidate. The low VIX environment supports risk-taking in cyclical tech.
Risk: A breakdown below the SMA200 or industry cyclical downturn invalidates the setup.

Alphabet Inc. (GOOGL)
Scanner what happened: GOOGL traded cheaper than its industry peers on trailing P/E at 71/100 (setup 82).
Why it printed:
- Trailing P/E 16.8 is below the interactive media peer median 17.5 (4% cheaper).
- Market cap still sits above the peer median ($1.96T vs $9.9B).
- Cap is more than 2× the peer median — quality/size filter.
- VIX regime is low (17.31).
- Relative volume 1.16x.
Invalidation: Multiple re-rates up to the peer median P/E, or a daily close back below SMA200; stop ~317.50.
Same-setup analogs on 21 names (n=265, n_eff=86.4, 95% CI 47–68%): 58% 21-session net hit rate against a 59% base rate, average +3.0%, vs SPY +1.2%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM confirm at 71/100: Alphabet qualifies under cheap_vs_industry with a P/E slightly below peer median and massive market capitalization. Increased relative volume indicates buyers stepping in near current levels. The low VIX backdrop provides a supportive environment for mega-cap accumulation.
Risk: Re-rating failure or a daily close below the SMA200 invalidates the setup.

Qualcomm (QCOM)
Scanner what happened: QCOM pulled back to its 20-day average inside an uptrend at 59/100 (setup 70).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=49.6 is a reset, not a chase.
- Close is -1.4% from SMA20 (pullback/tag).
Invalidation: Stop ~164.16 (2x ATR) or daily close back below SMA50.
Continuation of the 2026-10-05 alert, not a new one — this is print 3 of one run: the same call has gone out twice since 2026-10-02 across 19 calendar days, most recently on 2026-10-05. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-02, not to today.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM haircut at 59/100: Qualcomm presents a mild pullback in an uptrend, but lagging relative strength keeps conviction muted. Relative volume is subdued, suggesting a lack of aggressive accumulation. It remains a viable watch item rather than an immediate priority.
Risk: A break below the SMA50 or a failure to reclaim short-term moving averages invalidates the thesis.

Applied Materials (AMAT)
Scanner what happened: AMAT sat in a deep drawdown during VIX stress at 58/100 (setup 68).
Why it printed:
- Close is -44.7% from the 52-week high — a deep drawdown.
- Drawdown reaches 35% from the high.
- RSI_14=31.0 is washed out.
- Trailing P/E is 16% cheaper than the industry-peer median.
- TTM EPS is still growing through the drawdown (14.0% YoY).
Invalidation: Stop ~370.02 (2x ATR) or TTM earnings turning negative.
Same-setup analogs on 21 names (n=221, n_eff=58.2, 95% CI 50–74%): 63% 21-session net hit rate against a 59% base rate, average +3.6%, vs SPY +1.4%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM haircut at 58/100: Applied Materials satisfies the deep drawdown and growth criteria, but the weak relative strength indicates persistent selling pressure. While earnings growth through the drawdown is encouraging, price action lacks immediate momentum. Patience is warranted before committing capital.
Risk: A breach of the 2x ATR stop or a contraction in TTM earnings invalidates the setup.

Apple Inc. (AAPL)
Scanner what happened: AAPL closed near a 52-week high at 55/100 (setup 65).
Why it printed:
- Within 1.2% of 52-week high.
- Still in a bullish SMA regime.
- Beating SPY by 1.3% over 21d.
- VIX regime low.
Invalidation: Failed breakout if close back below prior 52w high and stop ~331.05.
Same-setup analogs on 21 names (n=739, n_eff=101.1, 95% CI 49–67%): 58% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.7%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM wait at 55/100: Apple hovers near its 52-week highs within a steady uptrend, registering a standard breakout attempt. However, the lack of heavy relative volume makes this close insufficient to alter an existing schedule. It is best treated as a hold rather than a fresh chase.
Risk: A failed breakout occurring on a daily close back below the prior high invalidates the setup.

Meta Platforms, Inc. (META)
Scanner what happened: META pulled back to its 20-day average inside an uptrend at 50/100 (setup 60).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=47.4 is a reset, not a chase.
- Relative volume 1.14x.
Invalidation: Stop ~648.57 (2x ATR) or daily close back below SMA50.
Continuation of the 2026-10-08 alert, not a new one — this is print 2 of one run: the same call went out on 2026-10-08. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-08, not to today.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM wait at 50/100: Meta shows a routine pullback above rising moving averages with a mild RSI reset. Negative relative strength over the past month dampens the appeal of this specific print. The setup is valid but lacks urgency.
Risk: Loss of the SMA50 support level invalidates the bullish trend structure.

Advanced Micro Devices (AMD)
Scanner what happened: AMD pulled back to its 20-day average inside an uptrend at 52/100 (setup 60).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=45.1 is a reset, not a chase.
- Relative volume 0.88x.
Invalidation: Stop ~499.62 (2x ATR) or daily close back below SMA50.
Continuation of the 2026-10-13 alert, not a new one — this is print 2 of one run: the same call went out on 2026-10-13. Today is not a fresh signal; the setup was already true when the desk last printed it. The analog below measures the first session of a run, so its hit rate belongs to 2026-10-13, not to today.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM wait at 52/100: Advanced Micro Devices demonstrates a standard pullback in a broader uptrend with subdued relative volume. While moving average alignment remains bullish, underperformance relative to the broader market suggests waiting for stronger confirmation. No immediate action is required.
Risk: A decisive close below the SMA50 invalidates the uptrend thesis.

Lam Research (LRCX)
Scanner what happened: LRCX sat in a deep drawdown during VIX stress at 52/100 (setup 60).
Why it printed:
- Close is -40.5% from the 52-week high — a deep drawdown.
- Drawdown reaches 35% from the high.
- RSI_14=31.8 is washed out.
- TTM EPS is still growing through the drawdown (40.6% YoY).
Invalidation: Stop ~232.51 (2x ATR) or TTM earnings turning negative.
Same-setup analogs on 21 names (n=221, n_eff=58.2, 95% CI 50–74%): 63% 21-session net hit rate against a 59% base rate, average +3.6%, vs SPY +1.4%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM haircut at 52/100: Lam Research displays a deep drawdown value setup with strong underlying earnings growth, offset by poor relative strength. High relative volume indicates heavy turnover, but the steep downward momentum demands strict risk controls. It is a secondary candidate compared to cleaner peers.
Risk: A break below the 2x ATR stop or a deterioration in TTM earnings invalidates the thesis.

Microsoft Corporation (MSFT)
Scanner what happened: MSFT pulled back to its 20-day average inside an uptrend at 48/100 (setup 55).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- 21d relative strength vs SPY is 7.0%.
Invalidation: Stop ~510.50 (2x ATR) or daily close back below SMA50.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM wait at 48/100: Microsoft shows solid relative strength and an uptrend structure, but the low relative volume during this pullback suggests quiet digestion. The setup has fired, but the print offers no compelling reason to adjust ongoing positioning today. Monitoring from the sidelines is appropriate.
Risk: A break below the SMA50 or a failure of the relative strength trend invalidates the setup.

Marvell Technology (MRVL)
Scanner what happened: MRVL pulled back to its 20-day average inside an uptrend at 48/100 (setup 55).
Why it printed:
- Price above SMA50 and SMA50 above SMA200 (bullish regime).
- VIX regime is low (17.31).
- RSI_14=46.8 is a reset, not a chase.
Invalidation: Stop ~212.97 (2x ATR) or daily close back below SMA50.
Same-setup analogs on 21 names (n=1021, n_eff=114.1, 95% CI 51–68%): 60% 21-session net hit rate against a 59% base rate, average +2.8%, vs SPY +1.5%.
The 95% interval includes 50%, so this is the desk's narrative setup rather than a demonstrated edge.
LLM wait at 48/100: Marvell Technology exhibits a routine moving average pullback with an RSI reset in a bullish regime. Underperformance versus the S&P 500 and light volume limit the enthusiasm for this print. It remains a low-conviction watch item.
Risk: A daily close back below the SMA50 invalidates the bullish structure.

Overall: mixed hybrid tape. Read each card's scanner grade, LLM stance, and hybrid conviction separately — not as one market call.

RESEARCH ONLY — not advice.

Scanner names what happened on the tape. The model grades that observation. Hybrid conviction is what publishes. The model cannot add a name that did not fire. Research only — not a timing call.

add reminders

Five add reminders

Live add reminders: K1, K2, K3, K4, and K5. Exclusive listing, strongest first. K6, K7, K8, K9, and K10 repeat those shapes on the name's industry tape and can print on the same close. K11 and K12 are that name's own RSI and can print on the same close. Not a named setup. Does not score. Now vs 52w is close versus the 252-session high. This-name median is that same drawdown on this ticker's prior fires of the same gate.

Quiet today: K1 Weekly RSI (fear) · K2 Weekly RSI (elevated) · K3 Daily RSI (fear) · K4 Daily RSI (elevated) · K5 Daily RSI (calm) · K6 Weekly RSI (industry stress) · K7 Weekly RSI (industry elevated) · K8 Daily RSI (industry stress) · K9 Daily RSI (industry elevated) · K10 Daily RSI (industry calm) · K11 Weekly RSI · K12 Daily RSI.

Quiet watch: no desk name is approaching. VIX must already be 30.