Recommendations
0 to acquire · 2 to dispose
| Ticker | Name | Close | Action | Action Price Where to act — a buy-limit on a pullback or a buy-stop above resistance, not the current price. Sanity-check vs Close: on a held name, an action price stranded above the close means the queued level was set when the stock was higher — stale, awaiting reprice. | Stop The risk-defined exit, set below the entry for a long (~1.2–1.8×ATR). A held long whose Stop sits at/above the Close is already below its stop — a stale box awaiting reprice. | 1wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | 4wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | 12wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| KLAC | KLA Corporation | 192.80 | SELL 30% | 192.80 | 175.00 | 193.11 (+0.2%) | 196.53 (+1.9%) | 206.17 (+6.9%) | KLA is the book's biggest position by far - 27% of the portfolio - and the price has been drifting lower for weeks while the protective stop was walked down from without a single share ever being sold. That is a risk line running away from a position that never shrank, so we are cutting 30% to bring the holding back to a sane size and banking a 9.7% gain on that slice. The company itself is fine - record quarterly revenue up 15% and a strong guide - but it trades near 53x forward earnings, sits 37% below its 52-week high, and fell 10% on its own good results, so this is about position size and risk, not about the business. ADR-0020 Trail Erosion is the sole origination: live stop 175.00 vs Peak Stop 272.00 = -35.66% with zero size reduction ever; the 08-04/08-05 fires were never executed so this remains the FIRST fire and is veto-immune to US-17/18/34/22/30. Signal leg (STRONG SELL - REDUCE cv68, Sell_Rank 2) is a lone print after a NO-TRADE gap - US-18/34 bars it from originating, carried as supporting only. ADR-0012 armed (Gain_ATR 1.32, +9.70%, T12_Progress 93.52 - passes US-26) but NOT fired: no 2nd warning code, conviction rose, Est_12W 206.17 alive vs the 196.66 threshold, only one NO TRADE. Valuation amplify (MoS -4.41 would bump SUPPRESSED by G2 - Est_12W is 2.06x the 99.91 DCF anchor. Trail on the retained 700 stays 175.00 (1.5 ATR below close = 173.44, never loosen); watermark re-bases to 175.00 on execution. US-37 STALEMATE session 6 - record TAKEN or SKIPPED. · news |
| SPCX | Space Exploration Tech Corp | — | SELL 50% | 108.27 | 135.00 | n/a | n/a | n/a | SpaceX has fallen another 13.6% to 108.27 and is now nearly 20% below the exit line we wrote for it, and today roughly 911 million insider shares - more than three times the entire tradable float - become eligible to sell. The quarter itself was good (revenue 7.8bn against 6.8bn expected, Starlink at 12 million subscribers), so this is a supply problem rather than a business problem, but a supply problem of this size deserves to be met with a smaller position. We sell half and keep half. WARNING: price is news-sourced, not model output - SPCX has no row in any scan snapshot, so there is no Est_1W/4W/12W, no Target_Price and no conviction, and none is invented. Exit reason: accelerating decisive break of the pre-committed 135.00 manual stop (-19.8%, deepening from -7.8% a session ago), US-27 directly on point, L4 satisfied (settled regular-session close). THIS IS THE 08-05 CARD REPRICED 124.46 -> 108.27, i.e. one session of non-execution has already cost on these - US-37 in its purest form. Sized 50% not 100% because the price is unverifiable by the platform, US-27 as narrowed opens a case rather than mandating a full exit, and the print was a beat. HARDENED TRIGGER (the old one was overtaken by price - SPCX closed below 110 BEFORE the unlock): the remaining 80 are a MANDATORY FULL-EXIT candidate at the next EOD unless SPCX posts a settled close back above 120.00 after the 08-06 unlock; independently, if SPCX appears in the scan CSV with a non-null Close below 135.00 it auto-unmutes and the remainder becomes a full-exit candidate immediately. Manual stop on the retained 80 stays 135.00, deliberately above the market. US-37 STALEMATE session 2 - record TAKEN or SKIPPED. · news |
| GOOGL | Alphabet, Inc.- Class A | 362.43 | HOLD | — | 356.00 | 363.77 | 364.85 | 390.14 | Alphabet had an ugly session - it opened near 383, ran to 384, then broke all the way to 357 and closed at 362, down 4%, on heavy volume. The trigger was a DeepMind leadership shake-up in which chief scientist Jeff Dean and senior fellow Sanjay Ghemawat are leaving to start their own company, arriving on top of a capital spending guide raised to 195-205bn and negative free cash flow of -5.9bn. That is real news and we are not ignoring it, but the position is up 195% from cost and the model gives no sell instruction, so we hold and let the stop do the work - it now sits just 1.8% below the price. Label is NO TRADE cv0 (Conflicted pillars, Momentum flipped Bearish 24 against Trend Bullish 69) and US-17 bars a bare NO TRADE from originating. ADR-0012 armed and genuine (Gain_ATR 13.93, +194.66%, T12_Progress 92.90 - passes US-26) with ZERO ticks fired: no 2nd code; conviction is a flip to NO TRADE cv0, a label change not a comparable conviction decay - declined here for exactly the reason it was declined on KLAC on 08-05, and it must be declined consistently or tick (b) becomes a back door around US-17; Est_12W 390.14 alive vs the 369.68 threshold; only one NO TRADE. No trail raise available (1.5 ATR below close = 336.65). US-31: the 32-day-stale overlay override (<345) has no originating force. RE-ARM restated and now genuinely live: two consecutive settled STRONG SELL - REDUCE prints, or any settled close below 356.00, fires the 30% trim without debate. · news |
| NET | Cloudflare, Inc. | 292.96 | HOLD | — | 272.00 | 293.73 | 302.95 | 323.69 | Cloudflare reports second-quarter results tonight after the close, so we hold through the event rather than pre-positioning around it. The position is up 13.6% from cost, the trend is intact above every moving average, and the company has beaten expectations in each of the last four quarters by an average of 11.6%. The stop sits at 272, comfortably above our cost, so a disappointing print gets handled by the risk line rather than by guesswork tonight. Q2 guide 664-665m is marginally below the 666.1m consensus; EPS consensus 0.27 (+28.6% YoY); ~1,100-person workforce reduction announced with the shift to an agentic-AI-first operating model; DDOG reports 08-06 BMO and sets the cloud-spend tone hours earlier. Catalyst-aware hold: no pre-trim, trim only on a confirmed lower-high or a decisive break. WATCH cv56 is not a REDUCE so NET is not in the sell pool. No raise available (1.5 ATR below close = 268.10 < 272.00). WARNING - 6th consecutive session: usportfolio.json carries TWO NET rows (150/0) and the loader takes the last, so this position prints Held=N with null Avg_Cost/Gain_ATR/T12_Progress and is invisible to the scan and the monitor going into its own earnings. De-duplicate before the next run (systemic finding #3). · news |
| NBIS | Nebius Group, N.V. | 218.99 | HOLD | — | 183.00 | 217.94 | 244.89 | 305.09 | Nebius slipped 3% with the rest of the AI complex but remains slightly above our cost and well above every major moving average. The model's AVOID label means do not buy more, not sell - it is not an exit instruction on a position we already own. We hold into the 12 August results with the stop at 183. Third consecutive session of conviction decay -> 54) and AVOID is not a REDUCE, so NBIS is not in the sell pool. Still +3.49% over the 211.60 cost, holding EMA20 203.64 / EMA50 206.41 / EMA200 156.09, RSI 53.4, Consolidation_Bullish with Trend 64. Stop_Price prints 218.99 = the close exactly - the US-29 artifact, rejected by hand. Not armed (T12_Progress 71.78, Gain_ATR 0.29). No raise available (180.95 < 183.00). The 12wk band [127.69-621.00] is 4.9x wide and near-uninformative - weight the 1wk. CRWV 08-11 is the closest read-through ahead of NBIS Q2 on 08-12. |
| MU | Micron Technology, Inc. | 893.19 | HOLD | — | 770.00 | 892.85 | 1006.86 | 1048.03 | Micron was flat on the day and stays modestly above our cost. The model has had no clear read on it for six sessions running, which is not a reason to sell - we keep the position and the 770 stop, which sits 14% below the market. The memory-pricing news flow around Samsung and SK Hynix capex still supports the thesis. 6th consecutive NO TRADE cv0 (US-17 bars it from originating), Indeterminate pillars, Trend 58 / Mom 43 / Vol 29. Above EMA100/EMA200 but still under EMA20 898.71 / EMA50 890.35 - the reclaim is incomplete. Not armed (T12_Progress 85.23, Gain_ATR 0.41). 1.5 ATR below the close is 771.01, effectively the standing line - no material raise. FLAG: the 12wk estimate rolled 1131.45 -> 1048.03 (-7.4%) on a flat session and the band [450.95-2151.36] is 4.8x wide. Re-based Trail Erosion -18.75% stays live but VETOED (US-17 six NO TRADEs; US-30 price +16.0% above the line and above cost); if MU ever prints a real sell signal that eroded watermark amplifies it. AMAT FQ3 08-13 and TSMC monthly ~08-10 carry the HBM/WFE read-throughs. |
| OKTA | Okta Inc. | 147.04 | HOLD | — | 140.00 | 147.89 | 151.31 | 165.42 | Okta was essentially unchanged and remains 5% above our cost. The signal went quiet rather than negative, which is not a sell case. We hold with the stop at 140 - a level that is above our cost, so the position is now protected into a profit rather than a scratch. HOLD cv63 -> NO TRADE cv0 (Conflicted, Momentum flipped Bearish 24 against Trend Bullish 69) - the same pillar shape GOOGL printed tonight, and the 6th session of a decaying run (51/64/68/63/0). US-17 bars it from originating. Above EMA20 140.63 with EMA50 129.04 / EMA200 103.36 far below. Not armed (T12_Progress 88.89, Gain_ATR 1.15). No raise available (137.63 < 140.00). WATCH: a second consecutive NO TRADE arms ADR-0012 tick (d) if this position ever becomes verifiably armed. |
| DELL | Dell Technologies Inc | 462.70 | WATCH | 462.70 | 407.00 | 461.29 (-0.3%) | 494.47 (+6.9%) | 612.65 (+32.4%) | Yesterday's Dell buy order was a stop-entry at 464.24 and the stock opened at 466.60, so it will have executed - but the holdings file still shows no Dell position. If it filled we own at about 466.60, currently marked at 462.70, a loss of roughly We are not adding more: the setup that justified the purchase has already changed character, and the stock's own signal has weakened. We raise the protective stop to 407 and need the fill confirmed or the order confirmed cancelled. DISPOSITION OWED (ADR-0019/US-37): record TAKEN (trade buy) or SKIPPED (trade skip us DELL) - an unrecorded fill is worse than an unexecuted card because the monitor, the Trail Erosion watermark and every held-name path silently ignore a position. No add: STRONG BUY cv88 on BreakoutUp_52W (-2 cell) decayed to BUY cv68 on FailedBreakdown_Up_20D (-7 cell, n=7, hit_1w.143), one of L1/US-16's three momentum-chase traps; adj = 68 - 10 (MoS -2.59) - 7 = 51, sixteen points under the floor. Stop 404.50 -> 407.00 = 1.5 ATR below the close, NOT the model's 444.14 which sits only 4.0% under a name with an 8.0% ATR. Quality_Pass True (F=8, Bargain True) - still the only quality-passing name in the buy pool. |
Outcome & track record
Accuracy at the time of this record.
117 Graded decisions Past recommendations old enough to score against what actually happened (target hit / stopped out / closed). This is the evidence base — the bigger the number, the more history behind the calls. Click to inspect every one. See the decisions →
-0.132 1-week calibration How much of the eventual 4-week move typically shows up in the first week (0 = almost none, 1 = most of it). Use it to read the 1-week estimates: a low number means prices drift toward the target slowly, so don't over-read a soft week-1.
52 Published records How many dated end-of-day verdicts exist for this book — the length of the public track record you can browse in the Archive below.