Recommendations
0 to acquire · 0 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 |
|---|---|---|---|---|---|---|---|---|---|
| — | AI - HK Leveraged ETFs | — | NO ACTION | — | — | n/a | n/a | n/a | Book stays 100% cash at HKfor a 41st session. The decisive fact of the week is a non-confirmation: NVIDIA delivered the best AI-capex print available - revenue US22bn against US37bn expected, a US$108bn next-quarter guide, roughly 70% growth guided for next year and explicit commentary that memory supply stays a bottleneck into 2028 - its own shares rose 9%, and across the 27-28 August stretch both of these leveraged Korea memory trackers FELL, 7709 by 2.8% and 7747 by 2.4%, each closing near its low and below its own open on the thinnest volume of the month. A tape that will not rally on its own best news has not bottomed. Volatility is the one thing still improving - ATR/price fell to 11.9% and 10.7%, the lowest of the window on both, which genuinely reduces the decay drag these daily-reset funds carry - but compression into indifference is quiescence, not accumulation. Valuation is blind on both (ETF, no intrinsic anchor), so this is a technical and news judgement only. Nothing bought, nothing held to sell. OPERATIONAL: both print FailedBreakout_Down_20D (7709 10th observed, 7747 19th), consensus Consolidation_Bearish, regime Ranging - a bearish family, so HL-4 screens them out before conviction is consulted. Conviction de-converged 7709 (new book-window low, takes out 39 on 08-25) and 7747; 28 and 14 points under the 65 floor. HL-3 binds on both: Stop == Entry == Close (36.44/72.40), Target_Price blank, Order_Type Market, Exit_Strategy immediate - 10th and 19th observed zero-width risk band, so no risk-defined entry exists at any conviction. FINDING 4 ESCALATED: Target_12Week is pinned at 0.5699-0.5701 x close across BOTH tickers and all four snapshots to 4dp, i.e. T12 = 0.570 x Close mechanically - which makes HL-2's 'T12 >= close' gate unsatisfiable by construction, permanently ~43% away. Moot today (the other three gates fail badly); flagged for core/predictions.py and for the next learnings review, NOT self-applied - the operational bar leans on bullish label + second confirming print + usable stop + conv >= 65. DATA GAP: no 2026-08-27 snapshot exists for any HK-family book (US books have full 08-27 coverage), so all deltas span two sessions - third gap of the window after 08-20 and 08-24. L1 inert-but-noted: both flipped to a VWAP DISCOUNT for the first time in the window (7709 +3.47%->-3.74%, 7747 +7.81%->-1.24%) - a filter satisfied with no trigger to filter. Valuation blind (NA_ETF both, FetchedAt 2026-07-05, val_adj 0, quality gate off). Scorecard 0 graded of 103, no actionable cells, calib_1w null -> f=0.30 (L5/HL-6). Counterfactual optional_trim n=22 unchanged, still mis-signed for a cash-only book. Overlay 54 days stale. NO pre-positioning into the 31 Aug Jackson Hole HK reaction, 31 Aug China NBS PMI, or the 1 Sep Korea August semiconductor export line (L6/HL-6) - the Korea print is evidence for the second-print test, never an entry trigger. |
| 7709 | XL2CSOPHYNIX | 36.44 | WATCH | 36.44 | 36.44 | n/a | n/a | 20.77 (-43.0%) | The 2x SK Hynix tracker gave back both of the gates it won last session and made a new conviction low for the book. It closed 36.44, down 2.8% across the two-session stretch, only 8.8% up its daily range and 5.4% BELOW its own open of 38.50 - and it did that in the window that contained NVIDIA's blowout quarter and explicit commentary that memory supply stays tight into 2028. When the single most favourable catalyst available to a leveraged HBM proxy produces a lower close on 21% less volume, that is the market declining to participate, not a pause. Conviction fell, the lowest this book has recorded, and the engine's bracket turned from HOLD back to EXIT. The one genuine positive is volatility: ATR compressed a third straight session, 4.70 to 4.34, taking ATR down to 11.9% of price - real relief on a daily-reset fund, but a precondition, never a trigger. Not buyable, nothing held. OPERATIONAL: LOST EMA5 37.08 (-1.71%) and EMA10 37.00 (-1.50%), both of which were held on 08-26; EMA20 gap WIDENED 40.96/-8.44% -> 40.21/-9.37%, and in ATR terms 0.74 -> 0.87 ATR, reversing last session's largest gate improvement. Still -35.85% below EMA50 56.80 and -41.26% below EMA100 62.04; no EMA200 prints (insufficient listed history). Pillars trend 40 flat, momentum (entire 08-26 gain given back), volatility 27 flat. RSI 43.30->42.53. Volume 109.4M (-20.9%). 10th observed FailedBreakout_Down_20D print and 10th observed zero-width risk band - Stop == Entry == Close 36.44, Target_Price blank, Order_Type Market, Exit_Strategy immediate: HL-3 binds and defeats an entry before the falling-knife veto is reached. HL-4: Action_Rationale reads 'Flat; bearish signal (conv=37); do not open (long-only)' - the acquire floor screens bullish actionable labels only. HL-2 re-arm gates: no bullish label, no second confirming print, no usable stop, conviction 28 under the floor; Target_12Week 20.77 sits 43.0% below close but that gate is SUSPENDED-PENDING-FIX (T12 = 0.5699 x close, a mechanical artifact - see Finding 4). L1 inert: VWAP premium +3.47% -> -3.74%, a first close below VWAP for the window - filter satisfied, no trigger supplied. Off the 08-17 window peak 40.80 by -10.7%; still +29.7% above the 08-11 low 28.10. Declined-print marks all deepened: 07-14 @65.20 -44.1%, 07-20 @52.58 -30.7%, 07-27 @54.50 -33.1%, 08-07 @31.06 +17.3% (the one miss, narrowing). NO pre-positioning into the 1 Sep Korea August semiconductor export line (horizon band +/-6% raw, ~2x in vehicle terms) or the 31 Aug Jackson Hole reaction - L6/HL-6. · news |
| 7747 | XL2CSOPSMSN | 72.40 | WATCH | 72.40 | 72.40 | n/a | n/a | 41.27 (-43.0%) | Conviction ticked up and everything that matters got worse. The 2x Samsung tracker closed 72.40, down 2.4% over the stretch, below its own open and 19% up a narrow range - on volume of 5.53M, down 32% and a NEW LOW for the entire window, taking out the 8.19M floor set two sessions ago which had itself taken out 8.10M. Participation is collapsing through successive floors in the same week NVIDIA guided next quarter to US$108bn and called memory supply a bottleneck into 2028. It remains below every moving average that matters and has not attempted a fourth round-trip at the 200-day since losing it. The one real improvement is volatility, and it is a large one: ATR fell 9.70 to 7.77, nearly 20%, taking ATR to 10.7% of price - the lowest ratio of the window and a material reduction in the decay drag a daily-reset 2x fund carries. That earns continued watching, not capital. OPERATIONAL: below the ENTIRE ladder for a sixth observed session - EMA5 74.84 (-3.26%), EMA10 75.93 (-4.65%), EMA20 78.09 (-7.29%), EMA200 85.54 (-15.36%), EMA50 93.73 (-22.76%), EMA100 99.35 (-27.13%). Pillars trend, momentum (entire 08-26 gain given back), volatility off the Bearish label; bracket stays HOLD. RSI 45.68->44.98. Conviction, still 14 under the 65 floor, and the momentum pillar - the one that would have to lead a genuine turn - fell. 19th observed FailedBreakout_Down_20D print and 19th observed zero-width risk band: Stop == Entry == Close 72.40, Target_Price blank, Order_Type Market, Exit_Strategy immediate - HL-3 binds. HL-4: Action_Rationale reads 'Flat; bearish signal (conv=51); do not open (long-only).' HL-2 re-arm gates: no bullish label, no second confirming print, no usable stop; Target_12Week 41.27 is 43.0% below close but that gate is SUSPENDED-PENDING-FIX (T12 = 0.5700 x close, mechanical - Finding 4). L1 inert: VWAP premium +7.81% -> -1.24%, first close below VWAP in the window - a filter satisfied with nothing to filter. Off the 08-17 window peak 89.98 by -19.5%; still +10.2% above the 08-11 low 65.68. Declined-print marks: 07-14 @95.30 -24.0%, 07-27 @83.42 -13.2%, both deeper than last session. NO pre-positioning into the 1 Sep Korea August semiconductor export line or the 31 Aug Jackson Hole HK reaction - L6/HL-6, bands doubled on a 2x vehicle. · news |
Outcome & track record
Accuracy at the time of this record.
0 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.
0.000 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.
50 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.