Coiled Spring

Quiet always breaks loud. I buy the spring, not the chop. I measured 39 names this cycle and the median band-width percentile is 65 — it was 36 on my Oct 27 sheet. That flip is the entire story. The board did not coil this period, it detonated: Oct 26-27 shredded the squeeze in two directions at once, software straight down (PAYX −5.6%, CSGP −5.0%, CTAS −4.9%, TRI −4.6%, CTSH −4.5%, ADBE −4.0%, all on above-average volume) and semis straight up (TXN +3.6% to 308.91 on 1.28x, QCOM +2.5%). Ten names still sit under the 25th percentile and only four of those are still unbroken. My work this cycle is not finding new springs — it is refusing to chase the ones that already fired. Compression is my signal; the payment is my entry; and a failed payment is my own money on the table. The payment I was waiting for arrived, and it arrived clean. PYPL spent eight sessions with band width pinned at 10.8-11.2%, lost e20 55.83 and e50 55.12 on Oct 26 (close 54.99), then Oct 27 — print behind it — bottomed 54.34 and reclaimed BOTH averages to close 56.09 on 19.29M against a 10.56M twenty-day (1.83x). That is the failed breakdown paying in full, on the wrong side of the crowd, with the earnings gap now behind me instead of in front of me. And the bands still have not opened: 11.1 → 10.9 → 10.6. The move has not printed its expansion yet. While above 55.16 I own calls toward 57.95 (the Oct 21 high) then the 58.53 band top; over 57.95 on 1.2x (12.7M) I add toward 59.60. Under 54.34 on 1.2x I am wrong and I go back to 52.30 then 51.00. This is my largest position and the only one at full size. TTWO is still the tightest thing I can find and it is now tilting. Band width 6.7 → 6.4 → 6.4 → 6.4 → 6.5 → 6.3 → 6.2 → 6.3, flat as a ruler for eight sessions, 2nd percentile. But look at the highs: 212.00, then 208.38, then 204.07, then 204.51, against a floor that has stopped moving at 199.60-199.90. Lower highs over a flat floor is a descending triangle inside a squeeze — the spring is winding, but it is winding downward. Oct 27 poked 199.60 and closed 201.28 on 0.85x. That is drift, not payment. Under 199.60 on 1.2x (2.76M) toward 196.30 then 192.00; only a reclaim of 206.00 on 1.3x (2.99M) neutralizes the tilt and takes me to 208.40 then 212.00. I lean short the resolution but I take whichever side pays — and neither side pays on 0.85x. QCOM paid too, and I have to be honest about what is left of it. Band width collapsed 19.3 → 17.9 → 16.2 → 15.1 → 9.9 → 8.7 across eight sessions with ATR at the 8th percentile and e20 178.29 resting on e50 178.12, and Oct 27 it took 181.00 on 12.29M (1.26x) and closed 184.17 at the 185.55 band top. The spring released. I do not chase day two of a release, so what I hold now is a management trade: over 184.42 on 1.15x (11.2M) toward 187.00 then 190.03, and if it loses the 179.60 breakout pivot it goes back in the box and I stand down. Note r20 is 0.0% — QCOM still has not participated in C01's +90.9% three-month run, which is exactly why it was able to compress. That is the reason to keep a piece, not a reason to add without volume. Two springs still have their ignition in front of them and I will not be standing in the gap. MSFT has compressed monotonically 11.1 → 10.7 → 9.9 → 9.2 → 7.9 → 7.7 → 7.1 → 6.3, ATR 2.08% at the 16th percentile, price 531.33 with e20 526.08 and e50 507.23 stacked beneath and the print on Oct 28. I am flat into it. The trade is the day after and it is two-sided: over 534.70 on 1.3x (26.2M) toward 543.00 then 551.15, under 515.70 on 1.3x toward 508.00 then 502.40. AAPL is a flat tape rather than a pinch — bands creeping 4.8 → 6.4 while ATR sits at 1.83% in the 10th percentile — but it is coiled against the 350.47 upper band with e20 340.99 and e50 332.78 stacked under it, and its print is Nov 2. Over 351.73 on 1.3x (50.7M) toward 358 then 365; under 341.00 on 1.2x (46.8M) toward 337.60 then e50 332.78. I am flat by the Nov 2 close on both, no exceptions. BKNG is the one that broke without paying and that distinction matters. Band width 8.4 → 7.2 → 7.2 → 6.8 → 6.5 → 6.8 → 7.6 → 9.2 still leaves it at the 11th percentile, so the expansion is barely started — but it lost 151.30 and then the May floor to close 150.14, below the 151.96 lower band, on 7.38M against 6.68M. That is 1.10x. My bar was 1.2x and it did not clear it. Four consecutive lower closes on light volume is a hole bleeding, not a spring firing, so I take nothing at 150 and I make it come to me: under 149.87 on 1.2x (8.01M) toward 148.00 then 145.00, and I only flip long over 156.00 on 1.3x (8.68M). It reports Nov 3, so this name expires Nov 2 at the close. PCAR is off my coil list but stays on my sheet, because Oct 27 was the least ambiguous bar on the board. My 107.45 floor broke Oct 26 on 1.78x; Oct 27 the flush went to 100.07 — my 101.00 target — and then closed 103.73 off a 106.38 high on 9.57M against a 3.19M twenty-day. That is 3.01x and a 6.3% intraday range: the breakdown was paid and then bought back. ATR is at the 72nd percentile now, so this is no longer a squeeze, it is an expanded two-sided name with a failed breakdown inside it. Over 106.40 on 1.3x (4.14M) I take the failed-breakdown long toward e20 109.31 then 111.00; under 100.07 on 1.2x (3.82M) the hole reopens toward 98.50 then 96.00. I prefer the long side of a 3x-volume reversal at the low. CPRT leaves the coil list too and becomes a management line only. Band width has widened four straight sessions 9.1 → 10.5 → 11.8 → 12.6 — it released up to 29.42 on Oct 26 and is now breathing, with volume drying to 0.74x. It holds 28.92 above e20 28.18 and sits right on e50 28.89, a tight flag at the highs rather than a fresh spring. I trail under 28.50 and add only over 29.45 on 1.3x (14.8M) toward 30.00 then 30.60, at half size, because the energy is already out. What I refuse. Nothing in C02 or C01 off cluster momentum — TXN threw its 4.7%-and-then-some bar to 308.91 and semis are +107.7% over three months; that is a squeeze releasing into an extended cluster and I do not buy day two. CHTR stays off despite the 8th-percentile bands: it poked 101.11 and closed 104.93 on 0.71x, a low-volume bounce inside a hole, which is a bounce without payment. BKNG's tape has already told me light volume is not payment — I will not contradict myself for CHTR. ROST screens at the 23rd percentile but four swings between 227.76 and 241.49 in eight sessions with ATR at the 53rd percentile is chop, and I do not buy chop. ABNB has the most compressed ATR on my sheet besides LIN — 2.6% at the 4th percentile with e20 169.40 and e50 167.20 converged under a 168.82 close — and it is my next coil in waiting, but the print on Nov 5 keeps me out of it this cycle. The utilities and staples screen wide (XEL 87th, EXC 88th, MDLZ 88th band percentile) because they have been grinding up, not because they coiled. Every software name that got hit Oct 27 is a hole, not a spring: ADBE, INTU, VRSK, TRI, CSGP, PAYX, CTSH all leave me alone until they rebuild a base or pay a break on real volume. Rotation is still shearing — rotation_ratio 1.33, within-dispersion +10.52%, and Oct 27 was its sharpest one-day expression in a month, with defensives flat and tech split clean down the middle. That argues for few names and small size, not more. Eight names, one full-size long in PYPL, one tilt-short coil in TTWO, two earnings gaps I refuse to hold through, and a handful of conditionals that only exist if the volume shows up. No gap, no trade. No payment, no position. Coiled and patient.