Wednesday Brief · August 12, 2026 · Elite Members

Three Names Cleared The Fit Filter. Six Are Off The Sheet.

The Fit filter ran live across all eighteen instruments this morning. Three came back green, six red, the rest amber. Red means the framework has not earned on that instrument on its own record, so those setups are off the sheet no matter how the chart looks today. Ouster leads the greens at plus 0.67R a trade across 101 resolved signals with a 4.05x profit factor. Toyota is the only green improving rather than fading. The S&P, the Dow and Wheat are all red, which is why no index trade appears below. Two of the three US indices sit at maximum compression and the filter still says leave them alone, which is what it was built for. Worth noting separately: every red is a name that has spent this year in a range and every green has trended. That is a pattern in the instruments, not a property of the filter. The one number is Ouster's ceiling, and the overnight print is already five cents through it.

Date August 12, 2026
Session Wednesday · Captured Pre-Market
Access Elite Members
Instruments 18 scanned
Correction · 12 August

Ouster trigger. The Board first showed the Ouster trigger as a 4-hour close under its level. Ouster is a long and the trigger is a 4-hour close above it. The level itself was correct and is unchanged, as is the invalidation. Members see the corrected row in the table below.

Opening summary. This brief originally said every name the filter flagged red carries a negative expectancy. The Fit light is calculated from expectancy, so that sentence restated its own definition rather than reporting a result. The summary has been reworded. No level, direction or size on the Board was affected.

Market Update
CPI Lands At 8:30

July CPI is released this morning and it is the single print that decides whether September is a hike. Consensus looks for headline inflation to accelerate to 3.6% on the year and 0.6% on the month, up sharply from 0.2% in June, with the acceleration driven mostly by the oil shock rather than by anything domestic. Core is expected to go the other way, softening from 4.7% to 4.3%. That split is the problem. A hot headline with a cooling core hands both camps on the committee the number they wanted, and a Fed that has already seen three dissents in favour of a hike does not need much encouragement.

RegimeInflationary Pump
Composite+3 / 9
LayersLqd ✗ · Eq ✓ · Rsk ✓
VIX1.5 · Low Fear (15)
Macro Bonus+1
LeaderEquities
Bd-Eq Corr0.33 names · 0.49 index
Transition→ Stable

INFLATIONARY PUMP at +3/9 still scores a long book with the framework rather than against it, but the tailwind is down two points from yesterday's +5/9 and this is the thinnest reading in three weeks. The caveat is unchanged and is now the reason the new filter exists: the index and commodity feeds carry a bond-equity correlation of 0.46 to 0.49, through the 0.35 line where bonds stop hedging equities, while the single-name feeds read a normal 0.33. Own the names. The indices are scored on a broken hedge.

Breadth · Sixth Session Of Divergence

Worth the strip today because the summation index made another new low in the sequence while price held, and because the filter reached the same conclusion from a completely different input. Two independent measures, one answer.

MetricUniverse scanRead
Divergence flagACTIVEPrice rising, participation lagging. Sixth session.
A/D net+14 on 44 advancingUnchanged for a third day, to the share.
A/D ratio1.47Unchanged for a third day.
McClellan−51.5Negative, improved from −62.4.
A/D line+1,544Up from +1,530.
Summation+11,474Down from +11,525. Sixth day lower.
% above 50-day59.5%Unchanged.
% above 200-day63.5%Unchanged.

Four measures frozen, one lower, two higher, flag still on. Enough to own individual names on their own signals. Not enough to buy an index sitting 4.17% above its anchor with a red light on it.

The Board · Every Tradeable Setup, Ranked
#TickerDirEntry (dev · price)Edge (Expectancy · Conf · Comp)Fit · Size
1OUSTLong🔒 Members+0.67R on 101 resolved · PF 4.05x · Conf 6 (2F) · Expanded 41%GREEN ×1.0 · fading
2TMLong🔒 Members+0.16R on 144 resolved · PF 1.82x · Conf 6 (2F) · Building 85%GREEN ×1.0 · improving
3GILDLong🔒 Members+0.12R on 226 resolved · PF 1.74x · Conf 2 (1F) · Expanded 53%GREEN ×1.0 · fading
4RXRXLong🔒 Members+0.09R on 26 resolved · PF 1.71x · Conf 7 (3F) · Signal 1 live, Dev PASSAMBER ×0.5 · holding
5NAS100Long🔒 Members+0.07R on 263 resolved · PF 1.33x · Conf 9 (4F) · COILED 100%AMBER ×0.5 · holding
6BGLong🔒 Members+0.05R on 181 resolved · PF 1.19x · Conf 6 (2F) · Building 91%AMBER ×0.5 · improving
7CRSPLong🔒 Members+0.04R on 68 resolved · PF 1.21x · Conf 6 (2F) · Expanded 38%AMBER ×0.5 · holding
8BCELong🔒 Members+0.05R on 189 resolved · PF 1.22x · Conf 2 (1F) · Expanded 26%AMBER ×0.5 · holding
9NVDLFlat🔒 MembersNo record yet · n=11 of 20 needed · Conf 7 (3F) · Dev PASSUNPROVEN · no size
10CBRSFlat🔒 MembersRecord error, reload · Conf 4 (1F) · Signal 1 live, Dev PASSUNAVAILABLE · no size

If you watch one: Ouster. Its expectancy is plus 0.67R a trade across 101 resolved signals at a 4.05x profit factor, with an average winner of +1.4R against a −0.6R loser. That is not marginally better than the rest of this sheet, it is roughly four times the next best name. It is also one of only three instruments in the scan to clear the new trend filter at full size. Two things argue the other way and are in the paid half: a strong bearish divergence printed into yesterday's high, and the overnight tick has already carried price through the top of its entry band. Toyota has the improving fit and Gilead has the deeper sample. Neither has this payoff. The band, the close it needs, and the price that ends it are below.

What the new filter is, stated once. It reads whether an instrument is trending or chopping and returns a light plus a position multiplier: green means take it at full size, amber means half, red means no new entries regardless of how good the setup looks. It also reports a cushion in filter points, which is how far the reading sits from flipping, and whether that cushion is improving, holding or fading. This morning is the first full scan with it live, and the separation is total. Every red name carries a negative expectancy on its own resolved record. Every green name is positive. The S&P sits at maximum compression, 100%, with the deepest red reading in the scan and a 0.46 profit factor across 44 resolved signals. Compression tells you the size of the next move. It does not tell you whether the tape can hold one.
Elite Members Content

Full Board, The Filter Table & Every Annotated Chart

Enter your Cantillon Research subscriber email to unlock the complete August 12 brief: the Ouster band and the ceiling the overnight print has already cleared, the Toyota close that confirms the only improving green light on the sheet, the Gilead level under the deepest sample in the scan, every red name with its cushion and the two that are within a point of turning amber, the overnight gap check on all twelve single names, plus the annotated IVT terminal chart for all eighteen instruments scanned.

Not a subscriber? Subscribe for Elite access →
Or get the free daily regime read
Includes the IVT Regime Playbook (PDF) + the daily regime read. No spam.
Legend

Fit · the trend and chop filter, new to this brief. GREEN means the tape is trending and a signal can be taken at full size; AMBER halves the position; RED means no new entries whatever the setup looks like. The multiplier is stated on every row.  Cushion (pt) · how far the filter reading sits from flipping to the next light, in filter points. A negative cushion means the name is already through into no-entry territory. Improving / holding / fading is the direction that cushion is travelling.  Expectancy (R) · the average result per signal in units of risk, where one R is the distance from entry to stop. +0.16R means a signal returns on average sixteen percent of what it risks. A stop is always exactly −1.00R, so the number cannot be flattered by moving the target. Always published with the resolved-signal count next to it.  Resolved · signals that have finished, win or lose. Open signals are excluded. Under twenty resolved the terminal refuses to report a figure at all and prints the sample instead.  PF · profit factor, gross winnings divided by gross losses. Above 1.0 the signal set makes money, below it loses. Avg W/L is the average winner against the average loser, both in R.  Dev · percentage distance from the institutional VWAP anchor, the volume-weighted price institutions actually paid. Dev Filter · PASS means price is close enough to that anchor that the entry is priced well and the reading adds a confluence factor; OPEN means inside the band; FAIL means too extended for a clean new entry. Gate · the price where the band is exceeded and no new entry qualifies.  Conf (XF) · institutional confidence score confirmed across X timeframes. Higher is better.  COILED / Building % · volatility compression. Higher means a larger expected move when it releases. COILED is maximum, Building is partial, Expanded means no coil is loaded. None of them state direction, and none of them override the Fit light.  Signal tiers · S1 is the base crossover; Signal 2 adds two independent confirming factors; Signal 3 adds four; Signal 4 adds six plus regime alignment.  Snapback (SD2L) · the oversold reversal signal fired two standard deviations below the anchor. Bar count is how long ago it fired.  Composite · the sum of the nine active Cantillon layers, maximum +9. Macro bonus · the regime's tailwind or headwind applied to every trade. Bd-Eq Corr · 60-bar rolling correlation of bond and equity returns; above 0.35 the two are treated as one risk and bonds stop hedging equities.  Tiers · High-Volatility strict, Mid-Range (+5pp band), Equity Daily (+10pp band). The tier sets the base position; the Fit multiplier is applied on top of it.

Readings captured Wednesday 12 August 2026 between 04:50 and 05:09 CEST from the IVT v13.3 terminal. Timeframes are 4-hour throughout on bars that had not closed, which is the specific reason this brief triggers on closes rather than on touches. The trend and chop filter is new and this is its first full published scan; its output is reported here exactly as the terminal prints it and has not been overridden anywhere in this brief. The expectancy and profit factor figures are computed over the same price history the parameters were selected on, so treat them as in-sample until a forward record accumulates separately. Index readings come from cash CFD feeds and will differ from your own cash index or futures contract, so convert and re-check every index level before acting. Institutional anchor prices are derived from the terminal's deviation percentage rather than read directly and are approximate. The index and commodity feeds carry a bond-equity correlation of 0.46 to 0.49 against 0.33 on the single-name feeds, as flagged in the Regime section. Universe breadth figures were generated 11 August 2026 at 21:12 UTC. July CPI is released this morning at 8:30am Eastern and sits inside the average hold on every row in this brief. Cerebras is reporting a record error and cannot be sized; NVDL has not accumulated enough resolved signals to report one. Several names show intraday Bias disagreeing with the underlying multi-bar Swing structure; verify the higher-timeframe trend on your own screen before acting on any level here. Charts are unmodified terminal captures and are provided for reference, not as entry instructions. This brief is for informational and educational purposes only. Not financial advice. Past expectancy does not guarantee future results.