Value Zones and Focus Levels: What Actually Carried the Edge
The Indicator draws three Value Zone bands around the overnight anchor, sized on implied volatility. A natural trading idea: wait for a rejection or an acceptance at a band, and once the zone is cleared, enter at a Focus level. We tested exactly that — and learned which half of the idea mattered.
The short version
The IV-sized width of the zones carried no information: bands of another day's width did just as well. The sequence — price breaks a band, holds beyond it, pulls back to a Focus level, continues — was the best thing in the study, at +5.8 points per trade over 532 trades. It is fragile, its confidence interval touches zero, and it is now a frozen rule measured forward rather than a claim.
Rebuilding the zones honestly
The zones are anchored on the NQ price at the 18:00 ET session open and sized on the one-day expected move: the anchor times the ATM straddle's implied one-day sigma from the QQQ chain, with the Indicator's 0.25 premium and multipliers of 0.5, 1 and 2 for VZ1, VZ2 and VZ3. Through August 7 we had the chain; after that we calibrated the VXN index to it (ratio 0.81 on the overlap). 72 trading days of geometry, 49 with the product's Focus levels logged alongside. Same execution rules as our other studies: entry at the next bar's open, stop on the wick, forced exit 15:55 ET, 1 point of cost.
Four ways to trade a zone
| Family (stop 40, target 80, all bands) | Trades | Win rate | Pts / trade | t-stat |
|---|---|---|---|---|
| Accept beyond a band, continue immediately | 1,085 | 60% | +1.9 | 0.6 |
| Accept, then retest the band edge | 397 | 40% | −6.0 | −1.8 |
| Accept, hold, pull back to a Focus level | 532 | 45% | +5.8 | 2.0 |
| Reject at the band, fade toward the anchor | 206 | 48% | +5.7 | 1.2 |
The Focus-pullback family was best on VZ3 (+9.4 on 152 trades), positive on both sides, positive in June, July and August, and strongest between 10:30 and 12:00 ET — the same window that carried the battle-zone result in our previous study. Retesting the edge itself lost money. That distinction is the whole article: the level that just broke is not where the trade is; the product level behind it is.
Two controls that changed the conclusion
Bands of another day's width. We kept each day's anchor but swapped in the sigma of a random other day, 60 times. The pullback family made +7.3 on average with the wrong widths, against +5.8 with the right ones (p = 0.97). The implied-volatility sizing of the zones is not what the market reacts to.
Focus levels of another day. Same trick on the Focus levels: +1.6 with random-day levels against +5.8 with the real ones (p = 0.33). The product level is worth about four points per trade over a random price in the same area — real in direction, not yet significant.
Stripping the idea down
If width does not matter, maybe only direction does: pull back to a Focus level in the direction of the day's drift from the anchor, with no "the break holds" condition. That version made +1.6 to +4.8 points with t below 1, and a bias drawn at random per day did as well. The condition that carries the result is that the broken level holds during the pullback. Drop it and the edge is gone.
The mirror image is the one genuinely significant number in the study: pulling back to a Focus level and trading against the day's drift — fading a Focus level toward the anchor — lost 5 to 7 points per trade with t between −1.8 and −2.2. Do not fade a Focus level against the drift.
What we did with it
Rule 2 on our paper track record: a 1-minute close 10 points beyond VZ2 or VZ3, price holding beyond the edge, first pullback to a Focus level between the edge and the acceptance close, stop 40 beyond the Focus, target 80. Parameters frozen on September 6, decision at 200 out-of-sample trades, flagged as fragile because touch 15 with stop 40 is the only pair that worked. Progress is on the level statistics page.
Shape what the map computes for you
"Your map" lets you choose what the Indicator keeps and how the Focus levels are ranked — the same profile on TradingView and every native platform.
Start the 7-day free trial — $9.99/moThe honest caveat
Everything above is in sample on fewer than a hundred sessions, found by a search that looked at four families, three bands and several exits. Our later grid study (58,000 trades, every level family) puts a t-statistic of 2 squarely inside what random levels produce. The zones stay on the chart because they describe priced risk; nothing here makes them a signal.