Options lab · Marauri study account
Live· as of 2026-08-03 13:58:20A pre-market read on SPY for the $2,000 Marauri study account — the call, the strikes, the entry gate, and what to watch after, built for the 9:35–10:00 AM ET window.
Already in a position? Manage it →Market regime
Trending up
Trend tactics work: buy strength, give winners room, trail stops up.
SPY is in an uptrend with normal volatility (13% annualized, 45th percentile of its recent range). Descriptive, not predictive.
The engine's call · 2026-08-03
STAND DOWN: 56% confidence, below the 70% floorToday's confidence
lowThe engine expects the market to rise today, so its instrument is a call option — a contract that profits if SPY climbs. So far its calls have been right 9 of 14 decisive days — too few to lean on the percentage yet.
SPY at signal
$753.31 (+0.9%)
Calls right so far
9 of 14
If you always guessed up
56%
Track record 14 of 20 decisive sessions (70%) — the engine grades every call at the close and won't trust its own percentage before the bar. How it grades itself →
2 of 4 inputs agree: 5-min price structure (EMA 9/21), Factor tape (SPY/GLD/USO/TLT/UUP) point CALL with no input against.
context only · not the entry signal
Loading the live composite…
The same four-input composite as the morning signal, re-run on current ~15-minute-delayed data every 5 minutes. Context, not a trade trigger — the disciplined entry is the morning read in the 9:35 to 10:00 window. Chasing an intraday flip is how whipsaw losses happen.
Breakeven
$754.34
+0.1% from spot $753.31
Max risk / contract
$34.00
premium $0.34 · defined
Recommendation
Measuring
Sample too thin: n=14 of 20 required. No capital — data-collection only.
Confidence
Provisional14 of 20 observations — provisional until the bar is cleared. 95% precision ±0.38R.
See how NFAI is learning →Pre-tax edge
+0.50R
After-tax · 27.07%
+0.36R
Net vs hurdle · 95% lower
-0.12R
Downside: -0.50R per loss, realized 50% of trades.
The edge is a two-outcome distribution: it pays +1.50R on a win (50% of the time) and -0.50R on a loss. Pre-tax expectancy is +0.50R per trade. The instrument classifies as short-term, taxed at a blended 27.07%. Because a strategy’s own losses offset its own gains in the same year, tax scales the whole edge symmetrically — it shrinks the edge toward the hurdle, it does not flip its sign. The gate then compares the 95% lower bound of the after-tax edge (not the point estimate) against the +0.10R friction hurdle.
Federal long-term 15%, Pennsylvania flat 3.07%, local 0% (Philadelphia’s reach over trading gains is situation-specific — supply your own). These are the values you provide; the module does the arithmetic they imply.
Arithmetic on rates you supply — not tax advice, not a filing position. The gate never surfaces a recommendation without the after-tax number attached.
Refreshes every 5 minutes from delayed market data. Not a real-time stream. If the poll fails it holds the last good read and marks it stale.
Normal upward term structure: fear is structural, not acute.
Put wall
695
support
Call wall
754
resistance
Max pain
744
settlement magnet
Gamma peak
750
heaviest hedging
Nearest magnet to spot:
Last graded session · 2026-07-31
Sat out. Predicted no call; the market closed up +0.3%.
The engine sat out; the market closed up +0.3%. Abstention is graded in the no-trade ledger, not excused.
day type: gap and go · acceptance: rejected down · gap: +0.4%
Conditioned record: every slice is still measuring — the largest (gap and go) holds 6 of 10 graded days. No conclusion is drawn from a slice below its bar.
By regime: in Trending up (today’s tape): 3 of 6 — measuring — the other regimes are still measuring. Gating by regime waits for a slice to reach the parameter bar.
Per-input record: factor tape 9 of 10 graded · price structure 8 of 10 graded · VIX direction 7 of 10 graded · put/call 5 of 10 graded — 9 pre-v2 sessions ungradeable at input level.
Earned influence: All signals at the equal baseline — none has cleared its sample bar, so the live engine weights them identically and this layer changes nothing yet.
Shadow columns: credit dispersion 5 of 10 graded · Nasdaq vol premium 5 of 10 graded · 10-year yield 1 of 10 graded · HY credit 0 of 10 graded · real yield 0 of 10 graded — recorded and graded, not voting; a column earns a live vote only after beating the always-up baseline at 20+ graded votes.
Live real-world market data. Paper-trade study, not financial advice — options carry real loss of the full premium. Part of the NFAI Engine forward test.
Confidence keys to how many of the four inputs agree: 4 of 4 is high, 3 is the 70% entry floor, 2 is low, 1 or a split is no-trade. Neutral and unavailable inputs lower it rather than inflate it. The factor tape is a transparent heuristic, not a fitted forecast. The always-up baseline exists because SPY drifts up — a benchmark an honest engine must beat, not a strawman.
Signal 2026-08-03 · 9:59 AM ET · graded nightly against the session record, append-only, hash-stamped
No entry today — the conditions aren't met. The call still gets graded at the close.
Long call at $754.00. Risk is the premium and nothing more; the position turns profitable past breakeven. At-expiry payoff, before fees.
calls · August 3, 2026 · 0DTE · within 2% of spot, ranked by open interest
| Strike | P(ITM) | P touch | Delta | Gamma | Theta | IV | Open int. | Volume | Mid | Rank |
|---|---|---|---|---|---|---|---|---|---|---|
| C 754 | 0% | 0% | 0.182 | 0.083 | -0.31 | 15.2% | 12,676 | 40,506 | 0.34 | MEDIUM |
| C 753 | 100% | 0% | 0.285 | 0.102 | -0.59 | 16.0% | 11,514 | 51,062 | 0.63 | HIGH |
| C 750 | 100% | 0% | 0.611 | 0.098 | -1.08 | 18.8% | 11,334 | 23,928 | 2.23 | LOW |
| C 755 | 0% | 0% | 0.106 | 0.059 | -0.16 | 14.8% | 10,628 | 25,230 | 0.17 | LOW |
| C 745 | 100% | 0% | 0.886 | 0.038 | -0.28 | 23.8% | 10,588 | 481 | 6.45 | LOW |
Greeks are exchange values from CBOE (~15-minute delayed), not computed. P(ITM) is the Black-Scholes risk-neutral chance of finishing in the money (N(d₂)) and P(touch) is the chance the strike is reached before expiry, both from the strike's IV — risk-neutral estimates, not guarantees. Theta is per share per day.
Open interest frames SPY between 695 (put wall, support) and 754 (call wall, resistance), with max pain at 744. Into expiry, price often gravitates toward the heaviest open interest.
Open-interest-derived from the CBOE chain. Not a paid dealer-gamma (GEX) feed: it reads option open interest and Greeks, not dealer inventory, so it cannot sign dealer positioning. Treat as a magnet map, not a forecast.
True dealer GEX (signed gamma exposure) needs a paid feed of dealer inventory. The positioning above is an open-interest estimate; for signed dealer gamma, look up SpotGamma or Unusual Whales.
SpotGamma →How much more the market is charging for options than the stock has actually been moving. A big positive number means options are expensive right now — you pay for that on the way in.
+0.2 pts
How far the market itself expects SPY to move by this expiry. A trade needs to beat this to make money — it is the bar the market has already set.
0.41%
The round-trip cost of the bid/ask spread on a typical near-the-money contract. It comes straight off any edge before the trade has done anything.
3.5%
Implied volatility: how much movement options are priced for, at about a month out.
12.7%
How much SPY has actually moved over the last 20 trading days, annualized. Compare it to implied vol: the gap between them is the vol premium.
12.4%
How much more expensive downside protection is than equivalent upside. Equity options normally carry some of this; an unusually large gap means the market is paying up to hedge.
+4.6 pts
Volume above resting open interest
New activity relative to what is already on the books. Unsigned on purpose: the delayed chain reports volume, not who initiated it, so this says how much is changing hands — never which way it leans.
Derived from the same CBOE chain the strike ranker uses and the daily closes already stored — no extra feed. These describe the OFFER, not the direction: none of them votes in the recommendation above, and two of them (term structure, skew) are recorded as shadow columns that must beat the always-up baseline over 20 graded sessions before they ever do.
Verdict
Degraded
Two independent free feeds are asked what SPY costs — the quote, and the option chain read through put-call parity. This is how far apart their answers are.
0.28%
How far SPY has moved since the option chain was timestamped. Both feeds run about 15 minutes behind, so a big number means we are pricing a market that has already moved on.
-0.01%
Advisory only. The engine records what it thinks its own evidence is worth — a 5-point haircut today — but does NOT apply it to the confidence above, because that judgment has no graded record yet. It earns the right to move a number the same way every other signal does: by being measured first. Chain-implied spot is put-call parity on the nearest at-the-money strike with two-sided quotes on both legs, discounting ignored (under 0.02% at these tenors). It cross-checks two independent free feeds; it does not certify either one. Tape drift measures movement since the chain's own timestamp, so it bounds how stale the pricing is, not which direction is right.
Curated by hand, not a live feed, so it stays consistent with the forward-test timeline. Update as dates firm up.
Data provenance
Prices and VIX from Yahoo public chart data. Option chain and Greeks from CBOE delayed quotes (feed time 2026-08-03 13:58:20). Both free, both about 15 minutes delayed.
Generated 2026-08-03 9:59 AM ET. Schema v3.
When it sat out: 2 days — 2 up, 0 down, 0 flat, average move 0.25% either way. Abstention is a graded decision, not a blank.
Derived at read time from the append-only, hash-stamped session records — the report cannot drift from the evidence, and no slice speaks below its sample bar. Weights change only through evidence-stamped recalibrations on System Health, never from a single day.
