How Apollo works

Apollo is an analysis assistant for short-term options on two ETFs: QQQ (Nasdaq-100, the focus ticker) and SPY (S&P 500). Before each session it publishes a plan of specific contracts:

• 0DTE plays: options that expire the same day, built for quick intraday moves.
• About 3-day plays: options a few days out, which give a setup more time to work and lose value to time decay more slowly.

Apollo is unbiased between calls and puts. Calls (bets on a rise) and puts (bets on a fall) are scored the same way, and Apollo picks whichever side the data qualifies. It doesn't lean bullish or bearish and doesn't trade on hunches.

Every play gets one number: Apollo's estimated probability that the price reaches the quick (scalp) target before it hits the stop, once the entry trigger has fired. The score comes from data, not gut feel:

1. Probability model (the anchor). A Monte Carlo / barrier calculation uses current implied volatility and the real distances, in points, to the target and the stop. This base rate usually lands between 35% and 65%.
2. Adjustments, each worth roughly −10 to +10 points, capped at ±25 in total:
   • Timeframe agreement: how many of the 1m, 5m, 15m, 1h, 2h and daily trends point the same way as the trade.
   • Key levels: how strong the trigger level is (confluence of VWAP, pivots, prior-day high/low, volume-profile levels, moving averages) and how much room there is before the next obstacle.
   • Options positioning: the dealer-gamma regime (negative gamma tends to extend moves; positive gamma tends to pin price) plus gamma flip points and call/put walls.
   • Flow and sentiment: put/call volume, VIX direction, breadth and volume on the trigger bar.
   • Catalyst and time risk: upcoming data or news, time of day (lunch chop, last-hour time decay) and spread width.
3. Calibration. Once there's a graded history, if picks in one score range won less often than their scores said, scores in that range get pulled toward 50%. Every change is logged in public.
4. Capped at 90%. Short-term options are never a sure thing, so Apollo never claims more than 90%.

Each play shows its score breakdown in one line, plus a separate (lower) probability for reaching the extended target.

After every close, each pick Apollo called that day is checked against real price data, with losses graded in exactly the same detail as wins:

• Did the trigger fire, and when? Was the quick target, the extended target or the stop hit first, and at what time?
• How far the option moved in its favour (max favourable move) and against it (max drawdown), how many minutes it traded above entry, and its % change 15, 30, 60 and 120 minutes after entry.
• The setting at entry: VWAP side, gamma regime, VIX, time of day, timeframe agreement and volume, plus the best possible entry and exit in hindsight.
• A result (win, loss, scratch if neither target nor stop was hit, or no trigger) and a short lesson.

The results feed the public Track Record, including win rate by confidence band, so you can check whether 70%+ picks really beat 55–69% picks. Ungraded picks show "Pending". Nothing is estimated or backfilled.