**TL;DR:** Free Windows/Mac app for paper-trading options spreads against historical option chains. Fills cross the real bid/ask, earnings reactions are real, nothing is visible before its date, and each trade's P/L is broken down by Greek. It has a roguelite structure on top, mostly so you'll actually put in the reps. I'm looking for playtesters, ideally people who trade spreads and will tell me where the numbers are wrong.
**The problem it's trying to solve**
Most paper trading has two flaws. It tends to fill you at the mid, which is a generous assumption. And it runs in real time, so you get maybe a dozen trades a month to learn from. That's a slow way to find out how a 30-delta bull put behaves through an earnings print, or what IV crush does to a position you were sure about.
**What's under the hood**
* **Historical data:** end-of-day option chains from DoltHub's free options dataset (2019 onward), plus price history, earnings dates and reactions, Treasury rates, and VIX. Days the dataset skipped are filled with a model and labeled as modeled.
* **No lookahead:** every request for data goes through a time-gated layer that refuses anything dated after the simulated clock. IV rank, indicators, and fundamentals use only data on or before "today." You can see an earnings date ahead of time, but not the result.
* **Execution:** market orders pay the actual bid/ask spread. Early assignment, expiration, and ex-dividend risk on short calls are all modeled.
* **P/L attribution:** every trade is broken down day by day into direction (delta/gamma), time (theta), volatility (vega), execution, and fees. Whatever the Greeks don't explain is shown as a residual, not hidden, so the parts always add up to the actual P/L.
* **Structures:** cash-secured puts, covered calls, verticals, iron condors and flies, calendars, straddles. Payoff diagrams with date and IV sliders, POP and expected move.
* **Trade Builder:** set up a trade on today's market, analyze it, and copy it as a thinkorswim-style order line. Nothing is ever sent anywhere.
**The game part, briefly**
A run is 12 monthly rounds, about 20–40 minutes, with a shop between rounds and powerups that change your *score*. The rule is strict: the game layer can change your score, never the market. Each powerup is tagged REAL (mirrors an actual edge, like better fills inside the spread) or ARCADE (game-only). Holding a loser past your planned stop costs you, and taking the stop is rewarded. Many of us learned that lesson the expensive way.
**Caveats**
* It starts on a simulated market (made-up tickers with modeled option chains) so you can play right away. Real historical data is a one-time \~16 GB download from inside the app (needs \~40 GB free and a few hours).
* Older years in the free dataset are sparser (some only Mon/Wed/Fri). The gaps are modeled and labeled.
* The builds are unsigned, so Windows SmartScreen or macOS Gatekeeper will warn you once.
* It's paper trading and a game, not financial advice. You don't need a brokerage account.
**What I'd like to know from testers**
1. Do the Greeks, POP, credit, and breakevens match your broker's analysis tab for the same trade?
2. Do fills, assignment, and expiration behave the way you'd expect?
3. Is it useful for testing a strategy, or just entertaining? Either answer is fine, but I'd like to know which.
**Download (free):** [https://github.com/TTTrainer/Spread-Trade-Game/releases](https://github.com/TTTrainer/Spread-Trade-Game/releases)
Feedback here or in GitHub issues. Criticism of the math is especially welcome.
