Free vs Premium Tier

What free users get, what premium unlocks, and why the free tier is still genuinely useful.

Tier Overview

Tier Price Who
Free $0 All new users
Waitlist $29/mo (locked for life) Early adopters who joined the waitlist
Paid $49/mo Standard subscription

Waitlist and paid tiers are functionally identical -- both are "premium." The only difference is price.

What Free Users Get

Free tier is not a crippled demo. It uses production-grade algorithms that cover the majority of retail Iron Condor trading:

  • SVI vol surface -- Stochastic Volatility Inspired model for smile fitting
  • Lognormal PoP -- probability of profit from the Black-Scholes framework
  • Yang-Zhang volatility -- OHLC-based estimator (more accurate than close-to-close)
  • CRR binomial tree -- Cox-Ross-Rubinstein option pricing
  • Bootstrap simulation -- non-parametric resampling from real historical returns
  • 1 OOS seed -- basic sanity check against overfitting

These are the same models used by most retail options platforms. For liquid ETFs like SPY, QQQ, and IWM, the free tier produces actionable results.

What Premium Unlocks

Premium provides advanced algorithms, more simulation modes, and deeper validation:

Algorithms

Component Free Premium
PoP model Lognormal Cornish-Fisher, Johnson SU, EVT
Vol estimator Yang-Zhang HAR-RV (multi-horizon realized vol)
Vol surface SVI (single slice) SSVI (arbitrage-free full surface)
Option pricer CRR (binomial) QuantLib, Heston, SABR
Entry scorer EV (expected value) Sortino (downside-risk weighted)

Simulation

Mode Free Premium Description
ev Default Available Deterministic expected value
bootstrap Available Available Block-bootstrap from historical returns
mc -- Available GBM Monte Carlo (log-normal paths)
garch -- Available GARCH-family parametric simulation (7 variants)

GARCH Variants (premium only)

Variant Description
gjr GJR-GARCH with leverage effect (default)
egarch Nelson log-vol, guarantees positivity
ms Markov-Switching GJR-GARCH with per-regime parameters
fhs Filtered Historical Simulation (GARCH + bootstrapped residuals)
vg Variance Gamma (fat-tail Levy process)
mjd Merton Jump-Diffusion (GBM + compound Poisson jumps)
tail Empirical Tail Sampling (worst-case return windows)

Feature Gates (premium only)

Gate Description
EVT tails GPD tail adjustment replaces fixed pop_discount
Stochastic IV Per-path IV shock sampling from historical distribution
Seasonal IV Monthly/quarterly IV seasonality adjustment
HMM regime Hidden Markov Model regime detection
Heston barriers Stochastic-volatility barrier breach pricing
VIX correlation SPY-VIX correlation for correlated IV shocks

OOS Validation

Aspect Free Premium
Seeds 1 3-10 (configurable in preset)
Simulation Bootstrap Full GARCH MC (same mode as in-sample)
Overfit gap Single-seed (noisier) Median over N seeds (robust)

Parameter Access

Free tier -- locked parameters

All algorithm and parameter choices come from the user's preset. These cannot be overridden:

Parameter Free value
pop_model lognormal
vol_estimator yang_zhang
vol_surface_model svi
option_pricer crr
simulation_mode bootstrap
entry_scorer ev
oos_seeds 1
Delta ranges From preset
All feature gates Off

Free users control only: ticker, max_capital, and top_k.

Premium tier -- overridable parameters

Premium starts with preset defaults but allows overrides:

Parameter Override range
pop_model lognormal, cornish_fisher, johnson_su, evt
vol_estimator yang_zhang, parkinson, realized_kernel, har_rv
vol_surface_model svi, ssvi, vanna_volga
option_pricer crr, quantlib, heston, sabr
simulation_mode ev, bootstrap, mc, garch
garch_model gjr, egarch, ms, fhs, vg, mjd, tail
dte_min, dte_max 1-90
Delta ranges Configurable
iv_rank_gate_mode soft, hard
iv_rank_threshold 0.0-1.0

Quotas

Tool Free Premium
create_strategy Limited daily calls Higher or unlimited
find_iron_condors Limited daily calls Higher or unlimited

Check remaining quota with get_quota_remaining(tool_name) before calling expensive tools.

Response Metadata

Every create_strategy and find_iron_condors response includes metadata showing what algorithms were actually used:

Free tier response:

{
  "tier": "free",
  "metadata": {
    "pop_model": "lognormal",
    "vol_estimator": "yang_zhang",
    "vol_surface_model": "svi",
    "option_pricer": "crr",
    "simulation_mode": "bootstrap",
    "oos_seeds": 1
  }
}

Premium with overrides:

{
  "tier": "premium",
  "metadata": {
    "pop_model": "cornish_fisher",
    "simulation_mode": "garch",
    "garch_model": "ms",
    "oos_seeds": 5
  },
  "overrides_applied": {
    "pop_model": "cornish_fisher",
    "simulation_mode": "garch",
    "garch_model": "ms",
    "oos_seeds": 5
  }
}

Why the Free Tier is Still Useful

The free tier covers the standard retail options workflow:

  1. SVI + lognormal is the same PoP calculation used by most brokerage platforms. For liquid ETFs, it produces reliable estimates.

  2. Bootstrap simulation uses actual historical returns (no parametric assumptions). It naturally captures the return distribution including fat tails, just without explicit modeling.

  3. 1 OOS seed catches obvious overfitting. While not statistically rigorous, it flags strategies where in-sample performance diverges wildly from out-of-sample.

  4. Full optimizer pipeline -- free users get the same Pareto optimization, the same multi-objective search, the same result structure. The difference is in the underlying models, not the optimization framework.

The natural upgrade path: "Your free search shows 73% probability of profit with one bootstrap validation. Want to see what the Cornish-Fisher model with 5 Monte Carlo seeds says?"

Error Handling for Tier Access

When a free user passes a premium parameter, the platform returns a TIER_BLOCKED error:

{
  "error": "TIER_BLOCKED",
  "message": "Parameters pop_model require premium (current tier: free).",
  "suggestion": "Upgrade to premium for access to this parameter.",
  "tier_upgrade_hint": true
}

Agents should explain the restriction and offer the upgrade path, not silently drop the parameter.

Cross-references