LIVE·WAITING…

fetching live quotes from Yahoo Finance…

--:--:--UTC

Learn/Bots/IV Solver (NN)

AI QuantsAIid · ai-iv-solver

IV Solver (NN)

Implied vol from observed option price — no Newton-Raphson needed.

▶ Try it now ↗ Browse all 27srcai quants/models/implied_vol/train.pyapi/api/iv

⟢In plain English

When you see an option's price, you can run the math backwards to figure out what volatility the market is implying. The classic way (Newton-Raphson) is finicky and breaks near zero. Our AI does it in a single forward pass — fast and stable.

No jargon. Just what this bot does.

⟢The longer version

Reverse the BS pricer. You have an option's market price and need to back out the implied volatility — what σ makes the formula spit out that price? Classically done with Newton-Raphson, which is fragile near zero vega and can take 20+ iterations. The NN does it in a single forward pass with ~0.9% relative error.

⟢The math

formula

inverse BS surrogate · ≈0.9% relative error

parameters

typeTypechoices: C, Pdefault · C

spotSpot $range — → —default · 100

strikeStrike $range — → —default · 100

daysDays to expiryrange — → —default · 30

rateRate %range — → —default · 4.5

priceObserved $range 0.01 → —default · 4.5

⟢Live demo

Real IV Solver (NN) bot, running on real Yahoo data when the symbol is available. Drag the params — the bot re-runs instantly.

symbolloading…

loading AMZN bars…

⟢Source code · public

This is the actual code the bot runs — not a re-explanation, not a simplified version. Whatever ships here is what executes when you press Run All in the workbench. Read it, copy it, fork it, build a better one.

●lib/quant/ai-bots.ts·lines 234–321

TypeScript · MIT-licensed

const ivSolver: BotDef = aiBot<IVReq, IVRes>(
  {
    id: "ai-iv-solver",
    name: "IV Solver (NN)",
    category: "ai",
    glyph: "ν",
    tagline: "Implied vol from observed option price — no Newton-Raphson needed.",
    formula: "inverse BS surrogate · ≈0.9% relative error",
    endpoint: "/api/iv",
    module: "ai quants/models/implied_vol/train.py",
    params: [
      { key: "type", label: "Type", kind: "select", default: "C", options: [{ value: "C", label: "Call" }, { value: "P", label: "Put" }] },
      { key: "spot", label: "Spot $", kind: "number", default: 100, step: 0.5 },
      { key: "strike", label: "Strike $", kind: "number", default: 100, step: 0.5 },
      { key: "days", label: "Days to expiry", kind: "number", default: 30, step: 1 },
      { key: "rate", label: "Rate %", kind: "number", default: 4.5, step: 0.1 },
      { key: "price", label: "Observed $", kind: "number", default: 4.5, min: 0.01, step: 0.01, hint: "the option's actual mid price" },
    ],
  },
  {
    request: (ctx, p) => {
      const lastPx = ctx.candles[ctx.candles.length - 1]?.c ?? 100;
      const type = (str(p, "type", "C") === "P" ? "p" : "c") as "c" | "p";
      return {
        price: num(p, "price", 4.5),
        S: num(p, "spot", lastPx),
        K: num(p, "strike", Math.round(lastPx)),
        T: Math.max(0.0005, num(p, "days", 30) / 365),
        r: num(p, "rate", 4.5) / 100,
        flag: type,
      };
    },
    build: (data, ctx, p) => {
      const lastPx = ctx.candles[ctx.candles.length - 1]?.c ?? 100;
      const type = (str(p, "type", "C") === "P" ? "P" : "C") as "C" | "P";
      const spot = num(p, "spot", lastPx);
      const strike = num(p, "strike", Math.round(lastPx));
      const t = Math.max(0.0005, num(p, "days", 30) / 365);
      const rate = num(p, "rate", 4.5) / 100;
      const greeks = priceOption(spot, strike, t, rate, data.sigma, type).greeks;
      return {
        signals: [],
        metrics: [
          { key: "iv", label: "IV", value: `${(data.sigma * 100).toFixed(2)}%`, tone: "info" },
          { key: "vega", label: "ν Vega", value: fmtNum(greeks.vega, 3) },
          { key: "iters", label: "Iterations", value: "1 (NN)", tone: "neutral", hint: "vs ~25 for Newton-Raphson" },
          { key: "err", label: "Surrogate err", value: "≈0.9%", tone: "info" },
        ],
        summary: `Solved IV = ${(data.sigma * 100).toFixed(2)}% in a single forward pass. The python NN replaces Newton-Raphson, which usually takes 20+ iterations and explodes near zero vega.`,
        beginner:
          "Reverse the option pricer — given the price the market is showing, what volatility makes the formula match? The neural net does it in one inference.",
        verdict: { side: "hold", text: `Implied vol ${(data.sigma * 100).toFixed(2)}%.`, confidence: 1 },
      };
    },
    mock: (ctx, p) => {
      const lastPx = ctx.candles[ctx.candles.length - 1]?.c ?? 100;
      const type = (str(p, "type", "C") === "P" ? "P" : "C") as "C" | "P";
      const spot = num(p, "spot", lastPx);
      const strike = num(p, "strike", Math.round(lastPx));
      const days = num(p, "days", 30);
      const rate = num(p, "rate", 4.5) / 100;
      const obs = num(p, "price", 4.5);
      const t = Math.max(0.0005, days / 365);
      let lo = 0.01, hi = 3.0, mid = 0;
      for (let i = 0; i < 50; i++) {
        mid = (lo + hi) / 2;
        const guess = priceOption(spot, strike, t, rate, mid, type).price;
        if (guess > obs) hi = mid;
        else lo = mid;
      }
      const sigma = mid;
      const greeks = priceOption(spot, strike, t, rate, sigma, type).greeks;
      return {
        signals: [],
        metrics: [
          { key: "iv", label: "IV", value: `${(sigma * 100).toFixed(2)}%`, tone: "info" },
          { key: "vega", label: "ν Vega", value: fmtNum(greeks.vega, 3) },
          { key: "iters", label: "Iterations", value: "50 (bisection)", tone: "neutral", hint: "TS fallback" },
          { key: "err", label: "Surrogate err", value: "≈0.9%", tone: "info" },
        ],
        summary: `Solved IV = ${(sigma * 100).toFixed(2)}% via TS bisection (NN unavailable).`,
        beginner:
          "Reverse the option pricer — given the price the market is showing, what volatility makes the formula match?",
        verdict: { side: "hold", text: `Implied vol ${(sigma * 100).toFixed(2)}%.`, confidence: 1 },
      };
    },
  },
);

what each piece means

id — unique key the workbench uses to find the bot.
params — the sliders + inputs you see on the cell.
run(ctx, p) — the function that gets called with candles + your params and returns the verdict.
verdict — the BUY / SELL / HOLD pill at the top of the cell.
metrics — the small stat boxes shown in the cell body.

use this code yourself

Copy the whole block above.
On /quant, click + Import your bot in the bot library.
Paste, hit save. It hot-loads into your workspace.
Edit any param defaults or logic to your taste — it's now yours.

⟢Specialty · when it shines, when it fails

✓ Shines when

·Real-time IV updates as the chain ticks. Newton's method has variable latency depending on starting point; the NN is constant-time.
·Building IV surfaces. Need to solve for σ across hundreds of strikes/expiries simultaneously — batch inference shines here.
·Robustness near low vega. Newton blows up when ∂Price/∂σ is small; the NN gracefully degrades instead.

✗ Fails when

·Out-of-distribution. The training set covered 5-200% vol; if the option price implies σ outside that, predictions degrade.
·Pricing arbitrage edge cases. When the market price violates put-call parity, no σ exists — the NN just returns its best guess instead of erroring.
·Deep ITM/OTM at near-zero time-to-expiry. Both Newton and the NN struggle; you need a different algorithm entirely.

⟢How to read its verdict

HOLD always — solver, not a trade signal. The IV it returns is the trade input. Use it as the σ input to the Wheel bot, the BS Solver, or your own pricing logic.

⟢Python service

This bot tries to call the FastAPI service first. When it's up, you get real model output. When it's down, the bot transparently falls back to a deterministic TS surrogate.

FastAPI·http://localhost:8000·/api/ivCHECKING…

data flow

BotCell.run()

User clicks Run on this bot in /quant

callApi()

POST to localhost:8000/api/iv

load_surrogate()

ai quants/models/implied_vol/train.py

predict()

Forward pass on the inputs you provided

BotResult

JSON returned, card flips green Source: Python NN

spin it upcd "ai quants" && uvicorn serve:app --reload --port 8000

⟢FAQ

Why is the answer slightly different from a Newton-Raphson run?+

Newton finds the exact σ to floating-point precision. The NN is an approximation with ~0.9% mean relative error. For most use cases this is well inside bid-ask noise; for arbitrage strategies you'd still want the exact solver.

Can I trust this when the option is far OTM?+

Below an option price of ~0.05 the function gets flat and any solver becomes unstable. The NN is no exception — its predictions become noisy at deep OTM.

⟢Related bots

𝛣

BS Surrogate (NN)

AI Quants

Neural Black-Scholes — price + 5 Greeks in one shot.

Black-Scholes Solver

Options

Pure math option pricer with all the Greeks.

💥

IV Crush Detector

Options

Quantify the post-event vol drop on long options.