Historical Max Pain Data API - Pin Probability, Pain Curves & Dealer Alignment Over Time

If you've spent any time in options content, you've seen the claim: "the market gravitates toward max pain on expiration." It's repeated by traders, blog posts, and YouTube channels. It's almost never tested at scale, because doing so requires historical max pain data - pain curve, max pain strike, pin probability - at minute resolution, joined to actual close prices across thousands of expiries. That dataset hasn't been buyable until now.

FlashAlpha's Historical API ships /v1/maxpain at every minute since 2018-04-16. Same shape as the live endpoint. This article walks the response, then runs the basic pin-rate test you've always wanted to see.

The Endpoint

GET https://historical.flashalpha.com/v1/maxpain/{symbol}?at=&expiration=

What Comes Back

{
  "symbol": "SPY",
  "underlying_price": 679.46,
  "as_of": "2026-03-05T15:30:00",
  "max_pain_strike": 684,
  "distance": { "absolute": 4.54, "percent": 0.67, "direction": "below" },
  "signal": "neutral",
  "expiration": "2026-03-05",
  "put_call_oi_ratio": 2.033,
  "pain_curve":    [ { "strike": ..., "call_pain": ..., "put_pain": ..., "total_pain": ... } ],
  "oi_by_strike":  [ { "strike": ..., "call_oi": ..., "put_oi": ..., "total_oi": ... } ],
  "max_pain_by_expiration": [ { "expiration": "...", "max_pain_strike": ..., "dte": ..., "total_oi": ... } ],
  "dealer_alignment": {
    "alignment": "moderate",
    "description": "...",
    "gamma_flip": 677.5,
    "call_wall": 685,
    "put_wall": 680
  },
  "regime": "positive_gamma",
  "expected_move": { "straddle_price": 1.57, "atm_iv": 19.6, "max_pain_within_expected_range": false },
  "pin_probability": 45
}

Five things to highlight in the response:

• max_pain_strike: the strike at which total option holder loss is maximized (i.e. dealer P&L is best).
• pain_curve: total pain at every strike - the full curve, not just the minimum point.
• pin_probability: a 0-100 score blending OI concentration, gamma alignment, expected-move bounds, and time-to-expiry. Not the same as max pain - it's the FlashAlpha pin model.
• dealer_alignment: cross-references max pain with gamma flip, call wall, put wall - gives you "alignment" labels (typically converging, moderate, diverging) for whether the dealer-positioning picture supports the pin or not.
• max_pain_by_expiration: max pain across every loaded expiry, not just the requested one. Use this for term-structure plots of the pin level.

The Test Everyone Wants to See

Does the close on expiration day actually pin to max pain? Run it.

import httpx, pandas as pd
from tqdm import tqdm

API_KEY = "..."
BASE = "https://historical.flashalpha.com"

# SPY weekly Friday expiries, 2024-2025
fridays = pd.date_range("2024-01-05", "2025-12-26", freq="W-FRI")

rows = []
with httpx.Client(headers={"X-Api-Key": API_KEY}, timeout=30) as c:
    for d in tqdm(fridays):
        # Max pain at 15:30 on expiration day, for that day's expiry
        mp = c.get(
            f"{BASE}/v1/maxpain/SPY",
            params={"at": d.strftime("%Y-%m-%dT15:30:00"),
                    "expiration": d.strftime("%Y-%m-%d")},
        )
        if mp.status_code != 200: continue
        m = mp.json()
        if not m.get("max_pain_strike"): continue
        # Close spot
        sq = c.get(f"{BASE}/v1/stockquote/SPY",
                   params={"at": d.strftime("%Y-%m-%dT16:00:00")})
        if sq.status_code != 200: continue
        rows.append({
            "date": d,
            "spot_1530": m["underlying_price"],
            "max_pain":  m["max_pain_strike"],
            "pin_prob":  m.get("pin_probability"),
            "alignment": m["dealer_alignment"]["alignment"],
            "spot_close": sq.json()["mid"],
        })

df = pd.DataFrame(rows)
df["close_dist_to_pain"]   = (df["spot_close"] - df["max_pain"]).abs()
df["close_dist_to_random"] = (df["spot_close"] - df["spot_1530"]).abs()

print(f"Median close-to-max-pain:  ${df['close_dist_to_pain'].median():.2f}")
print(f"Median close-to-15:30 spot: ${df['close_dist_to_random'].median():.2f}")
print(f"Pin success rate (close within $1 of max pain): {(df['close_dist_to_pain'] < 1).mean():.1%}")

Two summary numbers tell you whether the cliché holds. Median close_dist_to_pain versus median close_dist_to_random - if max pain is informative, the first should be smaller. Empirically, on SPY weeklies it's somewhere between modestly smaller and roughly equivalent - the "always pins to max pain" claim is overstated, but max pain is not noise either.

Stratifying by pin_prob tightens the test:

df["pin_bucket"] = pd.cut(df["pin_prob"], bins=[0, 30, 60, 100], labels=["low", "med", "high"])
print(df.groupby("pin_bucket")["close_dist_to_pain"].agg(["count", "median"]))

The high-pin bucket - when OI concentration, gamma alignment, and expected-move bounds all line up - typically shows materially tighter close-distance than low-pin. Which is the whole point of having a pin probability score: it tells you when the cliché holds and when it doesn't.

Pattern: Dealer Alignment as a Filter

The dealer_alignment block scores whether max pain agrees with gamma flip and the call/put walls. When alignment is converging, all four signals point at the same level - that's a real pin setup. When alignment is diverging, max pain disagrees with the dealer picture and the pin claim weakens significantly.

df_conv = df[df["alignment"] == "converging"]
df_div  = df[df["alignment"] == "diverging"]
print(f"Converging alignment ({len(df_conv)} expiries): "
      f"median close-to-pain ${df_conv['close_dist_to_pain'].median():.2f}")
print(f"Diverging alignment  ({len(df_div)} expiries):  "
      f"median close-to-pain ${df_div['close_dist_to_pain'].median():.2f}")

The converging subset typically shows materially tighter median close distance than the diverging subset on SPY weeklies - a measurable, exploitable signal that doesn't show up if you treat all expiries as equal.

Pattern: Max Pain Migration During the Week

Max pain isn't a static number - it shifts as OI builds and decays through the week. Track it intraday on a single expiry to see the migration:

from datetime import datetime, timedelta

# Trace the Friday Mar 5, 2026 expiry from Monday open through Friday close
expiry = "2026-03-06"
t = datetime(2026, 3, 2, 9, 30)
end = datetime(2026, 3, 6, 16, 0)
rows = []
with httpx.Client(headers={"X-Api-Key": API_KEY}, timeout=30) as c:
    while t <= end:
        if t.weekday() < 5:  # weekdays only
            r = c.get(f"{BASE}/v1/maxpain/SPY",
                      params={"at": t.strftime("%Y-%m-%dT%H:%M:%S"),
                              "expiration": expiry})
            if r.status_code == 200:
                m = r.json()
                rows.append({"t": t,
                             "spot": m["underlying_price"],
                             "max_pain": m["max_pain_strike"]})
        t += timedelta(hours=1)

trace = pd.DataFrame(rows).set_index("t")
trace[["spot", "max_pain"]].plot(title=f"SPY {expiry} expiry - max pain migration vs spot")

You'll see max pain step strike-by-strike as OI rebalances, and converge as the OI distribution stabilizes into the close. When max pain pulls toward spot in the final hours, pin probability typically rises; when it drifts away, the pin breaks.

Coverage and Caveats

Why max pain doesn't shift much intraday: open interest is end-of-day (one value per trade day). Max pain is a function of OI, so the strike itself stays constant within a trading day. What shifts intraday is the distance from spot to max pain (because spot moves) and the expected-move bounds (because IV moves). For multi-day max-pain dynamics, sample once per day; for distance and pin-context dynamics, sample intraday.

What This Replaces

• Computing pain curves yourself from raw historical chains → ships pre-computed.
• Building your own pin-probability score → use the FlashAlpha score directly, or refit on top of the response.
• Cross-referencing max pain with gamma flip and walls in custom logic → the dealer_alignment block does this for you.

Where you'd still build it yourself: bespoke pin models that disagree with the FlashAlpha methodology, or weighted-pain calculations (volume-weighted, vega-weighted) that aren't standard max-pain.

Related Articles

• Historical Options Analytics API - full overview
• Historical 0DTE Data - Pin Risk & Gamma Decay
• Historical Options Chain API
• Gamma Flip, Call Wall, Put Wall - Explained
• Backtesting GEX-Regime Strategies

Upgrade to Alpha API Spec