Spot on regarding expired contracts and survivorship bias — if you can't query NVDA240315C00850000 or SPY220617P00380000 after expiration day, you can't accurately backtest short-premium roll rules, earnings IV crush, or tail-risk hedges across past regimes. That is why /v1/options/history keeps expired OCC contracts addressable from 2019 forward with daily OHLCV, open_interest, implied_volatility, and Black-Scholes Greeks recomputed at each historical session's exact remaining time-to-expiry (T).

On your two questions:

1. Crumb / session breakages disappear entirely: Unofficial scrapers break mid-run because they have to negotiate browser cookies and a short-lived CSRF crumb token tied to an HTML session — whenever an edge node invalidates that cookie jar or changes the consent redirect, the entire run fails. With a dedicated API endpoint (api.apextick.com), auth is a stateless apikey header over standard HTTPS, so there is no session state or cookie handshake to expire mid-run.
2. Client-side retry & partial-batch isolation: A single bad or delisted ticker inside a 50-symbol batch call (/v1/market/quotes?symbols=...) does not fail the HTTP request — all valid symbols still return in data[] so a 500-ticker screener (10 batch calls) never aborts halfway through over one symbol. On the client side, the only retry logic you want for unattended production runs is a persistent requests.Session() (to reuse the TLS connection across batches) with a lightweight transport retry for transient network blips:

import requests
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry

session = requests.Session()
session.headers.update({"apikey": "YOUR_API_KEY"})
retry = Retry(total=3, backoff_factor=0.4, status_forcelist=[429, 502, 503, 504])
session.mount("https://", HTTPAdapter(max_retries=retry))

That combination (stateless header auth + connection pooling + 50-ticker batch isolation) makes a 500-symbol screener finish deterministically in ~1–2 seconds without mid-run auth failures.