# The skew matters more than the hedge decision

> One assumed parameter moves a hedging result by more than the gap between hedging and not hedging. We measured it instead, and it is not one number.

- Source: https://defiloops.com/blog/the-skew-matters-more-than-the-hedge-decision
- Published: 2026-09-16
- Category: Architecture
- Tags: options, hedging, volatility, data
- Author: DeFiLoops

---
If you ask what a put costs, the answer depends on a parameter almost nobody states: the
**skew**, which is how much extra implied volatility an out-of-the-money strike carries above
the at-the-money index.

It is usually assumed. A model takes the volatility index, applies a constant, and returns a
price. That constant moves the result by **more than the difference between hedging and not
hedging** — which makes it the single most consequential number in the answer, and the one
least likely to be shown to you.

## So we measured it

Against a live chain rather than a textbook. Two findings, both inconvenient for a constant.

**It is not one number across assets.** A 15% out-of-the-money put carried **17%** more
volatility than at-the-money on BTC, and **8%** on ETH. Those differ by a factor of two, so
each asset now uses its own.

**It is not one number across tenors either.** On BTC the markup decays steeply with time to
expiry:

```
    8 days    38.6%
  288 days     3.4%
```

<Callout type="warn" title="Which means a single skew is wrong twice">
  Apply one value across all tenors and you understate a short-dated option and overstate a
  long-dated one — in the same model run, in opposite directions. The error does not average
  out; it changes which structure looks best.
</Callout>

## The honest caveat on our own number

The measurement holds for **one day and one tenor**. It is a reading, not a constant we have
discovered, and the next reading will differ.

That is why every answer states the value it used rather than burying it. A hedging result you
cannot reproduce because you do not know what skew produced it is not a result, it is a number
with a confident font.

## What this is for

This sits in the read-only half of the system. The volatility history — Deribit's index back
to March 2021, plus option chains for strikes away from the money — exists so the agent can
answer "what would this have cost" with something better than an opinion.

**None of it trades.** Options are not reachable from here; a hedging step in a plan is an
instruction for you to carry out at a venue. The data's whole job is to make that instruction
specific enough to be worth following: the asset, put or call, the strike as a multiple of
spot, and how long.

<Callout type="note" title="And the finding people like least">
  Priced across every strike, spend and tenor since 2021, protection at the money has cost
  30–40% a year against the asset's own drift. That is the answer more often than any clever
  structure is, and a tool that will not tell you so is selling something.
</Callout>