Options for Botty - the second carry: selling volatility like insurance (measured, with visualizations)

What this is about

With the funding carry we have our first living structural edge: collect a structural payment directionlessly instead of guessing where BTC will go. The obvious question: does the same work with options? Short answer: yes, the direct analog exists, it is even larger than the funding carry - but it has a catch, and there is a hard infrastructure wall. This article explains both from the ground up, in plain language, and shows what we actually measured.

1. What an option even is

An option is a right, not an obligation. There are two kinds:

• Call = the right to buy something at a fixed price (the strike).
• Put = the right to sell at a fixed price.

Whoever buys this right pays a premium for it. Whoever sells it (writes it) collects the premium - and in return takes on the obligation to deliver if the buyer exercises their right.

PROFIT of a CALL BUYER at expiry
  ^
  |                      /
  |                     /      the higher the price above the strike,
0 -+--------------*----/----->  the more profit
  |   -Premium ___|Strike      Price at expiry
  v (max loss = premium paid)

The most important intuition: an option is insurance.

   BUYER of the option           SELLER of the option
   = buys insurance              = sells insurance
   pays a small premium          collects a small premium
   wins big in the crash         earns steadily & small
   (loss capped)                 BUT pays big in the crash

And just as an insurance company makes money in the long run (premiums are on average higher than claims), the option seller can collect a structural premium. That is the key to the whole thing.

2. The funding-carry idea, transferred to options

With the Funding carry at Botty — the first structural edge, and what really defines it live we collect the funding - a payment the market structurally owes. With options the corresponding structural payment is called the variance risk premium (VRP), and it comes from a simple fact:

The options market on average prices in more movement than then actually occurs.

The priced-in movement is called implied vol (IV) - for BTC measurable via the Deribit index DVOL. What actually occurs is called realized vol (RV). The gap between them is the premium:

Vol (annualized, %)
  ^
62| ..................  implied vol IV  - what is priced in
  | ..................          ^
53| ##################          | gap = VRP ~ +8.8 vol points
  | ##################          v   (positive on 73% of days)
  | realized vol RV - what actually happens
  +------------------------------>  Time

Why is the gap there? For the same reason as with any insurance: most people happily pay a small premium to hedge against a big crash. This demand for protection pushes the priced-in vol durably above what occurs on average. Whoever sells the protection collects the difference - as payment for carrying the crash risk.

3. Why "selling vol" is the direct carry analog

To collect the premium directionlessly (just as the funding carry is delta-neutral), you sell a combination of a call and a put at once - a straddle (both at the same strike) or a strangle (put below, call above the price):

PROFIT of a SHORT STRADDLE at expiry
  ^
 +|  ___ premium collected ___
0 -+-\--------------------/->  Price at expiry
  |   \                  /      <- if the price stays calm: small profit
 -|    \                /
  |     \              /        <- if the price breaks out sharply: big loss
  v      \____________/
        "collecting pennies in front of the steamroller"

That is the whole bet in one picture: often a small profit, rarely a big loss. As long as BTC stays calm (RV < IV), you keep the premium. When BTC explodes (RV >> IV), you pay up - potentially a great deal. This is called negative skew: the exact opposite of a lottery ticket.

The decisive difference from the funding carry: the funding carry has a capped downside (funding flips, you exit, small losses). Naked vol selling has a fat left tail - a single steamroller can wipe out half a year of premiums.

4. Delta-neutral: taking the direction out

A short straddle on its own still reacts to two things: the direction and the vol. But we only want the vol premium, not to bet on direction. The solution is called delta-hedging: you continuously hold an offsetting position in the perp (or spot) that neutralizes the directional sensitivity (the delta).

   Short straddle alone      reacts to  DIRECTION + VOL   (messy)
        +  hedge in the perp   (buys/sells against the direction)
        ----------------------------------------------------------
        =  only VOL left       (clean, directionless)  <- delta-neutral

If BTC rises, the straddle's delta turns negative -> we buy a bit of perp against it. If BTC falls, the reverse. What remains is the pure "implied vs. realized vol" bet - the clean carry. The hedge leg runs on a perp, which we already have (Hyperliquid).

5. Defined-risk: fencing in the steamroller (iron condor)

Naked vol selling with an unbounded tail is irresponsible. The solution: you buy cheap, far-out options as protective wings on top. The naked strangle becomes an iron condor - you give up some premium and in return cap the maximum loss:

PROFIT of an IRON CONDOR at expiry
  ^
  |      +---------------+   <- max profit (net premium),
0 -+-----+               +----->  as long as the price stays in the corridor
  |    /                   \
 -| __/                     \__   <- loss CAPPED by the
  v   |wing            wing|       purchased protective options
   purchased wing | purchased wing

Instead of "-74 vol points in a crash", the loss is now firmly bounded. This is exactly the structure we measured.

6. What we actually measured - the four scouts

Instead of speculating, we worked the question through in four steps - all on our existing 5 years of DVOL + BTC data, without a single options trade. Each scout answers one question and hands it to the next:

 +-----------------------------------------------------------------------+
 | 1. HARVEST       Does the premium exist?     YES Sharpe 0.89, tail -74|
 |      |           (delta-hedged var swap)         (large, but roller)  |
 |      v                                                                |
 | 2. DEFINED-RISK  Does it survive UNHEDGED     NO  Sharpe 0.26         |
 |      |           with a capped tail?              (direction dominates)|
 |      v                                                                |
 | 3. HEDGED        Does daily delta-hedging     YES Sharpe 1.5-1.9      |
 |      |           rescue the edge?                 tail only -4%       |
 |      v                                                                |
 | 4. GAP-STRESS    Does it survive the real     YES at 4h: Sharpe 1.4   |
 |                  intraday gaps (1m path)?         tail -5.5%, robust  |
 +-----------------------------------------------------------------------+

What the chain teaches: the premium is real and large (step 1). But a simply sold structure held to expiry does NOT collect it - the directional endpoint risk eats the edge (step 2: 79% hit rate, still nearly zero). The delta-hedge is the key: it strips out the direction and lifts the Sharpe to 1.5-1.9, while capping the tail on the side to ~-4% (step 3). And this even survives intraday realism: at 4-hour hedging (operationally trivial, fits our bot cadence) the Sharpe stays ~1.4 and the worst case at ~-5.5% (step 4).

Hedge frequency vs. Sharpe (intraday stress test, condor 1s/2s)
 Sharpe
 1.5|        * 4h  <- sweet spot
    |   *daily * 1h
 1.0|                * 15m
 0.5|
 0.0|
-0.5|                       * 1m  <- costs + noise hedging eat everything
    +------------------------------>  hedge more often ->

The collapse at 1-minute hedging is not a risk signal but a cost artifact: 43,000 rebalances per roll plus microstructure noise (bid-ask flutter) over-hedge on noise. Lesson: do not hedge too often. The real sweet spot is 4h-1h.

7. What else options open up for us

The VRP harvest is only one of five doors that options push open:

8. The hard wall: the venue gate

Here the pretty theory ends at reality:

• Hyperliquid has no native options. Crypto options = practically only Deribit (BTC/ETH, ~85% of the volume).
• We already read Deribit (the DVOL index, without a login) - but have no trading client, no keys.
• A Deribit execution layer plus Greeks management (continuously monitoring delta/gamma/vega/theta, expiry dates, margin, assignment) is a larger build than the deferred cross-venue carry. The options leg needs Deribit; the hedge leg would run on an HL perp (which we have).

9. An honest conclusion

The VRP harvest is the first structural edge after the funding carry to pass all gates - including intraday realism. Unlike scalping (died at the fee and adverse-selection floor, see Scalping & Market-Making for Botty - why both variants die (fee floor vs. adverse-selection floor)), this is not a research kill but a build decision: the edge is statistically real, blocked only by missing infrastructure.

We keep the residual uncertainty honestly open: we computed with flat IV (real crypto options have a put skew), DVOL is an index rather than a tradable instrument (real spreads are wider, especially on the wings), and n=62 rolls is a moderate sample. Before any capital ever flows, the scout would need to be repeated with real, skew-aware Deribit quotes - that is the one model gap that could still move the result.

But the direction holds: options are not a magic trick but a second structural carry axis - present, measurably attractive, and honestly held back only by the question of whether a Deribit client is worth the effort for a Sharpe-1.4 edge.

Minimum capital outlay - the Deribit floor (2026-06-20)

An important practical difference from the Funding carry at Botty — the first structural edge, and what really defines it live: the funding carry can be run arbitrarily small (even $60, as live on wallet 3). The options carry has a hard capital floor - not because of our strategy, but because of Deribit's minimum contract size. Smaller than one contract is not possible.

ETH is ~3x smaller and thus the cheapest entry for a first real trade - the VRP mechanics are identical on ETH (Deribit also has an ETH DVOL).

Two important clarifications

1. The smoke does not risk the max loss. An open-and-immediate-close (a mechanics test) only costs the spread (~$15-30), not the max loss. The capped max loss (~$360 ETH / ~$660 BTC) only bites once you hold the condor over days. Defined-risk means: you can structurally never lose more than this amount, no matter what the market does.

2. The hedge leg needs almost nothing. The delta-neutralizing HL perp hedge is tiny at open (~$65 notional); an HL wallet with ~$100 is enough for a smoke. For a sustained run it should stand at ~$300-500, because the hedge delta grows via gamma as the price approaches a short strike.

Framing

The floor of ~$430 (ETH) or ~$950 (BTC) is the honest entry threshold of this strategy. That is considerably more than the funding carry needs, but structurally unavoidable - it is the price of the fact that crypto options are only tradable on Deribit with a fixed minimum size. Whoever wants the options carry must accept this threshold; a 'mini-mini test' like with the funding carry ($60) is physically not possible here.