Cash-secured puts: what they actually pay, and what they cost

Selling a cash-secured put is not income. It is being short volatility on a stock you have agreed to buy, funded by cash you have agreed not to use for anything else. Everything good and everything bad about the trade follows from that sentence.

The good part is that the odds are genuinely in your favour, most of the time, by construction. The bad part is that the payoff is bounded above by the credit and unbounded below by the stock, so the arithmetic that keeps you solvent is not the win rate — it is the ratio between a typical win and an atypical loss. We are going to compute both on a real chain.

The mechanics, in one table

You sell a put. You receive a credit. You set aside strike × 100 in cash. Three things can then happen:

A real ladder

Read the two end rows together. Moving from the $400 strike to the $450 strike quadruples the credit ($3.13 → $13.05) and nearly quadruples the annualized return (6.0% → 22.1%), while the probability of profit falls by only 17 percentage points (89.1% → 71.9%).

Stated that way, the far strike looks obviously wrong and the near strike looks obviously right. That intuition is a trap, and it is the single most expensive mistake in put selling.

Why the 89% strike is the dangerous one

The $400 put wins 89.1% of the time and pays $313. The $450 put wins 71.9% of the time and pays $1,305. Now ask what each one loses in the bad state.

MSFT’s implied volatility over these 48 days works out to a one-standard-deviation move of 10.7%, or $49.75. A two-sigma decline — unremarkable, roughly a once-every-couple-of-years event for a mega-cap, and something MSFT has done more than once — puts the stock at $375.15.

The $400 put — the safe one, the 89% one — needs 6.9 clean cycles to recover from one bad month. At 48 days a cycle, that is most of a year of being right, erased by one drawdown. The $450 put, which everyone will tell you is the aggressive choice, needs 4.7.

None of which makes the $450 strike correct. It makes the comparison correct: judge strikes by what a bad outcome costs measured in units of good outcomes, not by the win rate. Our builder prints the probability of profit and the max loss side by side for exactly this reason.

Annualized return is a rate, not a promise

The 16.9% on the $440 row means: if you could roll this trade, at this credit, at this volatility, with no losing cycles, for a year, you would make 16.9% on the collateral. Each of those conditions fails in a specific way:

• Volatility mean-reverts. MSFT is at 29.5% IV on this strike. Sell it during a quiet stretch at 18% and the credit roughly halves. The high-IV entries you would love to repeat are precisely the ones that come with the moves that hurt.
• The collateral is not idle. $44,000 in a settlement fund earns something — call it 4% at current short rates. Your excess return over cash is more like 12.9%, and that is the number to compare against buying the index.
• Assignment breaks the cadence. Once you own the shares you are no longer selling puts on that capital; you are selling covered calls on it, at whatever the chain then offers. The annualized figure quietly assumes a cycle you have stopped running.
• You will not always redeploy immediately. Real accounts sit in cash for days between cycles. A 10% gap in deployment is a 10% haircut on the rate.

Annualized return is the right tool for one job: comparing two candidate trades of different durations. It is the wrong tool for projecting a year, and if a platform shows it to you without showing peak capital and days alongside, it is flattering you.

Duration: 20 days or 48?

Same underlying, same strike, same $11,000 of collateral. Only the expiry changes.

The 48-day put costs 1.62× the 20-day put for 2.4× the time. That is the square-root-of-time law showing up in live quotes (√2.4 = 1.55; the extra comes from the front expiry’s higher implied volatility). Short duration wins on paper: roll the 20-day put eighteen times a year and you gross $5,530; roll the 48-day put 7.6 times and you gross $3,726.

Eighteen rolls also means eighteen bid-ask crossings, eighteen chances to be assigned at an inconvenient moment, and a position whose gamma makes the last week genuinely uncomfortable. The 72.6% front-month IV versus 60.8% for September is not a gift — it is the market pricing a known event inside that window. Check what it is before you sell it.

The same delta on four different stocks

Same probability of assignment. Three times the return from KO to PLTR — on 2.7 times the implied volatility (22.2% against 60.9%). The market is not mispricing anything here; it is charging you consistently for how far the stock can travel. The relationship is remarkably tight once you normalize by spot.

So the choice among these four is not a return choice. It is a question about which assignment you want to live with: 100 shares of Coca-Cola at $85, or 100 shares of PLTR at $115 after whatever move got it there. Pick the stock first. The chain will tell you what it pays.

Sizing, stated as a rule

The number to size against is not the credit and not the margin requirement. It is the cash you must have available if every put you have sold is assigned on the same day, because in the one market that matters they will be.

• Add up strike × 100 × contracts across every open short put. That is your obligation, and it is 100% correlated across positions in a crash.
• Five MSFT $440 puts is $220,000 of obligation for $4,900 of credit. If that sentence is uncomfortable, you are too big — the credit is the small number.
• Assume you get assigned everything and the stock is 20% below your breakeven. If the resulting position would force you to sell, the size is wrong regardless of how the probabilities look.

When not to sell the put

1. 1Earnings inside the expiry, unless being short that specific event is the trade. The IV you are collecting is the event premium; you are being paid, but you are being paid for that.
2. 2You do not have the cash. Selling a put on margin is a naked put with a friendlier name and a margin call attached.
3. 3The IV looks too good. Screen the reason. Going-concern doubt, a pending deal, a borrow squeeze — the chain is usually right and you are usually the last to know.
4. 4You would not buy the stock here. The entire structure rests on this. If the answer is no, the credit is not compensation, it is bait.

Run the numbers yourself

If you would rather cap the tail than carry it, the same MSFT $440 strike as a bull put spread — short the $440, long the $430 wing at $7.52 — shows the trade exactly: the credit drops from $980 to $228, and the worst case drops from $43,020 to $772. You keep 23% of the credit and 1.8% of the tail. The optimizer will enumerate the rest of the chain against a target price and date if you would rather rank candidates than eyeball a ladder.

When one of these gets assigned, the wheel tracker picks it up as a cycle rather than an orphaned loss — see the full walkthrough — and the MSFT strategy page keeps current chain context in one place.

Caveats

Next: how delta and IV actually pick the strike, or what happens after assignment.