The SPY Options Desk

A bull put spread is defined risk premium selling. If SPY sits above the chosen strikes, we sell the higher put and buy the lower one. In one early April 2025 sketch we looked at selling the 543 put and buying the 538 put for a $0.24 credit. The spread is five dollars wide. On one contract the best case is $24. The worst case is $476. Scale that to six contracts and the numbers stop looking cute: $144 of possible profit against $2,856 of possible loss, with a breakeven at $542.76.

That payoff is the whole character of short premium. High probability of a small gain, and a much larger amount at risk if the tape is not normal. You can be right for weeks and still give a meaningful piece of it back in one session. Almost every rule we later wrote came from that fact.

SPY became the main 0DTE instrument because the options market is deep. Bid ask spreads are tight, which matters when you are hunting five, seven or ten cents of premium. There are many strikes and a daily expiry, so you can pick by delta instead of guessing a round number of dollars below the cash. The cost of that convenience is speed. SPY can move quickly, and on expiry day there is almost no time for a recovery.

The original thesis was modest. We did not need SPY to rally. Sideways was enough. A small decline that stayed above the short strike was enough. The problem was the small number of sessions when SPY did not behave like an average day.

One of those sessions arrived in June 2025. Headlines and tariff talk hit together. A spread that had looked comfortably out of the way became dangerous in a hurry. The lesson was not simply “pick strikes further away.” On 0DTE, probability itself changes extremely quickly. A put that looked like a small delta in the morning can pick up a large delta after a few dollars of cash move, because gamma explodes as expiry approaches. With weeks left, an option can live through a dip. With a few hours left, there may be no meaningful time to get well.

SPY can tag a level and reverse. If an algorithm treats every tick as an emergency, it will close healthy positions all day. If it waits too long, a real break becomes a large loss. Later versions asked for about 500 seconds of persistence, at least three confirming checks, a cooldown near 120 seconds, gentler treatment early in the trade, and a ban on new risk when too little time remained. In parts of the stack we tried an entry grace of about 7,200 seconds and an early delta multiplier near 1.25, so the first wobble was not treated as a crisis. The English version is simple: ignore noise, respect a genuine regime change. The code is not simple.

We also discovered a risk we had filed under administration. A bull put spread is defined risk only while both legs stay together. With Alpaca we saw cases where the broker liquidated one 0DTE leg near the close. In September 2026 it happened again with SPY sitting around the remaining strike and minutes left. Suddenly the book is not the spread you designed. Displayed P&L on the leftover leg can also mislead if you ignore the realised result on the leg already closed. Alpaca was attractive for API work. Interactive Brokers offered deeper multi-leg tools and rolling. Day trade limits, how each firm handles expiry, and how they liquidate all sit inside the strategy, not beside it.

QQQ is also liquid, with a deep options market and frequent expiries. It is more growth and technology than SPY, so it often moves more and cares more about rates and a handful of large names. That can produce fatter credits. Premium is never a gift. Higher premium is usually a market telling you it expects more travel.

We did not copy SPY strike distances onto QQQ. The honest comparison is volatility and delta. QQQ may need a wider cash distance to keep a similar probability. For a constructive view, a longer dated QQQ put spread can work if the index rallies, drifts, or even falls a little and still sits above the short strike at expiry. It can also travel much further than expected in a technology led risk off tape. In June 2026 we looked at a more aggressive 735 / 730 style structure. Moving closer always looks economically better on the quote. The extra credit is often small next to the extra chance of loss. Strike creep is how a 10 delta habit quietly becomes 15, then 20, until the market explains why the extra money was there.

A bull put spread is bullish. An iron condor tries to be more neutral: a put spread below the market and a call spread above it, collecting both credits if the underlying stays between the short strikes. A strong rally helps a put spread. It can hurt a condor. Two sided premium is compensation for two sided tails, and only if the extra credit is large enough and the tape is actually range bound.

An early example was a JPM condor into May 2025: sell the 222.5 put / buy the 215 put, and sell the 247.5 call / buy the 250 call, for about $0.76 net. The put side was $7.50 wide. The call side was $2.50 wide. Downside max loss was about $674 per condor. Upside max loss was about $174. The chain can look balanced while the money is not. Width, credit and risk have to be calculated, not inferred from how neat the strikes look.

By April 2026 we treated the condor as a secondary, not the engine. A two leg SPY bull put is simpler, cheaper to trade, easier to automate, and a rising market helps it. A four leg condor needs a genuine range. Both structures, and every DTE we used, still share a hidden fact: they are short volatility and short convexity. They like ordinary markets. They get stressed together when volatility jumps. Owning several credit strategies is not automatic diversification.

At one stage we sketched a risk budget of about 60 percent in 14 to 21 DTE SPY bull puts, 15 percent in 0DTE, and 25 percent in iron condors. That was a way of thinking, not a law of nature. The point was philosophical: daily tickets are not the same as quality. Win rate is the wrong headline. Ninety winners of $10 and ten losers of $150 is minus $600. The question is the whole distribution, especially the worst day, and whether several sleeves break at the same time.

Most 0DTE sessions are not disasters, which is why the strategy can feel comfortable. A 1 percent day is not ten copies of a 0.1 percent day. Option maths is nonlinear. Gamma rises, volatility can jump, liquidity can thin, fills on stops can disappoint, correlations can change. Average behaviour is a poor teacher for the tail.

We no longer think of SPY 0DTE bull puts as simply selling puts under the market. We think of a stack of probabilities: cash toward the short strike, implied volatility expanding, the move persisting, an economic event widening the day’s distribution, execution, broker liquidation, strategy correlation, and whether the whole book survives. You cannot delete those risks. You can get paid for them without letting one event erase a year of ordinary sessions.

Every improvement tempted us to believe the puzzle was solved: better deltas, better timing, better triggers, better calendars, better persistence, better profit taking, better regime work. Markets adapt. The aim is not a finished algorithm. It is a framework that can notice when its own assumptions have stopped working. Premium is pay for risk, not free income. If there is an edge that can be kept, it sits in when to sell, how much to take, when to get out, and when to stand aside.