I have seen many many of you lose thousands in a day.
I have seen many gain thousands in a day(and lose it all again the next).
Clearly many of you have the income to fund/cover those incredible losses. What is your primary income?
I have been trying to make short put credit spreads work as a real strategy. Built a backtest with one config tuned on SPY, then froze every parameter and ran it unchanged on QQQ, IWM, AMZN, NVDA from 2019 to 2026. Then added SPXW as an out-of-sample check using a separate data feed.
The point was to remove the three things that make most VRP backtests lie:
Here is what came out.
The full results table
| Symbol | Trades | Win rate | Sharpe | CAGR | Fill rate |
|---|---|---|---|---|---|
| SPY (tuned) | 335 | 71.6% | 0.23 | +1.05% | 13.7% |
| NVDA | 201 | 72.1% | 0.22 | +0.84% | 6.5% |
| AMZN | 139 | 71.2% | 0.13 | +0.42% | 3.4% |
| QQQ | 162 | 69.8% | -0.08 | -0.23% | 7.9% |
| IWM | 58 | 63.8% | -0.35 | -0.11% | 3.1% |
| SPXW (OOS) | 112 | 61.6% | -0.39 | -0.66% | 100% * |
* SPXW was daily-close resolution from a different feed, so its fill rate is 100% by design. Direction check only.
The win rate held everywhere. 64% to 72% across five very different underlyings. Two of four out-of-sample symbols lost money. SPXW lost money. The best Sharpe was on the symbol I tuned on. Every symbol the strategy had never seen did worse.
Why a 70% win rate makes nothing
The payoff is brutally asymmetric:
| Symbol | Avg win | Avg loss | Ratio |
|---|---|---|---|
| SPY | $369 | -$850 | 1 : 2.3 |
| QQQ | $311 | -$751 | 1 : 2.4 |
| AMZN | $407 | -$933 | 1 : 2.3 |
| NVDA | $385 | -$888 | 1 : 2.3 |
| IWM | $53 | -$130 | 1 : 2.4 |
0.70 × 1 - 0.30 × 2.3 = roughly zero. By the math, a 70% win rate at a 1:2.3 payoff is a coin flip. You win small often and lose big sometimes, and the two almost exactly cancel.
The win rate is the wrong number to anchor on. Every post that leads with "I have an 80% win rate strategy" and does not show the loss distribution is selling you the setup, not the result.
Only 3-14% of my orders filled, and the ones that did filled at worse prices
| Symbol | Proposed | Filled | Fill rate | Avg slip vs mid |
|---|---|---|---|---|
| SPY | 2,438 | 335 | 13.7% | -$0.037 |
| QQQ | 2,055 | 162 | 7.9% | -$0.042 |
| NVDA | 3,072 | 201 | 6.5% | -$0.044 |
| AMZN | 4,114 | 139 | 3.4% | -$0.040 |
| IWM | 1,896 | 58 | 3.1% | -$0.039 |
A backtest that assumes you fill at mid books all of these orders at a better price than reality gives you. Real life fills 3 to 14 percent of them and the ones that do fill cross at about 4 cents worse than mid, because the orders that actually fill are the ones the market is running through.
On 100-multiplier contracts that is a $4 headwind per fill before the trade even starts. Multiply by the fact that most posts are not even paying attention to fill rate and you can see where the fictional returns come from.
The exact fill logic I used (post limit at ask_edge, stale-quote guard, patient-then-cross exits) is open-source here: flashalpha-fill-simulator. Plug in your own fill model and rerun if you want to test the sensitivity yourself.
The profit target hits are mostly forced spread crosses
pt = my close-limit at the 50% target filled cleanly. pt_x = the target was hit but my limit did not catch, so I had to cross the spread to get out.
| Symbol | pt (clean) | pt_x (crossed) | sl | sl_x | expiry |
|---|---|---|---|---|---|
| SPY | 103 | 133 | 63 | 32 | 4 |
| QQQ | 29 | 76 | 37 | 12 | 8 |
| AMZN | 44 | 47 | 25 | 14 | 9 |
| NVDA | 68 | 71 | 43 | 13 | 6 |
| IWM | 12 | 21 | 7 | 13 | 5 |
On every single symbol, more profit-target exits required crossing the spread than filled cleanly at the limit. The "close at 50% target" that naive backtests book is the minority outcome in practice. This one effect, invisible in any mid-fill backtest, eats most of the gap between the win rate and the real result.
The high-conviction regime lost money
My signal outputs risk_on, neutral, reduce, or risk_off. risk_on means "deploy full size." Here is what each regime actually earned:
| Symbol | neutral P&L |
risk_on P&L |
|---|---|---|
| SPY | +$5,604 | +$2,174 |
| QQQ | -$3,000 | +$1,364 |
| NVDA | +$5,918 | +$266 |
| AMZN | +$4,669 | -$1,644 |
| IWM | +$145 | -$910 |
On four of five symbols, the regime my signal is most confident about earned less than the middle-of-the-road neutral state. Same pattern showed up independently on SPXW (95 risk_on trades lost $5,266; 17 neutral trades made $625).
That is the mechanism behind every "+5,000% backtest" VRP post that later blows up. The signal sizes you in hardest exactly when it shouldn't.
Where I'm landing
I went into this trying to find the version of put credit spreads that actually works. After the honest backtest, I cannot find one that beats passive SPY on a risk-adjusted basis once you pay realistic execution costs and refuse to refit per symbol.
So I am starting to think the defined-risk structure is the wrong call. With a spread you cap the win (the credit) and cap the loss (the long put you bought for protection), but the long put is so far out of the money in 20-30 delta land that it barely helps in the years that hurt, while it costs you real money in the years that don't.
I am now seriously thinking about going back to short straddles or strangles and managing the position by hand. Undefined risk, yes. But when the underlying moves you actually get to do something about it. Roll the untested side in, adjust the strike, flip to an iron when needed. With a spread you just sit there and watch the short leg get tested while the long leg does nothing useful until you are already at -80% of max.
Managing the position beats hoping a far OTM long put saves you, if you have the discipline to actually do it. The spread feels safer, but eight years of data says the safety was already priced in.
What I'm actually doing next
Strangles is the headline lead, but I'm not committing to it yet. There are bigger levers I haven't pulled, and I want to be honest with myself about the ranking.
Short-term plan: patient-execution variant first, sleeve-sizing math second, dumb VIX-floor signal third. Managed strangles is the test after that, not the headline pivot.
Caveats
TL;DR
The search for money God continues...
Bloomberg flagged a single put spread on Brent yesterday I can't stop thinking about. Someone bought the $91/$90 July puts for May 26 expiry — 134 million barrels notional. Pays out roughly $129M if Brent drops about 19% in a week.
That's not a hedge. Refiners hedge in 5-10M barrel chunks spread across the curve. six trading days out. Someone wanted exposure to a specific event in a specific window.
Then todays session, headlines soften on Iran and the broader market pumps. Brent doesn't crash, but the tail risk that spread was priced against suddenly looks a lot less likely.
The pattern I keep noticing in this cycle: every meaningful headline-driven move gets front-run by concentrated, oddly-specific flow a day or two prior. Hormuz tanker scare, Lebanon ceasefire pop, the tariff walk-backs. Each one had a print in the tape before the news hit. Sometimes it's a single block, sometimes it's a strike that lights up out of nowhere on a name that shouldn't be active.
I'm not making a claim about who's trading or why. It could be a macro fund with a good read, a desk hedging a position we can't see, or someone genuinely lucky. The point is the tape keeps telling you something before the headlines do, and the people who learn to read it keep ending up on the right side of these moves.
The actionable version of this isn't "front-run the news." It's "when you see concentrated short-dated flow that doesn't fit the current narrative, the narrative usually shifts within 48 hours." That's a watchlist trigger, not a trade signal.
Anyone else tracking this kind of flow systematically? Curious if there's a clean way to monitor unusual single-strike activity in commodities futures options without a terminal.
NKE at $44 is easy money. ~1.3x sales. Nobody's talking about it which is exactly when i want in.
down ~75% from the $170 highs, sitting near lows last seen in 2017. still a $65B household name, new CEO (Elliott Hill, who came out of retirement to fix it), and analyst PT consensus is $68 with 38 Buys vs 5 Sells. the 5Y chart looks like hell but shows the $42-48 zone has held as base support before.
looking at the $45C 7/17, currently ~$3.00. ATM, earnings ~5 weeks out on june 25 which gives the IV ramp leading in, plus 3 weeks of post-earnings runway to exit.
main risk is macro — iran, china, vietnam tariffs, hoka eating share are all known and priced in.
(yes it popped 4% today, i was watching this before — thesis hasn't changed)
thoughts?
Edit: for people who keep DMing me the screenshots are from Thesis . Not affiliated.
Risk-reward in options feels different from regular trading because being right on direction is not always enough.
You can be right on the move but wrong on timing. You can have a good thesis but get crushed by theta. You can hold for the target and still watch premium decay. You can take profit too early because the position moves against you emotionally even when the chart is fine.
That makes risk-reward harder to judge.
A stock trade can wait sometimes. An options trade usually cannot wait forever.
So I think options traders need to think about risk-reward in three parts.
Price target. Time risk. Volatility risk.
Without those three, the RR can look good on paper but behave very differently in real life.
Currently on my mac after every trading day I run two executables. "goldilocks" (dont ask about the name Lol) audit and journal.
Features of the audit consist of a walkthrough of IN-TRADE minute by minute option data and also two hours post exit data to formulate accurate bracket recommendations (which t1/t2 runner if applicable is best suited without hitting stop first). accounting for things like CALMAR (max drawdown) and the sharpe ratio with minuscule risk of ruin. Self optimizing as the sample grows. I only trade QQQ 0dte single leg long only (puts/calls)
Features of the journal are like what many have already seen before. off the executable a push to schwab api to recollect new trades (old is cached, to reduce future run times). Net P/L, option price history chart in trade and post exit data, notes that the md can recollect for future advice/changes to brackets, and a few other things you can see.
my question now is, what can be improved? what are some things you have implemented to aid in your trading, happy to hear what the community here has to share.
BOXX is basically a fund that shifts STCG -> LTCG for the risk free rate. It kind of behaves similarly to a money market fund, but technically it's an ETF so can't directly use it as collateral for selling CSP's. And need to hold BOXX for at least a year to benefit from LTCG, so assignment on CSP's would cancel out benefits of BOXX. Here is my plan:
Selling CSP's on margin, and either rolling the option near expiration if it's in the money, or liquidating my lots of BOXX that are LTCG a few days before assignment if I'm expecting to be assigned.
Cons I see with this plan:
For people living in low tax states who benefit from BOXX over traditional money market funds as collateral, what are the other downsides of this strategy, which is essentially switching your collateral for CSP's from money market funds to a more tax advantaged fund? Is there other instances I would need to pay interest on the margin I'm not considering or something?
Never heard of anyone discuss this strategy. Let me know what y'all think.
Me - gonna play but it's close. Gonna be long defined risk IC 2DTE.
Reasons
But, with IV cheap and a nicely constructed, reward/risk IC (taking advantage of the vol surface), I'm in. Not for a lot - just 1% of my net assets.
Any one else long gamma on $NVDA or just the normal short sellers out there??
I saw a video a while back that explained a general rule: price up + volume up = buyers & price down + volume up = sellers. The video didn’t mention IV, which I suppose could better estimate buys/sells of a specific contract.
Any validity to the claim? And does anyone know of a better estimate?
$INTC. $RKLB. $AMD. These tickers have been on a tear recently, with nary a call to be sold. I think the sellers are back in the game. We're back to the point where selling premium makes sense again. It was a rough couple months, because even though the HV was massive, IV was super suppressed. The only stocks that actually had elevated IV were ones that still climbed way outside the expected move time after time after time.
Now that things have kind of returned to some sense of normalcy, I am reveling in the premium.
The stop losses I know on an options trade would be pnl%, delta, time. Are there others that could be used as a stop loss? I have been primarily using pnl% at 200% though find that many trades get shaken out and would have ended up profitable if I had held. What stop losses are best to use and what are the parameters you like to set?
Hi I’ve been doing option trading for a year and all I do is watch important news that Trump post on his page Truth, for example he post “We agreed on a ceasefire of 3 weeks with Iran… blah blah” I just buy calls with 2DTA.
I though that was a good idea until 2 days ago that we posted something similar and stocks market just went up like in 30 seconds and It was late when I get in and I lost $2k, I don’t know if need to keep the same strategy or do something different, the thing is that the market react so fast when think like that happens.
Serious question for people who’ve been trading options for years:
How many of you are genuinely outperforming just buying and holding VOO once you factor in:
I’m not talking about screenshots from a 3-month hot streak. I mean actual multi-year performance.
It seems like almost everyone eventually ends up in one of these categories:
At the same time, there are clearly some traders here who:
So I’m curious:
Would especially like to hear from people trading through:
Interested in honest answers, not just gain porn.
I’m gonna say something that sounds harsher than I mean it:
A lot of traders including myself spend years staring at candles and never realize they’re mostly looking at the aftermath of a fight, not the fight itself.
And yeah, candles matter. I’m not saying candlestick charts are useless before someone grabs their pitchfork.
But if you put a candlestick chart and a live order book heatmap side by side at the exact same moment, you realize they are showing you two completely different layers of the market.
One shows you what happened while the other shows you what is trying to happen.
That difference matters more than most people think.
The candle is historical evidence
A candlestick bar is simple: Open. Close. High. Low and it is incredibly useful for structure, momentum, pattern recognition, context… all the stuff we traders already know.
But candles have one fatal limitation: They only show completed transactions.
By the time you see that candle close, the actual battle is already over.
Buyers and sellers already fought. Orders got matched. Liquidity got consumed. Somebody won that round.
How foolish my younger self was when I drew a support line from candles printed three days ago and thought it would hold.
The market doesn’t owe anyone a repeat performance just because a candle bounced there on Tuesday.
The participants that defended that level might be gone. You’re making a decision based on historical reaction and assuming the live order still supports that thesis.
The order book shows live intent
This is when everything hits me differently about how I am missing the whole trading thing by just looking at candles.
A raw DOM (Depth of Market) is useful, but it updates so fast. A liquidity heatmap solves that problem by turning the order book into something visual.
Instead of numbers flashing in a ladder, you see liquidity as stacked zones of resting orders.
Bright = heavy liquidity.
Dark = thin liquidity.
And suddenly I can answer questions candles can’t answer:
That’s live market intent. And yes, intent can change — that part matters too.
But at least now I am not trading with my eyes blinded and praying for the market.
Sounds good on the paper right, how does it work in reality? In fact, the order flow will explain two common scenarios in trading:
Wall support / resistance
Price approaches a thick, bright liquidity cluster.
That wall just sits there and when the price hits it and struggles because there is actual resting capital there absorbing flow.
That’s real support or resistance. Because the actual size is sitting there and forcing price to deal with it.
A line on a chart is an idea but a liquidity wall is money doing business.
Spoofing liquidity
This one traps traders all the time, including me.
A giant wall suddenly appears, order book looks massively skewed and retail traders see it and think:
“Oh wow, huge support. Safe long.”
Then the price gets close……and the wall disappears. Gone.
Like my discipline after saying “just one more trade.”
A huge fake order designed to influence positioning. Retail buys early because they believe support is there.
Then liquidity gets pulled and stop loss gets triggered. 5 minutes later, everybody posts “manipulation” on X like the market personally attacked them..
BUT
Let me kill the fantasy before someone says it: The order book is not a crystal ball that you look at and see the market.
Because walls get pulled, liquidity shifts and big players reposition constantly.
You are not getting some magical “predict the future” cheat code. That’s retail bullshit.
However, order flow changes the mindset from : “Will this historical level bounce?” to “Is there actual liquidity here right now and can the wall hold?”
That’s a much more mechanical framework. And in trading, mechanical usually beats emotional.
As the old joke goes:
The market is a device for transferring money from the impatient to the overconfident.
Actually, Buffett said “patient,” but honestly sometimes “not being an idiot for five minutes” works too.
Regardless, candles still matter a lot and I still use them to read market structure every day.
But candles alone do not show the live resting liquidity behind the move. That’s the blind spot.
Also, the heatmap doesn’t replace charts, rather it adds another confirmation layer.
It lets you see whether a support level has actual size behind it… or whether you’re trading a ghost level.
2 years trading options, mostly SPX and AAPL. last 6 months i got really into 0DTE because the gamma is fun and the cycles are fast. account hasn't moved much though so i finally pulled the data and split it by DTE.
0DTE: 187 trades, 41% win rate, -$4,200
1-7 day expiries: 64 trades, 52% wr, +$3,100
8-30 day expiries: 31 trades, 55% wr, +$2,400
on paper the 0DTE win rate isn't catastrophic. the kicker is i was sizing 0DTE 2-3x what i risked on longer-dated stuff because "they're cheap and they're quick." bigger size on lower win rate = the bleed.
the part that hurt: i kept telling myself 0DTE was "my edge" because i could see results faster. the data says 0DTE was just where my discipline broke. sitting at the screen all day, taking marginal setups, sizing emotionally because i could exit by close anyway.
last 3 weeks i cut 0DTE entirely, went back to 7-day spreads with strict fixed sizing. tiny sample but the account isn't bleeding for the first time in months.
anyone else look at 0DTE vs longer expiries and find a discipline drop on the short stuff? is it the contract itself or the format/pace that does it?
Can someone make sure this gets to the right people who actually have the influence to change it?
The "Rule of 390" (which designates you as a "Professional Customer" if you average more than 390 orders a day in a month) is completely outdated. Right now, exchanges count every single modification and cancellation toward that limit.
In today’s fast-moving market, if you are trading multi-leg spreads or trying to walk a limit price to get filled, you have to constantly edit or cancel/replace your orders just to keep up with the shifting underlying asset or market makers. If you run a simple script or just active manual adjustments, you can hit that 390 limit in a heartbeat without actually executing more than a handful of trades.
It makes zero sense that cancellations and edits count toward a metric that strips you of retail priority. The rule should strictly be based on fills only. If an order isn't executing, it isn't taking up liquidity or acting like an institutional market maker.
We need the exchanges and regulators to look at modern market structure and update this. If anyone here has compliance channels or institutional ties, please push this up the ladder. It’s actively punishing active retail traders for just trying to get a fair fill.
Options are already complicated enough.
Greeks, spreads, liquidity, IV changes, expiration risk, assignment risk. There is already a lot to manage before even thinking about the platform.
That’s why I think platform design matters more for options than people admit.
If the interface is unclear, it’s easier to place the wrong order. If risk is not displayed clearly, beginners misunderstand exposure. If fills are poor, spreads become more expensive. If position details are confusing, managing exits becomes stressful.
For simple trades, maybe platform design is just convenience. For options, I think it can directly affect decision quality.
What do you think makes a platform good for options. Clean order entry, risk display, fills, education, or something else?
I'm adding more data about trades at opening and daily to my query and analysis tool so I can get better insights from my trading history. Thoughts on what I should be collecting? I'm thinking the IVR and greeks, bid/ask at opening, daily close and closing. Anything else?
I’m looking for feedback from futures/options traders who understand CME Globex execution mechanics, ES futures options, Market Protect orders (perhaps Rule 588 if done in 8 minutes time frame required?). I am a very experienced trader, a licenced fiduciary wealth manager and financial planner in my day job (Series 7 and 66 licenses. )
Yesterday I opened and closed 7 ES futures call options (1 day DTE) within three minutes:
Product: ES FOP / E3B
Contract: May 19, 2026 7405 Call
Ticker: E3BK6 C7405
Exchange: CME
Platform/Broker: Interactive Brokers TWS
Here is the sequence:
I bought 7 CALL contracts at 15.90 STRIKE 7405 at approximately 15:01:20 ET, roughly 13 points OTM (Out-the-money).
I got very lucky (or did I? I saw unusual movement across my indicators, probably by the usual insider traders). About 45 seconds later Trump made a post about delaying his planned military strikes on IRAN. ES had moved sharply higher and the option premium was trading around the low/mid 30s. Bid/Ask is highly unstable/wide spreads when indices move that quickly. So I waited for it to stabalize at the usual spreads for about 20 seconds. You can see from the attached Time & Sales screenshots that the bid/ask did stabalize like I say it did. I submitted a Market Protect sell order to close the position.
Instead, my closing fills were:
15:03:15 — Sell 1 at 5.00
15:03:15 — Sell 1 at 4.50
15:03:15 — Sell 5 at 4.50
Average sale price: 4.571429
Multiplier is 50x, so if the fair exit was around 34, the economic difference is roughly:
34.00 × 50 × 7 = $11,900 expected proceeds
4.571429 × 50 × 7 = ~$1,600 actual proceeds
Difference: about $10,300
What makes this especially confusing is the Time & Sales / NBO data I pulled afterward. Around the exact time of my fills, the displayed market appears to show roughly 32.00 / 35.00 or similar. My 4.50 and 5.00 fills print at 15:03:15, while nearby quote data still appears to show a bid/ask nowhere near 4.50.
Then shortly afterward, the quotes appear to flicker down to bizarre levels like 0.95 / 34.50, 1.00 / 34.25, etc.
I contacted IBKR immediately. The rep reviewed Time & Sales and said:
“the market right before your order went in was 0.95 bid at 34.75 then no bid at 34.75 then no bid at your 4.50 then 5 bid where you traded it”
That explanation doesn’t make sense to me because the screenshots seem to show the 4.50/5.00 prints occurring while the displayed market was still around 32/35, and the 0.95 bid quotes appear after or around the same second. Obviously milliseconds matter, but the visible sequence looks extremely questionable.
I filed a trade adjustment request with IBKR and separate complaints with CME and IBKR. I understand CME Rule 588 review windows are extremely short, and I was already outside the formal bust window – which is only 8 minutes and I did not know that you have to call CME within the eight minutes to have the trade bust considered (learned something new and valuable).
Anyone experienced enough rever experienced something similar, or have the expertise to give me their opinions? Please chime in.
My questions:
Time & Sales + Underlying Time & Sales
Attached Proof:
- Screenshots of the Time & Sales/NBO around and at the execution.
- My trade confirmations
- The trade displayed in my entire day’s audit trail (redacted)
- The trade confirmation email
I’d appreciate serious feedback from anyone who trades ES options or understands Globex order matching. I’m angry, obviously, but I’m mostly trying to figure out whether this was a legitimate thin-market disaster or an execution/data problem that should be challenged.
The 19$ strike for November are in the money now, but liquidity is so low that they aren't priced correctly. I think the underlying could move explosively if buffet so much as hints at investing their cash surplus. Thoughts?
