Interpreting the NOPE: A Brief User’s Guide


I’ll keep it short, and also post a TL;DR on NOPE:

- Using high/low NOPE you can see:

(Potentially) intraday points where SPY reverses direction (Intraday)

(Potentially) detects false rallies/selloffs (Intraday)

(Likely) Estimates next day behavior [green/red] (using EOD NOPE)

(Probably) Predicts higher/lower monthly returns (using EOD NOPE)

(Likely) Predicts when the crash bottom is here/near (using EOD NOPE)

(Probably) Detects crashes soon to come (using EOD NOPE)

(Likely) Detects how much volatility to expect next day (using EOD NOPE)

— —

As of right now, ER plays using NOPE are on pause; we’re actively re-forward-testing our model, and I want to ensure that it actually is returning the results we expect rather than just being a weird data anomaly and costing people (including myself) money.

That said, if you’re on the Discord, or have seen me rant on twitter (, the major use case for NOPE (I will add links to my prior posts describing the purported effect) is to predict market anomalies, in the ideal case predicting crashes before they happen, and bottoms of crashes when they happen.

In general, to achieve this, we use the intraday and historical NOPE data (intraday from secret brokerage API, historical from ORATS) in order to examine trends. I want to reiterate that I have no ownership of (although the guy who runs it works on the project with me, obviously), and I only am linking it since well, the metric doesn’t exist yet elsewhere. If you want to build a tool using it, PM me and I’d gladly assist for no cost; I’m in this just for the research.

That said, generally the way I examine NOPE is by looking at the intraday behavior, which can be found at the following URL: (the date changes, so just put the right date for today when you use it). For ease-of-use reasons, all my analysis is done on SPY rather than QQQ or IWM, although it would probably work there too.

This page can be decomposed into three graph lines:

NOPE (Net Options Pricing Effect) — This gets updated every minute or so, and on SPY usually aligns with SPY’s price evolution intraday. I’ll talk about our hypotheses why (since price does not appear in the NOPE equation), but to be quite honest it’s still an open question.

NOPE_MAD(30) (NOPE Median Absolute Deviation over 30 days) — This is a simple statistical formula comparing how “weird” today’s NOPE value is compared to the prior 30 days. It has issues, but it’s where sigma comes from (a measure of deviation). In general, SPY moves between about -3 and +3 sigma.

SPY Price — Line graph of SPY’s price data.

Some things/common questions to clarify:

  1. The intersection of the lines probably don’t mean anything. — It’s not a well scaled graph, and it’s clear the Y axis is very different between NOPE, NOPE_MAD, and SPY price. Please don’t apply TA like golden crosses; it’s likely complete noise.
  2. Why did the black line stop? — This is running on a literal shoestring budget, because it’s literally against our morals to try to charge for this. Right now I think the website is running at like -$150/mo, but thankfully our team can afford it. That said, our infra and data connections ain’t good, plus it’s a side job/project, so sometimes shit breaks.

Interpreting the NOPE

Let’s segment this question for using it on SPY into two time periods:

1) Intraday

2) Next day (and beyond)


I want to start this off with the following warning: we have evidence of intraday behavior observed til about 9–8–2020 (when we started collecting the data) but no data past that. We have not statistically validated our intraday usage, but anecdotally/observation wise it seems to be quite useful.

Intraday, there was two major important concepts for the NOPE/NOPE_MAD line — reversions and divergences.

On average, SPY ends since about 2007 at roughly 0 NOPE — e.g. not skewed to put or call delta. This makes it very interesting intraday when it exceeds a certain threshold, because from what we’ve observed, this implies the price direction of SPY is likely to reverse.


In general, I use the “rule of 3 sigma” rule of thumb, but it’s again not an actual rule. For values around +3 sigma, in general (aka probabilistically) I expect that as a signal for SPY to reverse price direction intraday (aka go down, a good signal to short). This is because in almost all cases observed, SPY price will decline when NOPE_MAD values are elevated.

Similarly, at around -3 sigma, assuming no news catalyst (e.g. like… lockdowns), in general we expect SPY intraday to also reverse direction, which makes it a good signal to go long.

This hasn’t been proven yet, but multiple people have used this to pretty nice success, and you can see it matches in fact exactly on today’s chart. I have a secret hypothesis why this effect might work, but that’s between me and the future Nobel Prize committee.


The second major use case of SPY intraday we’ve seen is cases where NOPE and SPY “diverge” — that is, NOPE leads SPY in shape (going up while SPY goes down over a meaningful stretch of time, like more than 5 mins) or vice versa. From again anecdotal evidence, it seems NOPE tends to lead SPY; that is, when NOPE is dipping while SPY is rising, it indicates a false rally (or if NOPE is rising while SPY is dipping, this indicates false selloff). Again, not proven yet, but it tends to happen in very notable cases.

The most recent example I remember where the divergence was especially notable was the massive 3.5% selloff on Oct 28 (

As you can see, NOPE immediately dived at open, while SPY false rose. Then, SPY followed, and dollars were saved by me and others.

Divergences again seem to occur in extreme scenarios, and in general we view the intraday NOPE graph to more or less look like SPY’s price graph.

Next day (and beyond)

Here’s where it gets more interesting, and I do have data. In fact, I have SPY end of day data versus NOPE since about 2007. It unfortunately calculates NOPE at 3:45 PM EST, but this probably makes the results look weaker than they are, not stronger (so I’d expect it to be more correct in the real world).

This only considers end of day NOPE data, and likely again ends up seeing less than the real world (since its likely intraday, as we saw on Oct 12th, also matters).

The working hypothesis, starting from our earliest days looking at NOPE on SPY, were that it could function as an effective crash predictor. This is similar to the use case of the put-call ratio for the same purpose, although this seems to perform better in practice so far (it caught the COVID crash, the crash of Sep 3rd and Oct 12th, for instance).

I’m going to let the analysis results speak for themselves:

In cases of positive NOPE (>0):

- As NOPE increases (30 -> 40 for instance), the chance of the next day being red (SPY net movement < 0%) increases. This becomes nearly 100% at high magnitudes (like 90 end of day), but I don’t want to make that statement since the sample size at super high NOPE values is so low.

A graph comparing NOPE value (x axis) to chance of SPY being positive the next day.

- Positive end of day NOPE is tied to decreased next day volatility (estimated comparing net difference in % between high and low SPY data). This is likely similar to positive GEX, and in fact I chatted with SqueezeMetrics about it earlier today on Twitter.

- Moderate (0–30 or so) end of day NOPE is tied to increased 14 and 30 day returns (estimated by comparing day’s close with % change to lowest low in the next 30 trading days). This is because it likely just indicates healthy bullish sentiment.

- High positive NOPE (40+) correlated to worse 14 and 30 day returns (this is likely indicative of crash/correction events). This is because it likely indicates irrational exuberance.

In cases of negative NOPE (<0):

Interestingly, unlike positive NOPE, negative NOPE seems to show no correlation to next day direction of SPY (see the above graph).


- Negative end of day NOPE is tied to increased next day volatility (estimated comparing net difference in % between high and low SPY data). I also controlled by looking only at negative NOPE on green days, because otherwise it’d be heavily biased (since crashes usually have multi-day negative NOPE).

Most interesting:

- High magnitude negative (NOPE < -30 or so) end of day NOPE is tied to increased 14 and 30 day returns (estimated by comparing day’s close with % change to lowest low in the next 30 trading days). This is because it likely indicates a day near a crash bottom (so the next 30 days is going up).

- High magnitude negative (NOPE < -30 or so) end of day NOPE is tied to lower median time to bottom. Time to bottom in this case measures days until the lowest point in the next 30 trading days. In a crash, we’d expect that the day before the crash bottom has a time to bottom of 1 day, basically (and the first day of the crash to have the longest time to bottom).

Over all days, on average a lower low than today’s close will occur in around 11 days. In cases where NOPE = -50 or below, that time goes to 6 days. At -70 or below, that time goes to 5 days.

This is heavily indicative of crash bottom “near”/”like” behavior.

— -

That said, the major caveat here is the project is new. While the recent analyses I did are promising, I’m mindful that I haven’t done statistical testing on it yet, and it’s quite possible some of the analyses may have bias. I want to caution people to use this at the moment not as a play decider, but another indicator to observe and exercise appropriate caution with.

Thanks, and have fun.

- Lily

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store