90 Days of Graded Setups: What the Grade Distribution Looks Like

I run SnapPChart, so I get a version of the same question a lot: what grade will my setup get? The honest answer lives in the data, not in a pitch. So here is the actual distribution of grades across every valid chart the tool graded in the last 90 days, 3,170 of them, with nothing rounded up to make the product look generous. It is not a flattering picture, and that is the whole point. Most charts, on most days, do not clear the bar, and a grader that tells you to skip the majority of what you upload is doing the one job that actually protects an account.

Quick Answer

What the 90-Day Grade Distribution Looks Like

Every number below comes from the same pool: valid chart analyses over a rolling 90-day window, meaning the upload was actually a readable chart and the tool returned a grade. The table sorts from the top grade down, so you can see exactly how thin the top is and how heavy the bottom gets. This is the raw ai chart grading data, not a marketing average.

The takeaway is not subtle. Just over two thirds of everything graded landed at C or F, a hair under a third made it to a plain B, and the entire top of the range, B+ and up, accounts for just 1.3% of charts. If you have ever felt like the grades run harsh, the distribution says you are reading it correctly, and that harshness is the feature doing its job.

How Many Setups Get an A Grade?

Zero. Out of 3,170 valid charts in the last 90 days, none earned an A or A+. The highest anything reached was B+, on 40 charts, which is 1.3% of the pool. That is the real answer to how many setups get an A grade, and it is worth sitting with rather than explaining away.

A grader that basically never hands out its top mark is not a grader that failed, it is one that refuses to be a rubber stamp. If A grades were common, the letter would mean almost nothing, and a meaningless grade cannot keep you out of a marginal trade at the exact moment you most want to be in one. The scarcity is the signal. It tells you that when a grade does climb, it climbed because the setup earned it against a fixed bar, not because the tool was in a good mood. That is the same reason a grade is a read on the setup and not a forecast, a distinction the piece on why AI grades a chart instead of predicting the market goes into in more detail. The grade is honest about what it can and cannot see.

What Verdict Do Most Setups Get?

Grades are one view. The verdict is the other, and for a lot of traders it is the more useful one, because it answers the blunt question of whether the tool would actually enter. Every graded chart gets sorted into one of three buckets: a clean entry now, wait for a pullback, or skip because there is no clean entry to stand behind. Here is how the same 90 days split.

Skip won by a mile at 88.0%. Only 6.1% of charts came back as an actionable entry right now, and another 5.9% were setups worth watching for a pullback rather than chasing at the current price. Put together, the tool said some version of not yet or not at all on about 94 out of every 100 charts. A buy signal that fires on everything is noise, and this is the opposite of that, which is also why a graded verdict is a different animal from an alert that just tells you to buy. The restraint is the value.

Why Does the Grader Say Skip So Often?

Because most charts are not trades, and a tool built to protect an account should say so out loud. This is the loss-prevention idea that runs through everything SnapPChart does. The edge in trading is not in taking more setups, it is in the ones you are kept out of, and an 88% skip rate is that principle turned into a number. Every skip the tool respects is a trade you did not have to lose money on to learn from, which lines up with the whole point of building a filter for the trades that are not worth taking in the first place.

The base rate backs this up. Day trading is a genuinely risky activity no matter how any single stretch goes, and the SEC's own glossary entry puts it plainly in its overview of how day trading works and why it's flagged as extremely risky, a risk FINRA reinforces in its guidance on the risks of frequent intraday trading. If the population of traders that trades the most tends to do worst, a tool whose default answer is skip is pulling in the right direction. It is not there to feed you setups, it is there to thin them.

None of this requires you to take the number on faith. The grade itself is a summary of things the tool reads off the picture, including whether the reward-to-risk is worth the stop distance, and how a reward-to-risk ratio is defined is standard enough that you can check the math on any entry it hands back. If you want the full walkthrough of what goes into a grade before it becomes a letter, the breakdown of what the AI actually reads off a screenshot covers the pipeline end to end.

What This Data Means for How You Trade

The practical read is that a good grade is supposed to be uncommon, so calibrate your expectations to the curve. If you upload a momentum setup and it comes back a B, that is not the tool being stingy, that is the top of the realistic range for most charts, and the momentum reads it sees most often, the bull flags, VWAP reclaims, and clean breakouts covered in the momentum trading strategy playbook, still have to earn every letter. Waiting around for an A that the data says basically does not come is a good way to talk yourself out of the workable B in front of you.

There is a founder-honest version of this that I put in the companion piece to this report, the first-person log of grading every setup by hand for 30 days, where the same lesson showed up as behavior rather than a chart: most of the value was in the trades the bar kept me out of, not the handful it waved through. The aggregate here is the population-level proof of that anecdote. Fewer, better setups is the entire trade, and a distribution this top-heavy toward skip is what fewer, better looks like when you count it.

The other read is about consistency. A grade only helps if it means the same thing every time, and a scarce top grade plus a dominant skip verdict is what a consistent bar produces. That steadiness is the reason a grade earns a place next to your own judgment rather than replacing it, an argument the wider guide to AI trading makes at length, and if you just want the plain-English version of what the read is, the AI chart analysis overview lays it out without any of these numbers attached.

How to Read Your Own Grade

Once you know the shape of the distribution, a single grade is easier to place. These are the rules I would hold in your head when your own chart comes back with a letter on it.

• Treat a skip as the tool working, not failing you
  Skip was 88.0% of all charts. It is the default answer, not a rejection of you specifically. A skip you respect is a loss you did not take, and that is the point of running the grade before you click.
• Judge your grade against what actually clears
  No chart hit A in 90 days and only 1.3% reached B+. So a B is near the top of the realistic range, not a mediocre score below some common A. Calibrate to the curve, not to a perfect mark the data says is rare.
• The grade is one input, not the whole decision
  It reads the static screenshot and hands back a letter, an entry, a stop, and targets. It does not know if you took the trade or how it turned out. You still make the call, and the log of what you did is yours to keep.
• Upload the setup you would actually take
  The aggregate includes a lot of charts nobody would trade, which is why F is the biggest bucket. Feed it the setups you are seriously considering and your own distribution will look different from the population.
• Let the grade influence size, not just go or no-go
  A grade is not only a gate. Lean more weight on the cleaner reads and less on the marginal ones, so your account leans on your best setups instead of averaging everything you take.