Learning from others safely

It is 09:10 ET (21:10 ICT, winter). A paid channel you are thinking of joining posts its morning plan for NQ: one main level, a bias, a target. The sales page says the main level “gets hit 80% of days”. Twenty minutes later the bell rings, and by 09:52 NQ trades through the level. It looks like skill. Is it? This lesson shows you how to answer that question with a spreadsheet and two free baselines, before you pay anyone.

Key terms

Baseline (mức nền): what you would score on the same days with a rule that takes no skill Definition. “Always bearish”, “a coin flip” and “a level the same distance away” are baselines.

Mirror level (mức đối xứng): the price the same distance from the 09:30 ET open, on the other side Definition. Mirror = 2 × open − level.

Respect (giá tôn trọng mức): after price touches a level, it moves a set distance away before it trades a set distance through Definition. You write both distances down before you look.

Term you need: ATR. Average true range: the average size of a day’s move over the last 14 days. A level 0.08 ATR from the open sits about one-twelfth of a normal day away, so price reaches it almost by accident. EV1.05 shows why distance drives touch rates.

Chart first: a level near the open gets touched

How it works

One record, four tests

A morning plan makes four different claims, and each has its own free baseline:

1. Touch. Price reaches the level. Baseline: the mirror level, same distance from the open.
2. Hold. After a touch, price respects the level. Baseline: every round-number level near the open, scored with the same distances.
3. Direction. The bias is right at a fixed time (10:30 ET, the close). Baselines: a coin flip, always long, always short.
4. Target. The named target trades during the regular session. Baseline: a level the same distance from the open in the opposite direction.

A banner that merges these into one “hit rate” hides which test the number came from. Touch is the easiest test to pass and says the least.

Why touch rates are cheap

In our 65-session QQQ sample, a level 0.08 ATR from the open was touched on 80% to 86% of days whatever it was called Measured: no edgeNasdaq index (EV1.05 shows the distance math). That is the whole point of the chart above: the mirror line has no author, no bias and no story, and it scored at least as well.

Five questions before you pay

Ask these of any desk, chat room or course (EV3.04 teaches them in depth):

• Sample: how many sessions, and which ones? Under about 100, the interval is wide (EV3.01).
• Baseline: what did the free rule score on the same days?
• Costs: is the record in points before commission and slippage, or in money after them?
• Holdout: were the rules fixed before the days they were tested on?
• Selection: is every posted call in the record, or only the screenshots that worked?

A service that cannot answer the second question has not shown you anything yet House rule.

What the data says

Our lab scored one widely followed paid desk’s pre-market plans on QQQ (a fund that tracks the Nasdaq-100) over 65 sessions in late 2026. Each plan was taken as posted before the open, and every day was scored, not a selection. The directional bias and target columns were coded by our lab from the desk’s written morning posts; 60 of the 65 sessions used 5-minute bars, while 5 sessions used 1-hour bars (for touch and direction only).

Read it fairly. The desk did not beat its baselines on the three tests a sales page usually leads with. On the one test that comes closest to “the level mattered”, it scored above the pool PlausibleNasdaq index, but not reliably enough to pay for.

Worked example: scoring a long-bias morning on NQ

A synthetic day, labelled as such. The service posts at 09:10 ET: main level 21,040.00, bias long above it, target 21,100.00. You fix your hold rule in advance using your own illustrative numbers: 25 points away before 15 points through, within 45 minutes.

At 09:30 ET NQ opens at 21,012.25. You write down:

• Distance: 21,040.00 − 21,012.25 = 27.75 points (111 ticks of 0.25).
• Mirror: 2 × 21,012.25 − 21,040.00 = 20,984.50.
• Opposite target: 2 × 21,012.25 − 21,100.00 = 20,924.50.

What happened: NQ rallies to 21,040.00 at 09:52 and touches the level. By 10:05 it is back at 21,012.75, 27.25 points away, without ever trading 21,055.00, so the level held. At 10:30 ET (22:30 ICT) NQ is 21,031.75, above the open, so the long bias scores right at 10:30. It closes at 21,004.00, below the open: wrong at the close. The low of the day is 20,991.00, so the mirror was not touched. Neither 21,100.00 nor 20,924.50 traded.

Your sheet row: touch yes, mirror no, held yes, bias 10:30 yes, bias close no, target no, opposite no. One row says nothing. Twenty rows start to say something; sixty-five rows said what the evidence box shows.

The mirror and the failure

Short side. On a bearish morning the scoring flips. Say the post reads “short below 20,984.50, target 20,924.50”. The opposite target is the same distance above the open, and the direction baseline is always long as well as always short. Score it with the same sheet.

Where it fails. The touch test fails in a way that looks like success: the paid level scored 80%, and the mirror scored 86%. The direction test fails a second way. A desk that leans bearish in a falling month can look right without timing anything, which is why you compare against always-bearish on the same days, and why beating it in one month proves little either.

Do it on the platform

Prop-firm note. If you trade an evaluation or funded account, copying someone else’s calls can fall under a firm’s rules on coordinated or copy trading. Read the firm’s prohibited-conduct article before you mirror anyone (PROP1.04).

Common mistake: treating a base rate as an edge. “Hit 80% of days” is a base rate until you see what the mirror level scored on the same days.

Common mistake: judging a service on the screenshots it chose. Ask for every posted call, losers included, with the time it was posted.

Drill: score a service for 20 sessions

Pick any free or paid level service you already follow.

1. Before the next 20 sessions, write your hold rule (points away, points through, minutes) and your direction times.
2. Each morning, record the posted level, bias and target with the time of the post, then the 09:30 ET open, the mirror and the opposite target.
3. After each session, fill the seven columns from the worked example.
4. After 20 sessions, put each rate next to its baseline.

What we don’t know yet

• Whether the hold lift is real: not yet measured with enough touches (about 150 to 200 needed).
• Whether the big-gap split survives a forward test: not yet measured.
• How this desk’s levels behave on NQ futures directly: not yet measured; the study used QQQ.

Recap

• Split any record into touch, hold, direction and target, each with its own free baseline.
• A level near the open is touched on most days, so touch rates are mostly distance.
• In our lab’s 65-session study, one paid desk did not beat its baselines on touch, direction or targets; a hold lift was suggestive and unproven.
• Ask for sample, baseline, costs, holdout and selection before you pay.

Further reading

• Barber, Lee, Liu and Odean (2014), “The cross-section of speculator skill: Evidence from day trading”, Journal of Financial Markets 18. Few day traders show persistent skill once you compare them with chance.
• Freedman, Pisani and Purves, Statistics (4th ed.), the chapters on chance variation: why 20 sessions give wide intervals.
• Our lab: one widely followed paid desk’s QQQ levels, bias and targets scored against mirror levels, round-number levels and naive direction rules over 65 sessions; the touch, direction and target tests showed no edge, and the hold test showed a suggestive, unproven lift.

Next: EV2.06 applies the same baseline lens to the framework terms you may already use, such as fair value gaps and order blocks.

Educational content only. Not financial advice. Futures trading involves substantial risk of loss.