How to Find an Entry
To find an entry, mark a level first, name the single thing that has to happen at it, and decide the stop before the position exists. The level supplies the reason and the trigger supplies the timing, and neither works without the other.
Most entry methods are a signal in search of a reason. This one runs the other way round, because a level can be tested and a feeling about a crossover cannot.
Before you start
A chart with no indicators on it. Add them later if the test says they earn their place — the do indicators work page has that test.
One instrument, on the timeframe you actually trade.
A stop distance you can measure — the swing, or a multiple of average true range (ATR). You need this before step one, because it decides the size and therefore whether the trade is takeable at all.
The steps
1. Mark the levels before anything moves
Prior highs and lows, the session’s open, the busiest price. Marked while you are flat, so the chart cannot argue you into one.
2. Name one thing that has to happen at the level
A close back above it. A sweep and reclaim. One observable event, written down. If you need two conditions to feel confident, write both — but write them.
3. Place the stop before you place the order
Beyond the level by a measured amount, per stop placement. An entry whose stop is decided afterwards is a position, not a trade.
4. Size from that distance
Money you will risk, divided by the stop distance. If the answer is a position you would not sit through, the entry is not available to you at this level.
5. Take it or skip it, and record which
The skips are data. A record of only the trades you took is filtered by the judgment you are trying to test, which is the trading journal problem.
How to tell it worked
Not by the outcome of one trade. By whether the trigger beats entering on a schedule.
The measurement, run on 576 bars: every moving average convergence divergence (MACD) cross taken as a long, with the same stop and target, gave 14 wins and 11 losses across 26 signals. The same number of entries taken every 21st bar, ignoring the chart entirely, gave 14 and 12.
Indistinguishable. Which does not mean signals are useless — it means an entry trigger has to be tested against something before you can claim it adds anything, and almost nobody runs the control.
Run yours the same way. Count every firing across a long history, take the same number on a schedule, apply identical stops and costs. If the trigger does not beat the calendar, it is not an entry signal — it may still be a useful filter, which is a smaller and more defensible claim.
Expect it to fire rarely
The most-taught beginner entry — relative strength index (RSI) crossing back up through 30 — fired four times in 576 bars. On the slower view of the same history it fired zero times in 144.
The longest gap between two of those four was 329 bars: 57% of the whole history with nothing to do.
A tutorial shows six examples in ten minutes because it searched years and instruments for them. Your chart will not, and the gap between those two experiences is where most people conclude the method is broken.
Which levels are worth waiting at
Not all of them, and the difference is measurable.
A level works because orders rest at it, which means it works in proportion to how many people are looking at the same price. That is the crowding argument the support and resistance page makes, and it has a practical consequence most entry material skips.
Measured across five price series: the first return to a level held 6 times out of 11. Later returns held 4 of 13.
Both samples are small and both point the same way. By the fourth test the orders that were sitting there have been filled, and what is left is a line on a chart with a memory attached.
So prefer a level that has not been tested yet, and treat the third and fourth visit as progressively weaker rather than as confirmation that the level is important.
The original data
Across our study of 24,971 trading videos, 26 cover finding entries specifically. The median one gets 5,857 views, 77% never pass 50,000, and the median length is 11.6 minutes.
The corpus carries description text for only seven of those 26, which is too thin to say anything about how the topic is written, and this page does not.
The field size is the finding. 26 videos on how to choose an entry, against 698 on MACD and 844 on RSI — thirty times as much material on the tools as on the decision the tools are meant to serve.
When it fails
The signal fired in open space
A trigger with no level under it has no reason attached. It can still work; you just have nothing to check it against when it does not, and nothing to tell you whether the next one is the same trade or a different one.
You added conditions until it stopped firing
Each extra condition removes trades and does not reliably improve the ones left. That is a trade you are making rather than a free improvement, and it is almost always described as if it were free. Measured on the same 26 signals: adding a momentum filter cut them to 17 and the record went from 14-11 to 8-8.
The range made it fire constantly
The same trigger that is selective in a trend is noise in a range, and the costs of that are the frequency table.
You judged it from the right-hand edge
At the moment it fires, the fourteen that won and the eleven that lost are the same picture. That is why the rule, the stop and the size all have to exist before this bar closes.
Related
How to read the market is the session routine this entry sits inside.
How to backtest a strategy is how you run the control test properly.
And entry and exit is the wider decision, including the half that is harder than this one.
The change that mattered for me was reversing the order. I used to look for a signal and then find a reason; now I mark the level first and the signal is only allowed to decide when, not whether. It cut the number of trades by about half and that was the entire improvement.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.