Part 4: Trading and the system · Chapter 11 of 14 · 11 min

Entries at levels

Nine entry rules at levels (bounce, break and retest, on time-based levels, H4 swings and round numbers), with stop, target at 1 R, costs and one trade at a time, give mean R between −0.17 and +0.06, with errors of ± 0.08-0.15. None beats random direction on the same entries (−0.02 to −0.04). The bounce at round numbers (−0.17 ± 0.09) is the only one that stands out, on the negative side.

From the level to the trade

In the previous chapters we measured what the price does after a contact. A trade needs more: an entry, a stop and a target written beforehand. In the book there are three ways of entering at a level.

  • Bounce. At the close of the contact candle you enter in the direction of the expected bounce (sell at a resistance, buy at a support). Stop half an ATR beyond the zone, on the far side.
  • Break. At the close of the first candle beyond the zone (chapter 7) you enter in the direction of the break. Stop a quarter of an ATR inside the zone, on the starting side.
  • Retest. After a true break (chapter 8), at the first contact with the old level from the new side you enter in the direction of the break. Stop half an ATR beyond the zone, on the starting side.

For all: target at 1 R (the same distance as the stop), time exit after 24 candles, cost 0.3 $, one trade at a time (if one is open, later contacts are skipped), risk between 0.5 and 3 ATR. The comparison is with random direction on the same entries: the same entries, with the same risk, but buying once and selling once (on average), to see what the level's direction alone is worth.

The median risk is about 22 $, so the 0.3 $ cost is worth about 0.01 R.

Nine rules

LevelsRuleTradesMean R ± errorWinnersRandom directionFirst / second half
TimeBounce553−0.04 ± 0.0849%−0.04+0.07 / −0.14
TimeBreak551−0.03 ± 0.0849%−0.03−0.06 / −0.01
TimeRetest273−0.04 ± 0.1249%−0.03+0.02 / −0.08
H4 swingBounce366−0.11 ± 0.1046%−0.04−0.16 / −0.07
H4 swingBreak367+0.03 ± 0.1052%−0.03+0.10 / −0.03
H4 swingRetest178+0.06 ± 0.1554%−0.02+0.14 / −0.01
50 $ round numbersBounce419−0.17 ± 0.0942%−0.03−0.23 / −0.10
50 $ round numbersBreak425−0.04 ± 0.0948%−0.03−0.03 / −0.04
50 $ round numbersRetest175−0.11 ± 0.1545%−0.03−0.15 / −0.06

("Time" are the previous day, week, month and Asia levels.)

Three readings.

No rule has a clearly positive R. The two values above zero (H4 swings: break +0.03 ± 0.10 and retest +0.06 ± 0.15) have an error larger than the result, and in the second half of the sample they return to −0.03 and −0.01.

None beats random direction. On the same entries, buying or selling at random gives −0.02 to −0.04 R. The level's direction adds nothing. Round numbers aside, all the rules are inside the error of random direction.

Round numbers are the exception, on the negative side. The bounce at round numbers gives −0.17 ± 0.09 R over 419 trades, winning only 42% of the time, with random direction at −0.03. It says that at round numbers it was better not to buy the bounce. It is consistent with chapter 5 (round numbers hold 44%) and it is one rule out of nine: to retest, not to use in reverse.

Figure 1. A bounce that reaches the target (10 April 2026, UTC hours): yesterday's close (4,765.55 $) is a resistance, short entry at 03:00 at 4,763.73 $, stop at 4,781.05 $ (17.3 $), target at 4,746.41 $ reached at 06:00: +0.98 R.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 9 Apr 2026 15:00 UTC to 11 Apr 2026 02:00 UTC.
Read the chart in words

H1 gold chart of 10 April 2026 with the zone of yesterday's close, the entry, stop and target of a winning trade.

WhenWhatPrice
9 Apr 2026 15:00 UTCyesterday's close 4765.554,760.29 – 4,770.81
10 Apr 2026 03:00 UTCentry4,763.73
10 Apr 2026 03:00 UTCstop4,781.05
10 Apr 2026 03:00 UTCtarget 1 R4,746.41
10 Apr 2026 06:00 UTCtarget: +0.98 R4,746.21

Times in UTC.

Figure 2. A bounce stopped out (5-6 January 2026, UTC hours): yesterday's high (4,450.00 $) is a resistance, short entry at 23:00 at 4,437.77 $, stop at 4,462.53 $ (24.8 $) reached at 03:00: −1.01 R.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 5 Jan 2026 11:00 UTC to 6 Jan 2026 22:00 UTC.
Read the chart in words

H1 gold chart of 5-6 January 2026 with the zone of yesterday's high, the entry, stop and target of a losing trade.

WhenWhatPrice
5 Jan 2026 11:00 UTCyesterday's high 4450.004,445.81 – 4,454.19
5 Jan 2026 23:00 UTCentry4,437.77
5 Jan 2026 23:00 UTCstop4,462.53
5 Jan 2026 23:00 UTCtarget 1 R4,413.01
6 Jan 2026 03:00 UTCstop: -1.01 R4,465.95

Times in UTC.

The curve

The cumulative R of the bounce at time-based levels (553 trades) ends at −23.2 R, random direction at −23.5 R: the two curves go together. The maximum is +20.0 R, in the middle of the sample, then the curve falls: the first half is +0.07 R per trade, the second −0.14.

Figure 3. The cumulative R of entering the bounce at time-based levels (bars, −23.2 R over 553 trades) and random direction on the same entries (line, −23.5 R): the curve rises to +20 R in the first half and then falls.Data: gold, XAUUSDT perpetual future (real candles) · daily (D1) · from 15 Dec 2025 to 5 Oct 2026.
Read the chart in words

Daily gold chart from December 2025 to October 2026 with, below, the cumulative R of the bounce at time-based levels and of random direction.

Different stops and targets

Changing the stop and the target moves the numbers, not the picture:

RuleStop 0.25 ATRStop 0.5 ATRStop 1 ATR
Bounce, time, target 1 R−0.04 ± 0.08−0.04 ± 0.08−0.06 ± 0.09
Bounce, time, target 2 R−0.08 ± 0.13−0.10 ± 0.13−0.13 ± 0.14
Bounce, H4 swing, target 1 R−0.12 ± 0.10−0.11 ± 0.10−0.08 ± 0.11
Bounce, H4 swing, target 2 R−0.02 ± 0.15−0.02 ± 0.16+0.07 ± 0.17
Bounce, round numbers, target 1 R−0.15 ± 0.10−0.17 ± 0.09−0.04 ± 0.10
Bounce, round numbers, target 2 R−0.03 ± 0.15−0.05 ± 0.14+0.05 ± 0.15
Break, H4 swing, target 1 R+0.03 ± 0.10+0.03 ± 0.11+0.07 ± 0.12
Break, H4 swing, target 2 R0.00 ± 0.16+0.08 ± 0.18+0.21 ± 0.19

Two things. First: the sign changes with the parameters (the bounce at round numbers goes from −0.17 to +0.05), and every cell has an error of 0.09-0.19. If you look only at the last row, the break on H4 swings with a 1 ATR stop and a 2 R target gives +0.21 R: it is the highest of 24 cells, with an error of 0.19, and it is exactly the kind of number that, with many tests, chance produces. Second: widening the stop reduces the trades (from 551 to 241 for the break) and increases the error.

To measure a mean R with ± 0.10 about 380 trades are needed (the standard deviation of R is 0.99), for ± 0.05 about 1,520.

Figure 4. Exercise: a bounce trade (18-20 February 2026, UTC hours): yesterday's close (4,979.25 $) is a resistance, short entry at 4,976.14 $, stop at 4,994.41 $, target at 4,957.87 $; four candles marked with the letters A-D.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 18 Feb 2026 13:00 UTC to 20 Feb 2026 00:00 UTC.
Read the chart in words

H1 gold chart of 18-20 February 2026 with the zone of yesterday's close, the entry, stop, target and four candles marked with the letters A, B, C and D.

WhenWhatPrice
18 Feb 2026 13:00 UTCyesterday's close 4979.254,974.20 – 4,984.30
19 Feb 2026 01:00 UTCentry4,976.14
19 Feb 2026 01:00 UTCstop4,994.41
19 Feb 2026 01:00 UTCtarget 1 R4,957.87
18 Feb 2026 21:00 UTCPoint A4,987.38
19 Feb 2026 01:00 UTCPoint B4,978.53
19 Feb 2026 04:00 UTCPoint C4,988.00
19 Feb 2026 05:00 UTCPoint D5,000.83

Times in UTC.

What to do with it

  1. Write the trade beforehand: entry, stop, target, time exit. And the cost.
  2. Always measure in R and compare with random direction. A −0.04 R that equals a random −0.04 means the level gave nothing.
  3. Do not buy the bounce at round numbers as a rule. In our data it gave −0.17 R, but it is one rule out of nine.
  4. Do not look for the best combination. With 24 cells of stop and target, the highest is the effect of the choice.
  5. Compute how many trades are needed. 380 for ± 0.10: for a rule that trades once a day that is almost two years.

Procedure in five steps

  1. Define the level, the contact and the entry (bounce, break or retest).
  2. Write stop, target, time exit and cost.
  3. Run the trades in order, one at a time, and measure R with its error.
  4. Compare with random direction on the same entries and look at the two halves of the sample.
  5. Count the combinations tried and compute how many trades are needed for the error you want.

Exercises

Chapter exercises

1. On the chart the lines are the entry (4,976.14 $), the stop (4,994.41 $) and the target (4,957.87 $) of a short bounce trade on 19 February 2026 (UTC hours). Which letter is the candle where the stop is hit?

Chart exercise. Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 18 Feb 2026 13:00 UTC to 20 Feb 2026 00:00 UTC.
Read the chart in words
WhenWhatPrice
18 Feb 2026 13:00 UTCyesterday's close 4979.254,974.20 – 4,984.30
19 Feb 2026 01:00 UTCentry4,976.14
19 Feb 2026 01:00 UTCstop4,994.41
19 Feb 2026 01:00 UTCtarget 1 R4,957.87
18 Feb 2026 21:00 UTCPoint A4,987.38
19 Feb 2026 01:00 UTCPoint B4,978.53
19 Feb 2026 04:00 UTCPoint C4,988.00
19 Feb 2026 05:00 UTCPoint D5,000.83

Times in UTC.

2. The bounce at time-based levels gives −0.04 ± 0.08 R and random direction on the same entries −0.04 R. What does it mean?

3. Among 24 combinations of stop and target, the highest (break on H4 swings, 1 ATR stop, 2 R target) gives +0.21 ± 0.19 R. How do you treat it?

4. The bounce at round numbers gives −0.17 ± 0.09 R, winning only 42% of the time; random direction −0.03. What do you take from it?

Have a question or want to share your exercise?

Post in the community, or join the free signals room on Telegram.

Educational content, not financial advice. Trading involves risk.