Part 2: Fibonacci for trading · Chapter 6 of 14 · 8 min

Fibonacci on POI: is the zone inside the 61.8%-78.6% band?

Does an order block or FVG that falls in the OTE band of the previous leg earn more? Of the 291 book 4 POI, 53 touch it: mean R +0.04 ± 0.49 against −0.13 ± 0.20 for the others, a difference of 0.17 R that random draws match in 25% of trials, and fake bands give the same.

The question

Book 4 has 291 points of interest (POI): order blocks and FVGs of gold on H4, with the limit order at the near edge giving on average −0.10 ± 0.18 R (216 trades filled). A widespread idea is to join the two methods: among the POI, those sitting in the 61.8%-78.6% band (OTE) of the previous leg are more likely to hold. Let us test it.

How it is measured

Reference leg. When the POI is complete, take the last confirmed leg in the direction opposite to the retracement: for a demand POI, the last confirmed high (B) and the low that precedes it (A); for a supply POI, the opposite. The swings are the book's (H4, legs of at least 2 ATR, confirmed two candles later).

Position. The price of each edge of the POI zone becomes a retracement ratio of the leg AB (0 = end of the leg, 1 = start). The POI is "in the band" if the zone overlaps the band even in part.

The center of the 291 POI sits like this relative to the leg: 70 beyond the end of the leg (price is already making a new extreme), 82 between 0 and 0.382, 30 between 0.382 and 0.5, 27 between 0.5 and 0.618, 13 between 0.618 and 0.786, 20 between 0.786 and 1, 49 beyond the start. POI are born almost always in a move, far from a deep correction.

The result

For each band the mean R of the trades of the POI in the band is compared with those of the others (same rule as book 4: limit at the near edge, stop beyond the far side plus 0.25 ATR, 2R target, cost 0.30 $; ± is the 95% interval):

Band of the legPOITradesMean RThe othersDifference
OTE 0.618-0.7865334+0.04 ± 0.49−0.13 ± 0.20 (182)+0.17 R
0.382-0.58161−0.15 ± 0.33−0.08 ± 0.22 (155)−0.07 R
0.5-0.6186146+0.07 ± 0.41−0.15 ± 0.20 (170)+0.21 R
0.786-14428+0.05 ± 0.54−0.12 ± 0.19 (188)+0.18 R
Fake band 0.45-0.556650−0.02 ± 0.38−0.13 ± 0.21 (166)+0.11 R
Fake band 0.2-0.37252−0.17 ± 0.36−0.08 ± 0.21 (164)−0.09 R
Fake band 0.7-0.83724+0.11 ± 0.59−0.13 ± 0.19 (192)+0.23 R

The OTE band has a mean R of +0.04 on 34 trades, but with an error of ±0.49: the whole interval from −0.45 to +0.53 is compatible. The difference with the other POI is +0.17 R. If membership of the OTE band is assigned at random to 53 of the POI (2,000 times), the median difference is 0, the 5th percentile −0.39, the 95th +0.40: 25% of the trials do at least +0.17. The fake bands (0.45-0.55, 0.7-0.8) give differences between +0.11 and +0.23 R, one band (0.2-0.3) −0.09: the same order of magnitude as the OTE band, and the highest is in fact a fake band.

Figure 1 shows a POI inside the band that wins: a demand FVG of 2 August 2026 at 04:00 (zone 4,062.00-4,068.17 $), in the 0.67-0.74 band of the leg from 4,039.00 to 4,126.64 $ of 30 July. The order at 4,068.17 $ fills at 08:00, the stop is at 4,058.74 $ and the target at 4,087.02 $ is reached at 20:00: +1.97 R. Figure 2 shows one inside the band that loses: a supply FVG of 31 July at 04:00 (zone 4,094.70-4,107.04 $), in the 0.68-0.84 band of the leg from 4,120.44 to 4,039.00 $. The order at 4,094.70 $ fills on 4 August at 12:00 and the stop at 4,116.39 $ is hit on 5 August at midnight: −1.01 R. Figure 3 shows a POI outside all the bands that wins: a supply FVG of 4 February 2026 at 16:00 (zone 4,960.87-5,015.00 $), beyond the start of the leg from 4,884.61 to 4,444.00 $ (position 1.17-1.30), with an order at 4,960.87 $ and the target at 4,798.60 $ reached: +2.00 R.

Figure 1. Demand FVG in the OTE band (position 0.67-0.74 of the leg from 4,039.00 to 4,126.64): order at 4,068.17, target at 4,087.02 reached (+1.97 R).Data: gold, XAUUSDT perpetual future (real candles) · 4 hours (H4) · from 29 Jul 2026 20:00 UTC to 3 Aug 2026 08:00 UTC.
Read the chart in words

H4 gold chart from 29 July to 3 August 2026 with the 50%, 61.8% and 78.6% levels of the leg AB, the POI zone, the entry line, the stop line (4,058.74) and the target line (4,087.02).

WhenWhatPrice
30 Jul 2026 04:00 UTCPoint A4,039.00
30 Jul 2026 12:00 UTCPoint B4,126.64
2 Aug 2026 20:00 UTCtarget reached4,089.31

Times in UTC.

Figure 2. Supply FVG in the OTE band (position 0.68-0.84 of the leg from 4,120.44 to 4,039.00): order at 4,094.70, stop at 4,116.39 hit on 5 August (−1.01 R).Data: gold, XAUUSDT perpetual future (real candles) · 4 hours (H4) · from 29 Jul 2026 08:00 UTC to 5 Aug 2026 12:00 UTC.
Read the chart in words

H4 gold chart from 29 July to 5 August 2026 with the 50%, 61.8% and 78.6% levels of the leg AB, the POI zone, the entry line, the stop line (4,116.39) and the target line (4,051.32).

WhenWhatPrice
29 Jul 2026 16:00 UTCPoint A4,120.44
30 Jul 2026 04:00 UTCPoint B4,039.00
5 Aug 2026 00:00 UTCstop hit4,142.20

Times in UTC.

Figure 3. Supply FVG outside the bands (position 1.17-1.30, beyond the start of the leg from 4,884.61 to 4,444.00): order at 4,960.87, target at 4,798.60 reached (+2.00 R).Data: gold, XAUUSDT perpetual future (real candles) · 4 hours (H4) · from 1 Feb 2026 16:00 UTC to 5 Feb 2026 12:00 UTC.
Read the chart in words

H4 gold chart from 1 to 5 February 2026 with the 50%, 61.8% and 78.6% levels of the leg AB, the POI zone, the entry line, the stop line (5,042.00) and the target line (4,798.60).

WhenWhatPrice
2 Feb 2026 00:00 UTCPoint A4,884.61
2 Feb 2026 04:00 UTCPoint B4,444.00
5 Feb 2026 00:00 UTCtarget reached4,793.17

Times in UTC.

Procedure in four steps

  1. Define the reference leg with a written rule, known when the POI is complete.
  2. Compute the position of the zone and count the POI in the band: few cases, large error.
  3. Compare the trades inside and outside, with the 95% interval.
  4. Compare with fake bands and with a random draw of the membership.

Exercises

Chapter exercises

1. A demand leg goes from A = 1,000 to B = 1,200 (hypothetical numbers). A demand FVG has its zone between 1,080 and 1,090. In which band does it sit (0 = B, 1 = A)?

2. Of the 291 POI, 53 overlap the OTE band and 238 do not. What must be done before comparing the results?

3. The OTE band gives +0.04 R against −0.13 R for the other POI (+0.17 R), and 25% of random draws do at least +0.17. What do you conclude?

4. Fake bands (0.7-0.8, 0.45-0.55) give differences of +0.23 and +0.11 R, like the OTE band (+0.17). What does it mean?

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.