TRADING SYSTEMS
FX
EUR/USD low level frequency at a neighbor count of 26 is outside what is expected from randomness within a 99% confidence – the fact that the difference with the largest random series result is small plus the fact that the curve then drops close to the random series average at even larger neighbor counts suggests that these results are only spurious.
Figure 3. Comparison between candlestick low neighbor counts for different real data sets and their random. Real data results are shown in blue, random data results are shown in pink.
amount of convergence of the highs around certain levels is in no way special compared to what we expect from simple randomness given the distribution of returns of the real data sets. As a matter of fact it becomes clear through this analysis that the random data does have an important variety of cases where high values show a stronger appearance of “resistance”. Tere are cases where frequencies for high neighbor counts are significantly higher than those of the real data.
In the case of the lows – shown in Figure 3 – the phenomena is significantly similar. We do tend to
see a significant accumulation of high frequencies for high neighbor counts in the real data – which is what leads us to see support levels in the charts – but in almost all cases the real data remains within the possibilities of the random data. It is worth noting that in some cases we do see that the levels tend to get close to the higher band of the random data and in a few cases it even goes above it. Nonetheless this difference is small enough and infrequent enough to prevent us from making any strong differentiation between the random and real data. Even if some punctual distinctions could be made – for example that the
It is clear through these results that a general distinction between the presence of S&R levels in randomly generated data and real market data cannot be made within any reasonable confidence for EUR/USD, GBP/ USD, USD/JPY and GBP/JPY daily data, given the definitions of S&R and the size of the S&R zones mentioned above. It is also important to point out that the above does not imply that S&R trading has no value or that no statistically relevant S&R measurements can be found. However what the above research does show is that the definition of S&R used should be considered carefully since there is an important probability that a given definition ends up generating results that are in practice indistinguishable from those attributable to randomness. A trader interested in using S&R zones will benefit from testing the intended definition using this type of analysis to figure out whether the results are really outside of what is expected from chance within a high enough confidence.
Daniel Fernández
Founder of
Asirikuy.com and
Mechanicalforex.com
FX TRADER MAGAZINE October - December 2017 61
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70
