![](bg7.png)
and taxes. Unfortunately, the using of the MH indica-
tor generates a large number of transactions.
We simulated a real investment of 2000 into NAS-
DAQ for a start, and we used the fee of 1% for each
transaction (BUY or SELL) because it corresponds
with the reality. After the whole period (i.e.,1996-
2014), we finished with 7819.8, and we spent 34497.3
on fees. To make a picture clear, using Buy & Hold
on NASDAQ, we would finish with 16648.1, and we
would pay only 187.7 for fees.
Tax rules are different in different countries, e.g.,
between 15% − 25%, but it is evident that they re-
duce the profit, too. Usually, the tax has to be paid
immediately after the transaction is finished. So, the
reinvestment of profit is reduced.
However, the transaction fees depend on the stock
exchange provider. Big investors can use the strategy
of scalping that represents many thousands transac-
tions in a day. Such investors are classified as market
makers, and they are not charged by transactions fees.
The more detailed explanation is out of the scope of
our paper.
6 CONCLUSIONS
The goal of our investigation was to develop and test
a new indicator MH for technical analysis based on
chaos measure represented by Hurst exponent of the
underlying time series of prices.
We found a construction described in Section 4.1,
and we evaluated it in comparison to the strategy us-
ing MACD or Buy & Hold for data described in Sec-
tion 5.
Using hypothesis testing, we proved our hypothe-
sis that the new MH indicator developed in this work,
i.e., our non-linear method described in Section 4.1,
generates more profit compared to the MACD techni-
cal indicator and to the Buy & Hold investment strat-
egy.
On DAX, MH was 4.5 times better than
Buy & Hold and 7.2 times better than MACD. On
NASDAQ, it was 2.9 times better than Buy & Hold
and 16.8 times better than MACD.
In Subsection 5.2, we explained why the indicator
MH cannot be used as a money generating machine.
However, we believe that complex, non-linear sys-
tems with problematic stationarity are an important
research topic. More research has to be done to an-
swer questions about filtering of BUY- and SELL-
signals. In our future research, we will apply meth-
ods of genetic programming to improve it like in our
previous work (Kroha and Friedrich, 2014).
REFERENCES
Brooks, C. and Persand, G. (2003). Volatility forecasting for
risk management. Journal of Forecasting, 22(1):1–22.
Casdagli, M. (1991). Chaos and deterministic versus
stochastic nonlinear modeling. Technical report,
SFI Working Paper: 1991-07-029, www.santafe.edu
(download 02-26-2017).
Falconer, K. (1990). Fractal Geometry: Mathematical
Foundations and Applications. Wiley.
Fama, E. (1970). Efficient capital markets: A review of the-
ory and empirical work. Journal of Finance, 25:383–
417.
Gleick, J. (1987). Chaos: Making a new Science. Viking
Books.
Gneiting, T. and Schlather, M. (2004). Stochastic models
that separate fractal dimension and the hurst effect.
SIAM Rev, 46:269–282.
Gneiting, T., Sevcikova, H., and Percival, D. B. (2012). Es-
timators of fractal dimension: Assessing the rough-
ness of time series and spatial data. Statistical Science,
27:247–277.
Goldstein, D. G. and Taleb, N. N. (2007). We don’t
quite know what we are talking about when we talk
about volatility. Journal of Portfolio Management,
33(4):84–86.
Hausdorff, F. (1919). Dimension und
¨
außeres maß. Math.
Ann, 79:157–179.
Hurst, H. E. (1951). The long-term storage capacity of
reservoirs. Transaction of the American Society of
Civil Engineers, 116.
Kaplan, I. (2003). Estimating the hurst exponent - c++ soft-
ware source. 1:12.
Kri
ˇ
stoufek, L. (2010). Rescaled range analysis and de-
trended fluctuation analysis: Finite sample properties
and confidence intervals. AUCO Czech Economic Re-
view, 4(3):315–329.
Kroha, P. and Friedrich, M. (2014). Comparison of genetic
algorithms for trading strategies. In Current Trends
in Theory and Practice of Computer Science - SOF-
SEM 2014, Lecture Notes in Computer Science, vol-
ume 8327, pages 383–394. Springer.
Kroha, P. and Lauschke, M. (2012). Fuzzy and fractal tech-
nology in market analysis. In Studies in Computa-
tional Intelligence - IJCCI 2010, volume 399, pages
247–260. Springer.
Lo, A. W. and MacKinlay, A. C. (1999). A Non-Random
Walk Down Wall Street. Princeton University Press.
Lorenz, E. N. (1963). Deterministic non-periodic flow.
Journal of the Atmospheric Sciences, 2(20):130–141.
Mandelbrot, B. (1963). New methods in statistical eco-
nomics. Journal of Political Economy, 71:421–440.
Mandelbrot, B. (1997). Fractals and Scaling in Finance.
Springer.
Mandelbrot, B. and Hudson, R. L. (2004). The Misbehavior
of Markets. Basic Books.
Mitra, S. K. (2012). Is hurst exponent value useful in fore-
casting financial time series? Asian Social Science,
8:8.
Hurst Exponent and Trading Signals Derived from Market Time Series
377