Applications of technical analysis in the design of trading strategies on cryptocurrencies

S preprostimi orodji tehnične analize in s preprostimi postopki signaliziranja priložnosti lahko sestavimo uspešno trgovalno strategijo, ki bo zagotavljala dolgoročno pozitivne denarne tokove in s tem preživetje tudi v necvetočih obdobjih ter obdobjih nepričakovanih tržnih šokov. Zaradi heteroskedastičnosti finančnih časovnih vrst moramo upoštevati tržno volatilnost, ki predstavlja velikost cenovnega gibanja in posledično privzeto tveganje. Ampak kako vemo, da bo strategija dosegala dobre donose? Pravzaprav ne moremo, saj ne poznamo prihodnjih cenovnih vrednosti, lahko pa jo testiramo za nazaj in ob tem razvijemo metodologijo za merjenje uspešnosti, ki predstavlja rezultate trgovanja na zgodovinskih podatkih ter z njimi ugotovimo empirično učinkovitost strategije. Pri interpretaciji izidov virtualnega trgovanja moramo biti pozorni na predpostavke. Pomembni sta predvsem poznavanje zgodovine in zanemarjanje operativnih stroškov. Testiramo lahko samo strategije, v katerih so vključeni programsko določljivi vzorci in indikatorji. Na primer, preprosti postopki kot so razdeljevalna metoda razvrščanja s $k$-voditelji in regresija s pomočjo jedra določijo podpore in odpore, drseče sredine, ki so tudi sestavni del mnogih indikatorjev, določijo količino sredstev in vstopne signale. S testiranjem za nazaj izvedemo tudi morebitno optimizacijo, ki je pomemben del razvoja in analize določene strategije. Mnogo parametrov osnovnih pravil se da malenkostno spremeniti in rezultati postanejo veliko boljši. Z ugotovitvami analiz in z vsemi dodatnimi trenutnimi informacijami, dodatnimi orodji ter zdravim razumom si lahko ustvarimo prihodnja pričakovanja uspešnosti ter naložbene priložnosti.With simple technical analysis tools and simple procedures we can build a successful trading strategy that will provide positive long-term cash flows and ensure zero probability of failure in the periods of non-flourishing markets and in the periods of unexpected market shocks as well. In order to consider financial time series heteroscedasticity we must include volatility, which represents the size of the price movements and a capital risk. How do we know that the built strategy will achieve great returns? Actually we do not, because we do not know the future price movement. What we can do is to backtest our rules on the historical data and develop a methodology for measuring performance. The outcomes represent historical trading profits, which serve for empirical efficiency assessment of the built strategy. When we are interpreting the outcomes we must be careful about assumptions. There are two very important assumptions - knowing the history and the assumption of zero operational costs. Only strategies with technical analysis indicators and patterns identifiable by programming code can be backtested. For instance, simple algorithms like $k$-means clustering and kernel regression determine supports and resistances, while moving averages that are also a part of many indicators, define position sizing and entry signals. Backtesting is also used for optimisation. Optimisation is an important process in strategy building and/or strategy analysis. Plenty of original rules\u27 parameters can be slightly modified to provide much better outcomes. In the end, all analysis results, current information, additional tools, common sense etc. are used for future expectation about strategy performance and therefore for trading opportunities