Directional-change intrinsic time

Directional-change intrinsic time is an event-based operator to dissect a data series into a sequence of alternating trends of defined size $\delta$ .

File:Directional-change dissection procedure.png

The directional-change intrinsic time operator was developed for the analysis of financial market data series. It is an alternative methodology to the concept of continuous time.{{Cite journal|last1=Guillaume|first1=Dominique M.|last2=Dacorogna|first2=Michel M.|last3=Davé|first3=Rakhal R.|last4=Müller|first4=Ulrich A.|last5=Olsen|first5=Richard B.|last6=Pictet|first6=Olivier V.|date=1997-04-01|title=From the bird's eye to the microscope: A survey of new stylized facts of the intra-daily foreign exchange markets|journal=Finance and Stochastics|language=en|volume=1|issue=2|pages=95–129|doi=10.1007/s007800050018|issn=0949-2984|doi-access=free}} Directional-change intrinsic time operator dissects a data series into a set of drawups and drawdowns or up and down trends that alternate with each other. An established trend comes to an end as soon as a trend reversal is observed. A price move that extends a trend is called overshoot and leads to new price extremes.

Figure 1 provides an example of a price curve dissected by the directional change intrinsic time operator.

The frequency of directional-change intrinsic events maps (1) the volatility of price changes conditional to (2) the selected threshold $\delta$ . The stochastic nature of the underlying process is mirrored in the non-equal number of intrinsic events observed over equal periods of physical time.

Directional-change intrinsic time operator is a noise filtering technique. It identifies regime shifts, when trend changes of a particular size occur and hides price fluctuations that are smaller than the threshold $\delta$ .

Application

The directional-change intrinsic time operator was used to analyze high frequency foreign exchange market data and has led to the discovery of a large set of scaling laws that have not been previously observed.{{Cite journal|last1=Glattfelder|first1=J. B.|last2=Dupuis|first2=A.|last3=Olsen|first3=R. B.|date=2011-04-01|title=Patterns in high-frequency FX data: discovery of 12 empirical scaling laws|journal=Quantitative Finance|volume=11|issue=4|pages=599–614|doi=10.1080/14697688.2010.481632|arxiv=0809.1040|s2cid=154979612|issn=1469-7688}} The scaling laws identify properties of the underlying data series, such as the size of the expected price overshoot after an intrinsic time event or the number of expected directional-changes within a physical time interval or price threshold. For example, a scaling relating the expected number of directional-changes $N(\delta)$ observed over the fixed period to the size of the threshold $\delta$ :

$N(\delta) = \left( \frac{\delta}{C_{N, DC}} \right)^{E_{N, DC}}$ ,

where $C_{N, DC}$ and $E_{N, DC}$ are the scaling law coefficients.{{Cite journal|last1=Guillaume|first1=Dominique M.|last2=Dacorogna|first2=Michel M.|last3=Davé|first3=Rakhal R.|last4=Müller|first4=Ulrich A.|last5=Olsen|first5=Richard B.|last6=Pictet|first6=Olivier V.|date=1997-04-01|title=From the bird's eye to the microscope: A survey of new stylized facts of the intra-daily foreign exchange markets|journal=Finance and Stochastics|language=en|volume=1|issue=2|pages=95–129|doi=10.1007/s007800050018|issn=0949-2984|doi-access=free}}

Other applications of the directional-change intrinsic time in finance include:

trading strategy characterised by the annual Sharpe ratio 3.04{{cite book

| last1 = Golub | first1 = Anton

| last2 = Glattfelder | first2 = James B.

| last3 = Olsen | first3 = Richard B.

| editor1-last = Dempster | editor1-first = M. A. H.

| editor2-last = Kanniainen | editor2-first = Juho

| editor3-last = Keane | editor3-first = John

| editor4-last = Vynckier | editor4-first = Erik

| contribution = The Alpha Engine: Designing an Automated Trading Algorithm

| date = February 2018

| doi = 10.1201/9781315372006-3

| isbn = 9781315372006

| pages = 49–76

| publisher = Chapman and Hall/CRC

| ssrn = 2951348

| title = High-Performance Computing in Finance}}

tools designed to monitor liquidity at multiple trend scales.{{Cite journal|last1=Golub|first1=Anton|last2=Chliamovitch|first2=Gregor|last3=Dupuis|first3=Alexandre|last4=Chopard|first4=Bastien|date=2016-01-01|title=Multi-scale representation of high frequency market liquidity|journal=Algorithmic Finance|language=en|volume=5|issue=1–2|pages=3–19|doi=10.3233/AF-160054|issn=2158-5571|doi-access=free|arxiv=1402.2198}}

The methodology can also be used for applications beyond economics and finance. It can be applied to other scientific domains and opens a new avenue of research in the area of BigData.

References

50px Text in this draft was copied from {{cite journal |last1=Petrov |first1=Vladimir |last2=Golub |first2=Anton |last3=Olsen |first3=Richard |title=Instantaneous Volatility Seasonality of High-Frequency Markets in Directional-Change Intrinsic Time |journal=Journal of Risk and Financial Management |date=2019 |volume=12 |issue=2 |pages=54 |doi=10.3390/jrfm12020054 |language=en|doi-access=free |hdl=10419/239003 |hdl-access=free }}, which is available under a [https://creativecommons.org/licenses/by/4.0/ Creative Commons Attribution 4.0 International License].

Category:Data analysis