Stefanel Radu https://alpha-week.com/ en Directional Data Biases and Systematic Trend Following: Friends or Foes? https://alpha-week.com/directional-data-biases-and-systematic-trend-following-friends-or-foes <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--title--features.html.twig x field--node--title.html.twig * field--node--features.html.twig * field--title.html.twig * field--string.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'core/modules/node/templates/field--node--title.html.twig' --> <span>Directional Data Biases and Systematic Trend Following: Friends or Foes?</span> <!-- END OUTPUT from 'core/modules/node/templates/field--node--title.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--created--features.html.twig x field--node--created.html.twig * field--node--features.html.twig * field--created.html.twig * field--created.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'core/modules/node/templates/field--node--created.html.twig' --> <span>Tue, 03/14/2023 - 11:19</span> <!-- END OUTPUT from 'core/modules/node/templates/field--node--created.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--body--features.html.twig * field--node--body.html.twig * field--node--features.html.twig * field--body.html.twig * field--text-with-summary.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-body field--type-text-with-summary field--label-hidden field__item"><p><span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Living in a world where (almost) everything seems familiar gives us a sense of comfort and at least <em>partial</em> security. The result is that we perceive ourselves to be in a privileged position where we can peacefully anticipate what may happen next in life. Surely apples can only fall from their tree onto Earth in the same way each time – this makes sense to us. Physics appears to rely on such an observation as one of its fundamental principles<sup>1</sup> but dig a little deeper and it also caters for potentially significant fluctuations in the gravitational field<sup>2</sup>. For a long time, we believed that light could only travel in a straight line<sup>3</sup>, until physicists deduced<sup>4</sup> and then confirmed that it can bend around very dense objects such as black holes<sup>5</sup>.</span></span></span></span></span></span></span></p> <p><span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>There are many principles or suppositions that appear to define our <em>current</em> reality, whether physical or otherwise and, quite routinely almost, these statements tend to be viewed as absolute truths. Big words indeed, but we appear all too often to rely almost exclusively on experience gained from the past, frequently the <em>recent past</em>, to deduce what may come next. However, life can (and not infrequently does) change in ways unforeseen by our experience. Why does this happen, and should we worry? </span></span></span></span></span></span></span></p> <p><span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>The principles spanning our world are described by the often-subjective way we detect and use them in real life. Do they change? That is a big subject<sup>6,7</sup>, but it suffices us to say, for the purposes of this paper, that<em> phenomena</em> obeying such fundaments can occur in manners that are not always expected by prior empirical evidence. This is due to the <em>open nature</em> of our world which evolves as it ticks away, driving us towards new states and allowing us to witness and ideally learn from new, uncharted experiences.</span></span></span></span></span></span></span></p> <p><strong><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Data Biases in Markets</span></span></span></span></span></strong></p> <p><span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Markets are a perfect example of an open system as they interact with the ‘exterior’ by continuously exchanging information. There is an almost encyclopaedic list of components that can change in the marketplace in ways that alter the dynamic of the price action. Such factors include, sensitivity to interest rates, market liquidity, earnings, momentum, downside risk, geopolitical developments and many more - the list is ever-growing, which is consistent with the nature of an open system. Moreover, the market participants themselves change over time and when they do, they introduce or remove their own views of the instruments they trade. When the ensemble of those individual views generates <em>on average</em> a particular directionality in the price action over a sufficiently long period of time, a directional data bias is established in that market. We set forth below some examples.</span></span></span></span></span></span></span></p> <ol><li><span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>The 40-year bull run in the US Treasury Long Bond Future between early 1981 and the end of 2020 (see Fig. 1). This price action generated a bullish bias in the underlying distribution of returns across a 40-year period. The tendency of the market was to rise persistently within that time, with occasional corrections along the way.</span></span></span></span></span></span></span></li> </ol><p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="220e37df-1091-45dd-8135-5b5fc18b8858" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-A.png" width="550" height="254" loading="lazy" /></p> <p><em><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US">Fig.1. The 40-year bull run in the US Treasury Long Bond Future between the beginning of 1981 and the end of 2020. </span></span></span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US">Source: Bloomberg.</span></span></span></em></p> <ol start="2"><li><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>The 22-year bear run in the 100oz Gold Future between January 1980 and the end of 2001 (see Fig.2). Within this period the price of gold depreciated continuously, establishing a bearish data bias in the returns profile of this market during that time. Although several mini rallies against the prevailing bearish trend occurred, their presence in the dynamic of the price action did not nullify the bearish bias of this market over the timeframe shown.</span></span></span></span></span></span></li> </ol><p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="cfcdc897-1845-4ca7-9157-d61e5c7c398d" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-B.png" width="550" height="255" loading="lazy" /></p> <p><em><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US">Fig.2. The long bear run in the 100oz Gold Future Contract between late 1979 and the end of 2001. Source: Bloomberg.</span></span></span></em></p> <ol start="3"><li><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>The </span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>47-year bull run in the S&amp;P500 Index (see Fig.3) between the beginning of 1975 and the end of 2021. The price action is seen advancing across this time range despite several setbacks, such as crashes and corrections, that take place along the way. Notwithstanding those setbacks, the market returns exhibit a bullish data bias over this long period.</span></span></span></span></span></span></li> </ol><p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="6778623f-1cb0-410b-954d-4c1bcb3a411e" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-C.png" width="550" height="252" loading="lazy" /></p> <p><em><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US">Fig.3. The 47-year bull run in the S&amp;P500 Index between early 1975 and the end of 2021. Source: Bloomberg.</span></span></span></em></p> <p><strong><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Impact on Building Systematic Trading Models</span></span></span></span></strong></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>When building a systematic trading model, including trend following models, one often has the option of fitting such a model to the data available in the timeseries of the markets intended for trading. This is in fact one of the most frequently employed methods of putting together systematic strategies, at least as a first iteration. It follows naturally that the directional biases of the price series present in the underlying markets making up the trading universe could ‘spill over’ into the behaviour of such a model. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>What does this mean in practice? Let’s consider, for example, the case of the 40-year bull run of the US Treasury Long Bond Future. If the data available for modelling is only the period shown in Fig.1 (i.e., January 1981 to December 2020), a first iteration, easy-to-fit model would expect bull markets more frequently than bear markets. This property of expecting bull markets over bear markets would be amplified in the case of medium-term trend following where the average holding period tends to be of the order of several months. The directional bias of the price action shown in Fig.1 increases in strength as the time-horizon of the momentum measure increases. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Is this problematic? The answer to this question is not necessarily binary. It really depends on where we are in the cycle of the current market regime. For instance, if around the year 2000 we built a data-fitted medium-term trend following model relying strongly on the underlying bullish bias shown in Fig.1 and then let that model run for the next 20 years, such a model would be expected to do well. However, the same approach may not necessarily have worked well during the subsequent 2 years when the market turned bearish for most of this new period (i.e., late 2020 to the end of 2022 – see </span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Fig.1a). At a glance, the price action behaviour in 2021 and 2022 appears disconnected from the previous 40-year long price action.</span></span></span></span></span></span></p> <p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="a1fa8392-0778-487a-901e-9bf4e0b301b3" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-D.png" width="550" height="252" loading="lazy" /></p> <p><em><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.1a. Bear market behaviour emerging at the end of 2020, lasting for at least 2 years in the US Treasury Long Bond </span></span></span></span></span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Future (as of December 2022). Source: Bloomberg.</span></span></span></span></span></em></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Similarly, if we used only the data set (or a subset of it) associated with Fig.2 above describing the 22-year long bearish price action in the 100oz Gold Future to build a data-fitted medium-term trend following model, we could end up with a strategy that expects bear markets more often than bull markets as a result of relying on the strong bearish bias of the returns’ distribution. Such an approach would work well as long as the bear persists. The problem is that the subsequent 10 years after 2001 (the end of the bear run) produced a strong bull run (see Fig.2a) where a model with a strong bearish bias would likely not work optimally because it does not expect a persistent bull market regime.</span></span></span></span></span></span></p> <p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="02600f77-7dc4-4ca5-b195-8277057304f6" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-E.png" width="550" height="252" loading="lazy" /></p> <p><em><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.2a. 10-year long bull run in the 100oz Gold Future following the 22-year long bearish market regime that occurred </span></span></span></span></span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>previously (see also Fig.3). Source: Bloomberg.</span></span></span></span></span></em></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>In the case of the price behaviour shown in Fig.3 (the multi-decade bull run in the S&amp;P500 Index), the situation is slightly more mixed. Despite the visibly persistent long-term trend which produced the bullish bias in the </span></span></span></span></span></span><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>distribution of returns, there are also other types of behaviour that are noticed: there are several crashes (1987, 2000, 2008 and 2020) and some bear markets (e.g., the period 2000-2003). Despite this mix of regimes, a first-iteration data-fitting optimization for a trend-following model would still naturally result in a strategy that expects bull markets and may find it hard to adapt to or recognize in a timely manner bearish price actions, particularly of the type that occur after this long bull run, in 2022 (see Fig.3a below). However, despite their relatively less frequent presence in the timeseries of returns, the occasional bear markets or sharp corrections may give insightful clues as to how to more comprehensively approach the problem of modelling a market timing engine for markets that exhibit strong directional biases. See further below.</span></span></span></span></span></span></p> <p><img alt="DG Partners" data-entity-type="file" data-entity-uuid="bcb5ea22-f3e7-4451-9dc5-ac9a21ddae06" src="https://www.alpha-week.com/sites/default/files/inline-images/DG-Partners-2023-03-14-F.png" width="550" height="252" loading="lazy" /></p> <p><em><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.3a. Persistent bearish behaviour in the S&amp;P500 Index occurring in 2022. Source: Bloomberg.</span></span></span></span></span></em></p> <p><strong><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Dealing with Directional Data Biases </span></span></span></span></strong></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>We have seen in the examples discussed above how prolonged, directionally continual price actions can generate data biases in the distribution of returns. We have also seen how these biases may not prepare the systematic trader for a change to a persistent market regime not seen or seldomly seen in the training data set. If the training set <em>and </em>what unfolds in reality in future turn out to be similar in directionality, one could argue that the data bias would in fact be the systematic momentum trader’s best friend. In such a situation, incorporating the data bias into one’s modelling would be likely to create a precious edge. However, this will hold true only for as long as the market regime that generated the directional data bias in the training set also persists in live trading. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>This has proven particularly true for Fixed Income and Equities for several decades (as shown in Fig.1 and Fig.3 above), wherein both asset classes showed a strong bullish bias. Such long and persistent upward moves greatly benefited the so-called risk parity trade and encouraged its adoption and subsequent ubiquity. So, one may ask: where is the problem? The issue is that even long-term market patterns or regimes may run their course or unravel without warning when markets enter trading phases not seen in recent history, or never seen before. By way of recent example, the risk parity trade began to unravel in 2021 and later in 2022 it no longer worked. As a result, data-fitted systematic trading models that were expecting bull markets to a greater degree than bear markets in Fixed Income and Equities would prove less effective. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Is there a solution? And, more importantly, could one find a solution <em>ahead</em> of the shift in the market regime? Ideally, such a solution would be something that works independently of the historical market biases. One possible approach would be to look for inspiration in the manner in which scientists generalise the patterns seen in nature with theories or concepts that do not rely solely on empirical evidence (the corresponding feature in market price action being the directional bias). </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Such theoretical conceptualisation must not be disconnected from reality: it should still be able to explain any experimental results (e.g., the measured distribution of returns) but, critically, it should also be able to generalise such a distribution to incorporate patterns never seen before but which could in principle take place at some point in the future. The rather infrequent but highly noticeable corrections or bear markets occurring in the price action of the S&amp;P500 Index (see Fig.3) can be ‘expanded’ theoretically in a dynamical way to cover longer timeframes. The same holds true for the chart of the 100oz Gold Future in Fig.2 - the occasional bear squeezes and counter-trend rallies seen here can guide the researcher to infer the potential formation of persistent market regimes that may contrast with the bear market depicted in Fig.2. Moreover, such a conceptual generalisation of market price action behaviour could be designed to be equally disposed to expect bear or bull markets at any time and, in essence, not to depend on any underlying directional data biases that exist in the market universe.</span></span></span></span></span></span></p> <p><strong><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Concluding Remarks</span></span></span></span></strong></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>This paper considered the influence that directional data biases may have in developing systematic trend-following (momentum) models. We have seen that a directional bias can be an excellent companion if the training set and the subsequent live data share the same directional drift. However, a long-term data bias can be a serious impediment for a strategy relying on data-fitting to adapt to a new and persistent market regime that exhibits a medium-term direction contrary to the expected bias. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>We have also suggested how one might deal with these types of directional change. Conceptual in make-up, such a solution would be predicated on generalising the distribution of returns to include patterns never seen before by the market, but which could happen at some point in the future, with a view to diluting and even eliminating the data bias observed in the measured distribution of returns. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Such a conceptual approach to systematic trading is similar in nature to the manner in which physical theories constantly generalise measured observations (and other previously established views) to expect new patterns - new events that do not look likely to occur from the empirical evidence alone. </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Such theories may prepare us for the unexpected. But should we <em>always</em> expect the unexpected? Should we always think that changes that seem implausible today may occur at any time? Is such an outlook a little unsettling even, as it leads to no persistent state of equilibrium? </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>While a conceptual solution may well outperform a data-fitted model at a time when the market moves against its historical bias, that very solution could well underperform a strategy relying on data-fitting when the market behaviour may again re-align with its long-term directional bias. Is there perhaps a place for <em>both types</em> of strategy, data-driven and conceptual? </span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>We explained in a previous paper<sup>8</sup> that price actions seem to exhibit more nonlinearity than before, with market regimes previously taking many years to develop now being confined to the extent of just a few months. Within a given time range, such nonlinear behaviour has the potential to generate significantly more market regimes compared to historical averages. In such a dynamically changing environment, we could experience alternating market modes, some of them observing their long-term directional bias and some disobeying it. This prospect seems to argue in favour of needing both types of systematic momentum strategy: models relying on the directional data bias of the markets traded and models independent of such a bias. In such a construct, a portfolio-level utility function that seeks to optimize the mix of these <em>two classes of models</em> in a highly energetic market environment becomes worthy of attention and perhaps the next challenge to solve for systematic trading researchers. Those CTAs that can tackle such a complex task are expected to outperform in the future.</span></span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>**********</span></span></span></span></span></span></p> <p><strong><em>Stefanel Radu</em></strong><em> is Head of Research at </em><a href="https://www.bhdgsystematic.com/"><strong><em>BHDG Systematic Trading</em></strong></a><em> / </em><a href="https://www.dgpartners.co.uk/"><strong><em>DG Partners</em></strong></a></p> <p><sup><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>References</span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>1</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Isaac Newton, <em>The Principia: Mathematical Principles of Natural Philosophy</em>, translation from the Latin original by Bernard Cohen and Anne Whitman, University of California Press, 19 February 2016</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>2</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>R. Brandenberger, V. Mukhanov, and T. Prokopec, <em>Entropy of the Gravitational Field</em>, Physical Review D <strong>48</strong> 2443 – Published 15 September 1993.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>3</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>J.L Heiberg and H. Menge, <em>Euclidis Opera Omnia; Volumen 7</em>, Leipzig 1895.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>4</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Johann Georg von Soldner, <em>On the Deflection of a Light Ray from its Rectilinear Motion by the Attraction of a Celestial Body Which it Passes Close</em> by (Lenard, P.) in Annalen der Physik <strong>65</strong> (15): 593-604 in 1801.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>5</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>D.R. Wilkins, L.C. Gallo, E. Constantini, W.N. Brandt and R.D. Blandford, <em>Light Bending and X-ray Echoes from Behind a Supermassive Black Hole, </em>Nature <strong>595</strong>, 657-660 (2021) – Published 28 July 2021.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>6</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Heraclitus, <em>Fragments </em>– translation by Brooks Haxton, (Penguin Classics), 28 October 2003.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>7</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>Aristotle, <em>Physics</em>, by C.D.C. Reeve, Hackett Publishing Company, Inc (2 March 2018).</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>8</span></span></span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>S. Radu, <em>Noise and Nonlinearity: Challenge and Opportunity for Momentum Trading</em>, ALPHAWEEK, 12 August 2021.</span></span></span></span></span></span></sup></p> <p><sup><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>***</span></span></span></span></span></span></sup></p> <p><em>The views expressed in this article are those of the author and do not necessarily reflect the views of AlphaWeek or its publisher, The Sortino Group</em></p> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-tags--features.html.twig * field--node--field-tags.html.twig * field--node--features.html.twig * field--field-tags.html.twig * field--entity-reference.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-field-tags field--type-entity-reference field--label-hidden field__items"> <div class="field__item"><a href="/hedge-funds-managed-futures" hreflang="en">Hedge Funds - Managed Futures</a></div> <div class="field__item"><a href="/directory/dg-partners" hreflang="en">DG Partners</a></div> <div class="field__item"><a href="/hedge-funds-guest-articles" hreflang="en">Hedge Funds Guest Articles</a></div> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <drupal-render-placeholder callback="flag.link_builder:build" arguments="0=node&amp;1=9963&amp;2=bookmark" token="MOFNY7WTbAHaJFRrZpIE15zTTjGMOIXVE62QdZSf_iM"></drupal-render-placeholder> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-author--features.html.twig * field--node--field-author.html.twig * field--node--features.html.twig x field--field-author.html.twig * field--entity-reference.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field--field-author.html.twig' --> <a href="/author/stefanel-radu" hreflang="en">Stefanel Radu</a> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field--field-author.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-content-role--features.html.twig * field--node--field-content-role.html.twig * field--node--features.html.twig * field--field-content-role.html.twig * field--list-string.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-field-content-role field--type-list-string field--label-above"> <div class="field__label">Content role</div> <div class="field__item">Public</div> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> Tue, 14 Mar 2023 11:19:33 +0000 AlphaWeek Staff 9963 at https://alpha-week.com Noise and Nonlinearity: Challenge and Opportunity for Momentum Trading https://alpha-week.com/noise-and-nonlinearity-challenge-and-opportunity-momentum-trading <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--title--features.html.twig x field--node--title.html.twig * field--node--features.html.twig * field--title.html.twig * field--string.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'core/modules/node/templates/field--node--title.html.twig' --> <span>Noise and Nonlinearity: Challenge and Opportunity for Momentum Trading</span> <!-- END OUTPUT from 'core/modules/node/templates/field--node--title.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-image--features.html.twig * field--node--field-image.html.twig * field--node--features.html.twig * field--field-image.html.twig * field--image.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-field-image field--type-image field--label-hidden field__item"> <!-- THEME DEBUG --> <!-- THEME HOOK: 'image_formatter' --> <!-- BEGIN OUTPUT from 'core/modules/image/templates/image-formatter.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'image_style' --> <!-- BEGIN OUTPUT from 'core/modules/image/templates/image-style.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'image' --> <!-- BEGIN OUTPUT from 'core/modules/system/templates/image.html.twig' --> <img loading="lazy" src="https://www.alpha-week.com/sites/default/files/styles/article_full/public/2021-08/BHDG.png?itok=rQijnwNP" width="400" height="225" alt="BHDG" typeof="foaf:Image" /> <!-- END OUTPUT from 'core/modules/system/templates/image.html.twig' --> <!-- END OUTPUT from 'core/modules/image/templates/image-style.html.twig' --> <!-- END OUTPUT from 'core/modules/image/templates/image-formatter.html.twig' --> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--created--features.html.twig x field--node--created.html.twig * field--node--features.html.twig * field--created.html.twig * field--created.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'core/modules/node/templates/field--node--created.html.twig' --> <span>Thu, 08/12/2021 - 13:08</span> <!-- END OUTPUT from 'core/modules/node/templates/field--node--created.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--body--features.html.twig * field--node--body.html.twig * field--node--features.html.twig * field--body.html.twig * field--text-with-summary.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-body field--type-text-with-summary field--label-hidden field__item"><p><span><span><span><span><span>Noise and nonlinearity are, in our view, two important market characteristics that have been on the increase over the past decade and have produced increasingly complex market price actions.</span></span></span></span></span></p> <p><span><span><span><span><span>Not infrequently, difficulties in market direction trading are summed up by stereotypical phrases such as ‘there is too much noise in the market’ or ‘markets are too difficult right now’. What is really meant by this? Clearly, the frequency used by the trader making such statements to capture alpha is not working well. Over such a frequency, ‘noise’ appears to be setting in to such an extent that it ‘blurs’ the trader’s <em>understanding</em> of the market in question. At the same time, another trader, using a different timeframe to form a view of the same market’s direction succeeds and generates a positive return. In this example, what appears to be a hinderance for one person is overcome by another. Moreover, their roles can change in the future <em>without notice</em>: the formerly unsuccessful trader may start to experience successive wins over his or her frequency of choice, while the formerly successful one may start to experience losses over<em> his or her</em> timeframe. What is it in the markets’ dynamics that can give rise to such phenomena? <em>Noise and nonlinearity</em> are, in our view, two important market characteristics that have been on the increase</span></span><sup><a href="#_ftn1"><span><span><span>[1]</span></span></span></a></sup><span><span> over the past decade and have produced increasingly complex market price actions. Trading such price actions has in turn become more and more challenging. This paper aims to raise awareness of this increased complexity, which invariably impacts momentum trading.</span></span></span></span></span></p> <p><span><span><span><strong><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Noise</span></span></strong></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Let’s start off with noise. We hear a lot about it but what is it actually? Noise and information are antonyms. Our understanding of the reality that surrounds us is based on the information available to us. The more genuine information we can access the more knowledge we can infer and establish. Noise is what blurs our understanding and perception and gets in the way of our learning efforts. Having more noise in a signal means there is less information available for analysis and interpretation, making the forecasting of that signal difficult. In the case of markets, this means that decreased information has been carried across the price action. This is particularly relevant for the universe of markets traded by most CTAs which, as academics have pointed out, have become noisier</span></span><sup><a href="#_ftn1"><span><span><span>[1]</span></span></span></a></sup><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>More noise also means less predictability and less stability. After 2008, once central banks started to intervene, markets were constrained from moving as freely as before. That partial lack of freedom led to the proliferation of <em>sideways markets</em>, which is one manifestation of noise that naturally hinders momentum trading as market direction is unclear.  However, sideways markets are not the only (or indeed least favourable) instance where noise has been increasingly felt post 2008. There is an even more dangerous type: <em>low volatility noisy</em> <em>markets</em>. Reversals following a long-term low volatility trend tend to be quite sharp and they can produce large losses very quickly. Both such instances, sideways and low volatility noise are tough challenges for momentum trading or trend following.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Since noisy markets offer less information, they are generally harder to forecast because our access to their underlying dynamic is naturally more limited. This poses a serious challenge when trying to model such markets - but not an insurmountable one. CTAs try to build systematic trading models that are based mostly or solely on price inputs and follow the perceived market direction accordingly. However, if the inputs of these models are contaminated by noise, modelling may require notable enhancements. The work required involves signal processing aimed at “de-noising” or “de-blurring” the inputs. To de-noise the inputs appropriately, you must ensure that when you take out noise, <em>you take out just noise</em> and make certain you leave the structure of the market intact. This is where experience in signal processing becomes very important.  </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>There are many ways to take noise out of a signal, particularly from the price action, and in the right hands this procedure can work well, as long as the person who does this work can distinguish between structure and noise with surgical precision. There is a clear distinction between the two, but the challenge is to see this demarcation in <em>live trading,</em> as markets move. If during the noise reduction process one starts at some point to take out structure, this may in fact weaken the signal even further. Why is that?  A noisy signal already carries decreased information. Consequently, if you reduce what precious little information is available even further, you may find that you have hurt your signal so much that there is even less information available to you than there was originally in the noisy signal. That will likely make things much worse.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>This is why applying signal processing to markets is not a trivial matter. People have been applying de-noising techniques to markets for many years, and many have reported no improvement. In fact, some obtained even worse results. This does not come as a surprise - one must be exceptionally careful when undertaking these procedures.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>I was formerly a mine seismologist. In seismology there are certain methods that can be employed to help distinguish between noise and information <em>on the go</em>. Mine seismologists evaluate the risk of experiencing a hazardous seismic event (e.g., a roof collapse) that can lead to severe consequences in underground mines. Their work involves a lot of data interpretation and their data is almost always noisy. Endless seismograms generated by the movement of rock in the underground mine are continuously processed and interpreted with the aim of estimating where and when the next large event may strike.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Can such scientific techniques be used in markets as well? What do seismology and the study of market behaviour have in common? Not much, except for perhaps one subtle yet surprisingly important factor: both disciplines seek to forecast the behaviour of a particular type of <em>open system</em>. As is well known, an open system exchanges various properties with the exterior beyond its borders such as matter, energy, or information. Seismologists interpret seismograms generated by the movement of rock which dissipates energy <em>away </em>from where the action takes place. Traders seek to understand the direction of markets which are open <em>to receive and dissipate information</em> as they travel. Contrary to a closed system which preserves its energy for instance, an open system does not preserve much of its characteristics and its modelling requires a fundamentally different approach.  Moreover, to understand how a real-life open system evolves you normally need a team of researchers that is multi-disciplinary. That is the case in both seismology and systematic trading. Finally, time is of the essence in both disciplines as they only have one shot each time they try to forecast their respective phenomena under study: seismic events or market movements, respectively.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>An example of how measuring noise in this way can be very helpful in terms of momentum trading or trend following occurred after the so-called “Trump Rally” a few years ago. In late 2017, equity markets were continuously reaching new highs. Towards the end of January 2018, the rally became so strong that markets began to go almost vertical. At the same time the VIX gauge, which measures the volatility of equity markets, continued to stay near its historical lows, i.e., it did not indicate any concerns (see Fig.1). </span></span></span></span></span></p> <p><span><span><span><span><span>​<img alt="BHDG" data-entity-type="file" data-entity-uuid="88d84496-3761-43d4-adbf-af99c1cd1ff2" src="https://www.alpha-week.com/sites/default/files/inline-images/BHDG-2021-08-12-A.png" width="550" height="241" loading="lazy" /></span></span></span></span></span></p> <p><span><span><span><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.1. The S&amp;P500 Stock Index (top panel) and the VIX Index around the time of “Volmageddon” in January/February 2018. Although the VIX gauge is not signaling any danger in late January 2018, the S&amp;P500 price action shows extreme noise which points to price instability ahead. Very early in February 2018 the VIX Index explodes and the S&amp;P500 experiences a large correction. Source: Bloomberg.</span></span></em></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Measuring noise at that point in time with our techniques showed that even though the rally was strong in magnitude and the volatility was low, <em>the noise content was immensely high</em>. It was probably the highest amount of noise we had ever seen in the time series available in our research database at that time. From that perspective, it was clear that the price action could be highly unstable - a strong warning that direction might soon change course. What happened next? VIX exploded in early February 2018 in an event known as “Volmageddon”, the so-called volatility Armageddon. As a result, equities reversed sharply and fell apart, seriously hurting unaware momentum traders, including many trend followers. People were following the price action, it went through the sky, it felt like it was going to go to the moon and then, only days later, it collapsed.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>In this example, volatility did not signal any concerns in an <em>ex-ante</em> way: it was in fact re-assuring the markets that the continuing low volatility rally would persist. According to our measurement process however, the noise content was very high. This illustrates a crucial point: noise and volatility are not the same thing.</span></span></span></span></span></p> <p><span><span><span><strong><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Nonlinearity</span></span></strong></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Nonlinearity is often associated with chaotic behaviour. It refers to the increased sensitivity to small changes in the environment of the process being observed. In the context of markets, this high level of sensitivity relates to changes in the relevant market’s underlying conditions. We are not necessarily referring here to macro environments because often price actions do not respond to a macro theme. We are only referring to the price action itself that can change significantly. Often, we may not even know for certain the real reason or collection of reasons that caused a price change and for systematic trading that is not always important. The interpretation of a macro theme by market participants can be very different from time to time and markets can respond to such themes in surprisingly different ways.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>So, what is the issue here? Just to give you an example, one big manifestation of nonlinearity took place last year, in 2020. When it sets in, nonlinearity leads to the proliferation of distinct yet tradable market regimes that occur in quick succession. Just like noise, nonlinearity has also contributed to the increased complexity seen in the markets’ price action after 2008. When they are progressively nonlinear, markets start to exhibit many different regimes within shorter and shorter timeframes, which is obviously a hurdle for momentum trading and trend following. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Consequently, one needs a way to cope with and adapt to the proliferation of these different market regimes. The case in point is 2020, last year. The year started off with a confident rally in equities. By late February, when the pandemic was not yet declared but concerns about what might happen worldwide existed, equities started to lose their consensus and crashed in March. Despite this by June those same markets that crashed in March were already making new all-time highs. This was spectacular.  Now, markets and indices have crashed before (such as ’87, the “dot com” crash, 2008), but, in the past, for indices to come out of those depths to reach new historical highs took several years. Last year, just months after the crash, they were already exhibiting new highs (see Fig.2). That is a form of nonlinearity. It is important to note that this was not market volatility. Rather, it was a new market regime almost totally disjointed from what had happened in March. The crash happened at very high market volatility while the rally took place with falling volatility, so these were very different market regimes. And they occurred in quick sequence, straight after one another.</span></span></span></span></span></p> <p><span><span><span><span><span>​<img alt="BHDG" data-entity-type="file" data-entity-uuid="4061e38e-76ad-4974-857c-0175c6ddbc02" src="https://www.alpha-week.com/sites/default/files/inline-images/BHDG-2021-08-12-B.png" width="550" height="226" loading="lazy" /></span></span></span></span></span></p> <p><span><span><span><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.2. The Nasdaq100 Stock Index in 2020. The market crashes in March 2020 and reaches the bottom on 23 March. On 6 June the market makes a new all-time high, less than three months since the trough of the crash. Source: Bloomberg.</span></span></em></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Later in the year, we had yet another regime, the commodities rally, a sudden yet massive market movement with commodities belligerently coming out of their bearish price action of the previous decade (see Fig.3). During 2020, a momentum trader or a trend follower had to deal with at least those three market regimes and adapt accordingly!</span></span></span></span></span></p> <p><span><span><span><span><span>​<img alt="BHDG" data-entity-type="file" data-entity-uuid="51455836-4162-4cb7-87f2-fde637ced9c7" src="https://www.alpha-week.com/sites/default/files/inline-images/BHDG-2021-08-12-C.png" width="550" height="226" loading="lazy" /></span></span></span></span></span></p> <p><span><span><span><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Fig.3. The Bloomberg Commodity Index (BCOM Index) since January 2011 till present (2 August 2021). Notice the abrupt bull run that originates in 2020, a very disjoined market regime from the previous decade-long persistent bear market. Source: Bloomberg.</span></span></em></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>When nonlinearity expands in markets, momentum trading must consider the potential for markets to exhibit sudden changes of market regimes which cannot necessarily be captured by volatility models or classical approaches to trend following. You need something else to characterise markets when they change like that - something that deals with the underlying dynamics of the price action itself.</span></span></span></span></span></p> <p><span><span><span><strong><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Concluding remarks</span></span></strong></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>We have described here some of the challenges associated with noise and nonlinearity. In terms of the opportunity set here, how can a momentum trader still benefit from an environment that has become so complex? </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>The trader will have to recognise these complexities and not be intimidated by them. We pointed out that markets are open systems. As a consequence, they are expected to lose and receive information all the time without any warning or notice whatsoever. As long as one is prepared for that, mentally as well as scientifically, one has a chance of dealing with it. If one fails to recognise this problem, the chances of succeeding at forecasting the market direction in a complex environment are random. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Trends still exist. This is encouraging and important to point out, particularly for the managed futures sector. Markets continue to move in a directional way, but trends have become more complicated. One must therefore decode the underlying dynamic that defines the processes governing the formation of these trends. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Select signal processing techniques, such as those used in seismology, may to some extent help identify these market regimes, their onset, and their extent. Such methods can be automated and incorporated into a systematic market-timing engine that can also be used for capturing more complex trends. The secret is to be able to use signal processing in order to distinguish between noise and structure as surgically as possible. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>In terms of nonlinearity, there are specific techniques that can be borrowed from the field of nonlinear dynamics which can help to model the expected sensitivity of the price action when certain changes are detected in the market. Last year was a great example of where trend following had the potential to succeed using this kind of approach to extract at least three types of alpha from markets, including a precious crisis alpha in March 2020. </span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>In our view, those trend followers and other directional traders that recognise and adapt to these challenges are more likely to outperform in the future.</span></span></span></span></span></p> <p><span><span><span><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>**********</span></span></span></span></span></p> <p><span><span><span><strong><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span>Stefanel Radu</span></span></em></strong><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span> is Head of Research at </span></span></em><a href="https://www.bhdgsystematic.com/"><strong><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>BHDG Systematic Trading</span></span></span></em></strong></a><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span> / </span></span></em><a href="https://www.dgpartners.co.uk/"><strong><em><span lang="EN-US" xml:lang="EN-US" xml:lang="EN-US"><span><span>DG Partners</span></span></span></em></strong></a></span></span></span></p> <p><span><span><span><a href="#_ftnref1"><em><span><span><span>[1]</span></span></span></em></a><em><span><span> Salim Lahmiri, Gazi Salah Uddin and Stelios Behiros: Nonlinear dynamics of equity, currency and commodity markets in the aftermath of the global financial crisis, in Chaos, Solitons and Fractals, 103 (2017) 342-346</span></span></em></span></span></span></p> <p><span><span><span><em><span><span>This document has been prepared by BH-DG Systematic Trading LLP (“BH DG”). BH-DG is authorised and regulated by the United Kingdom Financial Conduct Authority and only provides services to professional clients. This document is provided for information purposes only and does not constitute investment advice. Commodity trading involves substantial risk of loss.</span></span></em></span></span></span></p> <p><span><span><span><em><span><span>***</span></span></em></span></span></span></p> <p><em>The views expressed in this article are those of the author and do not necessarily reflect the views of AlphaWeek or its publisher, The Sortino Group</em></p> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-tags--features.html.twig * field--node--field-tags.html.twig * field--node--features.html.twig * field--field-tags.html.twig * field--entity-reference.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-field-tags field--type-entity-reference field--label-hidden field__items"> <div class="field__item"><a href="/hedge-funds-guest-articles" hreflang="en">Hedge Funds Guest Articles</a></div> <div class="field__item"><a href="/hedge-funds-managed-futures" hreflang="en">Hedge Funds - Managed Futures</a></div> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <drupal-render-placeholder callback="flag.link_builder:build" arguments="0=node&amp;1=8580&amp;2=bookmark" token="o1dmANpD7uBsV_0UVovd9D2i2X2_N0MmAQHy6FiVsO4"></drupal-render-placeholder> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-author--features.html.twig * field--node--field-author.html.twig * field--node--features.html.twig x field--field-author.html.twig * field--entity-reference.html.twig * field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field--field-author.html.twig' --> <a href="/author/stefanel-radu" hreflang="en">Stefanel Radu</a> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field--field-author.html.twig' --> <!-- THEME DEBUG --> <!-- THEME HOOK: 'field' --> <!-- FILE NAME SUGGESTIONS: * field--node--field-content-role--features.html.twig * field--node--field-content-role.html.twig * field--node--features.html.twig * field--field-content-role.html.twig * field--list-string.html.twig x field.html.twig --> <!-- BEGIN OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> <div class="field field--name-field-content-role field--type-list-string field--label-above"> <div class="field__label">Content role</div> <div class="field__item">Public</div> </div> <!-- END OUTPUT from 'themes/gavias_vinor/templates/fields/field.html.twig' --> Thu, 12 Aug 2021 12:08:54 +0000 AlphaWeek Staff 8580 at https://alpha-week.com