Evaluating Moving Average Clouds for Equity Trend Continuation

This exploratory research note examines how moving average clouds interact with trend signals, support and resistance levels, and swing structure in equities. The observations are preliminary and do not imply a validated trading edge.

WFDQuant Research Notes | November 2026 An exploratory study of cloud alignment, support and resistance, and proprietary WFDQuant indicators

Introduction

Can familiar moving average relationships become more useful when combined with directional signals, historical support and resistance, and confirmed swing structure? This research note sets out a framework for investigating that question. It presents chart observations and testable hypotheses, not a validated trading strategy or evidence of a repeatable market edge. The starting material consists of four annotated charts and accompanying observations in Shares MA & SR. The screenshots show Agilent Technologies Inc CFD; the first identifies an H1 chart and the fourth an H4 chart. These are illustrations from an equity-linked CFD feed, not a study of executed stock trades. Individual equities and exchange-traded funds are proposed subjects for subsequent testing. Results cannot be transferred between those instruments without checking their data and execution conditions. The central research question is whether combining the 14, 28, 50, 144, and 200 cloud structures with WFD Combo, WFD Breakout, and WFD ZigZag adds useful information beyond simpler moving average methods. The four examples motivate the investigation; they do not answer it.

Conceptual origins and modified MMD Clouds

The research is partly inspired by the MFD Clouds concept developed by Polish trader Mariusz Maciej Drozdowski. This acknowledgment identifies an influence on the approach; the framework discussed here should not be read as a reproduction or validation of the original MFD methodology. Analysing the order, slope and separation of moving averages is an established part of technical analysis. Moving averages smooth historical prices and therefore respond with a delay. Their trend-following and interpretive role is described in TradingView's . Presenting those relationships as clouds does not make the underlying principle a WFDQuant invention. The separate WFDQuant research proposal concerns how those structures interact with proprietary directional signals, support and resistance references, and swing-based management. Whether that integration improves decisions remains an open empirical question. Multiple tools derived from the same price history may provide overlapping information rather than independent confirmation.

The cloud framework

The source observations use five cloud periods: 14, 28, 50, 144, and 200. The shorter structures describe relatively recent price behaviour; the longer structures provide a broader trend reference. These are periods on the selected chart timeframe, not necessarily trading days. In a bullish configuration, shorter moving averages may sit above longer ones; in bearish conditions, the order can reverse. Slope and separation provide additional descriptive context. Frequent crossings and overlapping clouds suggest a less orderly configuration. For this study, cloud alignment is a candidate market filter, not an entry instruction. This period set is not shown to be superior to alternatives. A reproducible test must specify the average type, price input, exact cloud boundaries, indicator settings, and definition of alignment. Period labels and screenshots alone do not define a complete algorithm. A useful distinction is between a visual description and a measurable rule. “The clouds look orderly” can motivate research. A test needs an explicit condition, such as a fixed ordering of specified boundaries and a predefined slope measure, evaluated only from information available at the decision time. Any thresholds must be chosen before the final evaluation.

What the WFDQuant tools contribute to the hypothesis

WFD Combo supplies the directional and buffer-zone signals discussed in the source notes. The proposed investigation focuses on whether a renewed signal after a correction carries information beyond an isolated initial signal. Signal colors and annotations in the screenshots are descriptive; the material presented here does not establish their relative strength. WFD Breakout supplies support and resistance references. These provide context for a possible pause, break, or corrective return. A level is a location to investigate, not a promise of a reaction. Tests must retain the time at which each level became available, including any higher-timeframe confirmation. WFD ZigZag supplies swing references that the notes propose using as potential areas for reassessing an open position. The relevant research issue is when a turning point becomes confirmed. A completed historical chart can show structure that was unavailable at the apparent turning point. These roles describe the proposed workflow. They do not establish the tools' accuracy or their incremental value. Record their exact versions, settings, and signal timing before testing.

Figure 1 Support and resistance and a return towards cloud 28

The first example illustrates a declining price structure with broadly descending clouds. The original annotations connect reactions near horizontal levels with corrective moves towards the shorter cloud region, particularly the blue 28-period cloud. Figure 1. Support and resistance interactions and corrective returns towards the 28-period cloud in the source's bearish example. Original chart and annotations extracted from page 1 of Akcje chmury I SR; no price data or markings have been redrawn. The source proposes that price reaching a historical support area may pause, move through it, or recover towards a cloud before a decline resumes. Figure 1 illustrates that interpretation retrospectively. It does not establish how often the sequence occurs, how precisely the level can be identified in advance, or how many similar setups fail. The testable question is whether a subsequent WFD Combo signal, following a predefined level interaction and corrective return, improves outcomes relative to signals without that sequence. “Interaction,” “return,” and the maximum waiting time must be defined consistently. Counting only examples that eventually continue would introduce selection bias.

Figure 2 Tangled clouds and the candidate no trade filter

The second example provides a necessary counterpoint. Price and shorter clouds cross repeatedly, and the configuration is less orderly than in the directional examples. Figure 2. Overlapping and repeatedly crossing clouds illustrate the source's proposed conditions for standing aside. Extracted from page 1 of the original PDF with the existing signals and annotations preserved. The notes suggest avoiding this type of configuration. That is a proposed filter, not a demonstrated improvement. Excluding tangled clouds might reduce unproductive signals, but it could also delay participation in emerging trends or remove profitable trades. Evaluation should therefore include every qualifying signal before and after the filter, together with the opportunities it excludes. Lower trade frequency, higher average trade return, and higher total return are different outcomes. A useful filter must be assessed against the objective it is intended to improve.

Figure 3 Corrective entries and renewed directional signals

The third example develops a sequence rather than a single arrow: an established bearish cloud arrangement, interaction with an H4 support reference, a corrective recovery, and further directional signals. The chart retains the visible H4 level labeled 129.270. Figure 3. Successive signals within a bearish cloud structure, including corrective returns towards the faster clouds. The circled markers and H4 reference belong to the source image on page 1; the sequence is an illustration, not a verified trade record. The source interpretation favours a later signal after the corrective return over an earlier opportunity. This is a reasonable comparison to test, but “later” does not automatically mean better. Waiting may offer a closer structural invalidation point in some cases while producing a worse entry or no entry in others. The research hypothesis is that cloud alignment, a defined support-or-resistance interaction, a corrective move, and renewed WFD Combo confirmation form a more useful entry framework than directional signals alone. A fair comparison must use consistent entry timing, risk allocation, and exit rules, including cases when the later signal never arrives. The annotations do not provide enough information to calculate a verified trade return. No profit figure is inferred from the screenshot or the spoken estimate in the notes.

Investigating cloud based position management

The notes also raise a management question: should a position close when price crosses the fastest cloud, or should it be allowed to take a deeper correction? The proposed alternatives include tolerating interactions with cloud 14, allowing a return towards cloud 28, and treating contact with cloud 50 as a possible exit condition. These are experimental choices. A cloud touch, an intrabar cross and a close beyond a cloud are different events and must not be used interchangeably. A comparison should hold the entries constant and vary the management rule. It should measure realised returns, adverse excursions, drawdowns, holding time and the extent to which open profits are given back. Wider tolerance may preserve participation in a trend while increasing losses during reversals. Cloud management is not a complete risk policy. Initial invalidation, position sizing, maximum exposure and treatment of gaps still require explicit rules. Additional entries suggested by repeated signals must be tested separately from a single-position approach because they change aggregate risk.

Figure 4 Swing structure as a reference for reassessment

The fourth chart broadens the discussion to potential exit areas. Its circles mark swing locations identified in the source discussion using WFD ZigZag, while horizontal lines show structural price references. Figure 4. Circled swing locations and horizontal levels provide candidate areas for reassessing an open position. Original H4 chart extracted from page 2 of the PDF. The image does not establish when each swing marker became confirmed. A previously confirmed high or low can be treated as a candidate management reference rather than a prediction that price must reach or reverse at that level. The objective proposed in the notes is to investigate participation in part of a directional move, without assuming you can capture its absolute endpoints. Timing is essential. Generic ZigZag implementations can update their latest structure as new prices arrive, as explained in TradingView's . That documentation does not describe or verify WFD ZigZag. You must independently verify the confirmation and revision behavior of the specific WFDQuant implementation. A valid test may use a swing level only after it becomes available. It must not assign advance knowledge to a pivot merely because that pivot is visible on the completed chart. Compare structural exits with fixed-risk targets and cloud-based exits under the same entry and execution assumptions.

Why extend the study to equities and ETFs

The proposal is relevant to traders exploring markets beyond currencies, but familiarity with a charting method is not evidence of transferability. The supplied CFD examples provide a starting point for questions about equity-linked price behaviour; they do not establish results for exchange-traded shares or funds. Separate studies should distinguish cash equities, ETFs and CFDs. Each needs appropriate prices, session definitions, costs and execution assumptions. Overnight gaps, liquidity, corporate actions and event exposure may affect the observed patterns and their tradability. Short strategies also need realistic assumptions about availability and costs, while CFD studies need the relevant financing terms. For stock and ETF samples, adjusted histories and any treatment of distributions must be documented. Define the instrument universe without selecting only today's survivors. Closely related ETFs and stocks also require care: many similar-looking observations may reflect the same underlying market move.

A research programme that can reject the hypothesis

The next stage should be able to determine that the proposed integration adds no useful information. A practical program has six parts. 1. Freeze the specification. Record indicator versions and parameters, cloud construction, timeframes, level availability, swing confirmation, entry and exit events, sizing and exposure limits. Distinguish signal time from execution time. 2. Build explicit comparisons. Evaluate a simple moving average baseline, WFD Combo alone, cloud alignment with Combo, and the additional level and swing conditions. These comparisons test each component's contribution rather than attributing all results to the full combination. 3. Separate the questions. Test entry selection, cloud-based exits, swing-based targets, and additional entries independently before combining them. Otherwise, you can't attribute an apparent improvement to a particular decision. 4. Use realistic execution. Include spreads, commissions, slippage, financing or borrow costs where applicable, and session gaps. Use conservative assumptions when a single bar contains both a stop and a target. A signal confirmed at a bar close must not receive an earlier fill. 5. Reserve unseen data. Develop rules on a training sample, use chronological validation, and preserve a final evaluation period. Include different volatility and trend conditions. Report sensitivity to nearby periods and thresholds, and account for the number of alternatives tried. 6. Report the whole outcome. Publish trade counts, net expectancy, exposure, turnover, drawdowns, holding periods, and uncertainty, alongside rejected and missed signals. Compare risk consistently. An attractive average from a small selected sample is insufficient. The key questions are whether aligned clouds change the distribution of Combo outcomes, whether corrective entries add value after costs, whether support and resistance add information beyond trend context, and whether the proposed management rules improve the chosen return and risk objective. Evidence should also show where the framework fails.

Preliminary conclusion

The four charts support a coherent research agenda: describe trend context with clouds, investigate directional signals within that context, and examine structural references for managing positions. They remain selected visual examples, with no systematic sample, verified execution record, or statistical performance assessment. Moving average ordering, support and resistance, corrections and swing analysis are established concepts. The WFDQuant contribution proposed here is integrating them with proprietary tools and explicitly testing that integration. Neither the period set nor the added indicators should receive credit for an advantage before suitable comparisons demonstrate one. The appropriate next step is to determine whether, and under what conditions, the combination improves outcomes over simpler alternatives. A negative result would be a useful research finding too.

Acknowledgement and source note

This research was partly inspired by the MMD Clouds concept developed by Mariusz Maciej Drozdowski. The WFDQuant observations, proposed integration, and hypotheses form a separate exploratory study. The four figures are original embedded chart images extracted from Shares cloud & SR, pages 1 and 2. We preserved their prices, signals, circles, and lines. The first three are small source screenshots; some labels are correspondingly difficult to read. We did not increase their resolution or reconstruct missing detail. English captions explain their role without altering the source images. This publication presents preliminary observations and research hypotheses for analytical and educational purposes. It does not constitute investment advice or a validated trading strategy. No verified performance advantage is claimed.