Steps

1. Fill, sign and send us the Client Registration Form with your personal data, choose your preferred Investment Plan and the amount you want to invest. Pay attention to enter the bank account data in your own name in the Client Registration Form, because this is the basic condition for concluding the Investment Agreement and we will transfer the interest and the principal at maturity to this account number.
2. Based on the Client Registration Form, we will send the Investment Agreement. Send a copy of the signed Investment Agreement back to us.
3. Deposit the Investment Amount to invest via a bank transfer to the bank account number specified in the Investment Contract. Make sure that the bank transfer is made from the account owned by the client, which was previously indicated on the Customer Registration form.
4. Once the Investment Amount arrived at our company’s account we send you a Confirmation of Your Deposit. The Investment Agreement is valid from the day the Investment Amount arrives in our company’s bank account.
5. The interest payment is due after one year or on a monthly basis – depending on your Investment Plan. The Company is unable to initiate a transfer to a third party. The interest payments and refunds due at the end of the contract will only be returned to the bank account of the person with whom we concluded the contract and from which account number the amount was received.
6. When your investment has matured, you have 3 options:
   • Withdraw the principal and the interest.
   • Reinvest only the principal and withdraw the interest.
   • Reinvest the principal and the annual interest.
7. In the case of reinvestment, you do not need to sign a new contract, you just have to declare that you are investing only the principal or the principal and the interest payment for the next year.

Secrets of Trading – Results of Our Fullscope Reserch

Key Characteristics of Tradable Instruments

A trading strategy cannot be evaluated without considering the characteristics of the instrument on which it operates. Transaction costs, financing costs and market liquidity can substantially change the result of a strategy that appears profitable in a simplified calculation.

Spread

The spread is the difference between the price at which an instrument can be bought and the price at which it can be sold. It represents an immediate trading cost because every position starts with a small loss equal to the spread.

Its effect becomes particularly important in strategies that trade frequently or aim for relatively small profits. A strategy can appear profitable before costs while producing a loss once the spread on every transaction is included.

Spreads are generally narrower in liquid markets and wider when trading activity is low, volatility increases or the market is outside its most active trading hours.

Overnight Financing Cost

Leveraged positions that remain open overnight may incur a financing charge, commonly referred to as a swap. Depending on the instrument and the direction of the position, the trader may pay or occasionally receive this amount.

The effect of financing costs increases with the holding period. A strategy that keeps positions open for several days, weeks or months may lose a significant part of its apparent return through accumulated overnight charges.

Market Capitalisation

For shares and equity indices, market capitalisation indicates the size of the underlying company or market. Larger markets usually attract more participants and support deeper liquidity, although market capitalisation and liquidity are not the same measure.

Smaller markets can experience sharper price movements, wider spreads and greater price impact when larger orders enter the market. Market capitalisation therefore helps assess whether an instrument is suitable for the intended position size and trading approach.

Market capitalisation does not apply in the same way to currencies and commodities, so these instruments require different liquidity measures.

Trading Volume

Trading volume shows how actively an instrument is traded during a given period. Higher activity usually allows orders to be executed more easily, with tighter spreads and less slippage.

Low trading volume can make entry and exit more expensive and less predictable. The price received may differ from the expected price, especially when the position is large relative to the available market liquidity.

Why These Characteristics Matter

The same technical signal can produce very different results on two instruments. Spread, financing costs and liquidity determine whether a theoretical strategy can be traded profitably under real market conditions.

For this reason, our simulations incorporated the relevant characteristics and costs of each instrument. Ignoring them would make the test results appear better than a trader could realistically achieve.

Tens of Millions of Indicator Simulations

We did not evaluate technical indicators using a few selected settings or favourable historical examples. We tested every widely used indicator across all meaningful parameter values and practical combinations.

Each indicator was examined across:

• 53 financial instruments;
• tick data and six candle periods;
• multiple price calculation methods;
• standard and Heikin-Ashi candles;
• different level calculation methods;
• multiple position opening structures;
• different Take Profit and Stop Loss levels;
• more than 20 years of market data.

Combining these variables resulted in millions of individual simulation runs.

A strategy was considered credible only if its performance remained consistent across different instruments and market conditions, continued after the optimisation period and survived real trading costs.

Results – When Optimised Performance Meets New Data

7 Instruments

To examine whether historical optimisation could produce lasting results, we selected AMA indicator settings for seven different financial instruments:

• JP225;
• US500;
• Brent crude oil;
• gold;
• USDJPY;
• FTSE 100;
• EURUSD.

For each instrument, the indicator parameters and trading rules were optimised using the learning period. The selected settings were then frozen and carried forward unchanged into the test period from January 2024 to August 2026.

This separation is essential. Strong performance during optimisation only demonstrates how well a strategy fits data that has already been observed. Its practical value depends on whether it can retain that performance when applied to subsequent data that played no role in selecting its parameters.

The first chart shows the cumulative returns generated by the seven selected strategies. The vertical dashed line marks the beginning of the test period in January 2024.

Before this boundary, the optimised strategies appear highly successful. By the beginning of 2024, their accumulated returns range from approximately 14% to more than 140%. Several curves show long and apparently convincing upward trends.

The picture changes after the selected settings encounter new market data.

During the test period:

• the JP225 strategy falls substantially from its previous peak;
• the US500 strategy gives back a considerable part of its accumulated return;
• the Brent strategy also experiences a pronounced decline;
• the FTSE 100 strategy loses much of its earlier gain;
• the USDJPY and EURUSD strategies deteriorate;
• and the gold strategy finishes broadly around its January 2024 level.

None of the seven strategies reproduces the strength of its learning-period performance. The important observation is therefore not that every strategy immediately collapses, but that the apparent advantage identified during optimisation largely disappears when the same rules are applied to new data.

Historical Success Was Parameter-Dependent

The selected AMA settings were not chosen randomly. They were selected precisely because they had performed well during the learning period.

This makes their subsequent deterioration particularly important. The profitable historical curves did not represent a stable relationship that continued automatically into the future. They represented parameter combinations that were well suited to the market behaviour already contained in the optimisation data.

When the market environment changed, the parameters remained the same.

The indicator continued to calculate exactly as designed, but the relationship between its signals and subsequent price movements was no longer equally favourable. Signals arrived at less advantageous moments, profitable movements were captured less effectively, and adverse trades eliminated part of the previously accumulated return.

Percentage returns can appear abstract, so the second chart translates the test-period results into account values.

Each of the seven strategies begins January 2024 with a separate hypothetical balance of $100,000. The entire account is allocated according to the relevant strategy, and each realised result is fully reinvested in the following trades.

By the end of the test period, the approximate account values are:

• Gold: $99,469.26
• EURUSD: $93,700.15
• USDJPY: $92,332.29
• FTSE 100: $85,631.60
• Brent: $79,392.14
• US500: $78,331.60
• JP225: $73,883.05

Only the gold strategy finishes close to its initial capital, the other six strategies finish below $100,000, with the weakest account losing approximately one quarter of its starting value.

The chart also reveals an important feature that the final values alone would conceal: the path was unstable. Some accounts initially rose above $100,000 and appeared successful before suffering major reversals. Evaluating them after only a few favourable months could therefore have produced a very different—and misleading—conclusion.

The central problem was not the calculation of the AMA indicator. It was the assumption that parameters selected from historical success would retain the same relationship with future price movements.

Historical optimisation can produce remarkably convincing charts. The real test begins only after the optimisation ends.

Gold Multi-MA Strategies: Optimisation on a Single Market

The second experiment focused exclusively on gold, traded as XAU/USD. Instead of comparing different instruments, we examined numerous variations of the same underlying technical concept.

Each strategy used three moving averages. We varied:

• the type of each moving average;
• the lag applied to each moving average;
• the combination of the three indicators;
• and the Take Profit distance.

The tested moving-average types included:

• Simple Moving Average (SMA);
• Exponential Moving Average (EMA);
• Smoothed Moving Average (SMMA).

The strategy interpreted the relative position of the three moving averages as a directional signal. Their ordering determined whether the system opened a long or short position, maintained the existing exposure, closed a position or changed direction.

The optimisation also tested three Take Profit levels:

• 0.5%;
• 1.0%;
• 1.5%;

This produced a large number of variations from a seemingly simple trading principle.

Why Three Moving Averages?

A moving average reduces short-term price noise and attempts to reveal the underlying direction of the market. Different calculation methods and lag settings respond to price changes at different speeds.

A more responsive moving average reacts quickly but may generate frequent false signals. A slower average produces smoother signals but may respond only after a significant part of the price movement has already occurred.

Using three moving averages creates a relative ordering. When the faster averages move above the slower ones, the strategy may interpret the configuration as an upward trend. When their order reverses, it may indicate a downward trend or an exit condition.

This logic appears intuitive, but its result depends heavily on the precise combination of average types, lag settings and exit levels.

The first chart presents the cumulative returns of the selected Multi-MA configurations. The vertical dashed line at the beginning of 2017 separates the learning period from the subsequent test period.

During the learning period, many strategies produced remarkably strong results. Some accumulated returns of more than 200%, while numerous other configurations reached approximately 100–180%.

If the analysis had ended at the optimisation boundary, several of these strategies would have appeared highly successful.

After 2017, however, their behaviour changed considerably.

Some of the historically strongest configurations retained part of their accumulated gains, but their curves stopped rising consistently. Others entered extended periods of stagnation or gave back a substantial proportion of their earlier profits. Several of the weaker configurations eventually fell close to or below zero cumulative return.

The test results therefore became highly dispersed:

• a small group preserved or moderately increased its historical gains;
• many strategies produced little meaningful progress;
• several suffered prolonged declines;
• and some finished the complete period with negative cumulative results.

The important comparison is not the final height of each line alone. Since the chart includes gains accumulated before 2017, a strategy may still display a large positive cumulative return while having performed poorly throughout the test period.

A line that rises to 200% during optimisation and later falls to 120% does not represent a successful test merely because it remains above zero. It represents a strategy that lost a substantial part of its previously accumulated advantage after its parameters were frozen.

Similar Rules, Radically Different Outcomes

All the strategies were based on the same general idea. They traded the same instrument, used three moving averages and interpreted their relative positions.

Nevertheless, relatively small differences in moving-average type, lag or Take Profit level produced significantly different long-term results.

One parameter combination happened to capture certain gold trends effectively, while another entered later, exited earlier or reacted differently during sideways periods. These small timing differences accumulated over hundreds of trades.

The dispersion between the curves demonstrates how sensitive an optimised strategy can be to its exact parameters. If the underlying trading principle were consistently robust, neighbouring configurations would be expected to produce broadly similar results. Instead, the outcomes range from substantial historical gains to complete loss of those gains.

The second chart translates the test-period performance into account values.

Each displayed strategy begins in January 2017 with a separate hypothetical account balance of $100,000. The results are fully reinvested, so every gain or loss affects the amount of capital available for subsequent trades.

During the early years, several accounts initially appear successful. Some rise above $120,000, and the strongest temporary peaks approach $140,000.

These gains do not remain stable.

From approximately 2020 onward, many accounts begin a prolonged decline. By 2023–2024, a large proportion of the strategies are below their original $100,000 balance. Gold’s strong movements during 2025 temporarily improve several results, but many of these recoveries are followed by renewed losses.

By August 2026:

• only a minority of the tested configurations remain above $100,000;
• many finish between approximately $60,000 and $95,000;
• the weakest accounts retain less than $60,000;
• and even the strongest final results are modest compared with the risk and fluctuations experienced along the way.

The spread between the final account values is approximately $60,000. This difference was created not by trading different markets, but by changing parameters within the same moving-average framework.

Temporary Profit Is Not Evidence of Robustness

The account-value chart also shows why evaluating a strategy at one arbitrarily selected date can be misleading.

A configuration examined near one of its peaks could appear highly profitable. The same configuration evaluated several months later might show a significant loss. Selecting the most favourable endpoint would therefore create an exaggerated impression of reliability.

Several strategies rose well above their initial capital before falling substantially below it. Their temporary gains did not protect them from later drawdowns because the relationship between the moving-average signals and the behaviour of gold was not stable.

Full reinvestment makes this deterioration especially visible. Once an account loses capital, subsequent returns are earned on a smaller base, making recovery increasingly difficult.

What the Gold Experiment Demonstrates

This experiment does not show that moving averages are incapable of describing past price behaviour. They clearly can, and optimisation can identify combinations that fit historical gold movements extremely well.

It demonstrates that this historical fit is not necessarily a persistent trading advantage.

The three moving averages continued to calculate correctly throughout the test period. The Take Profit rules also operated exactly as defined. What changed was the market’s response to the signals.

The same relative indicator positions no longer led to the same distribution of subsequent price movements. Parameters that appeared optimal before 2017 became less effective, while small differences between configurations produced increasingly divergent outcomes.

The Central Lesson

The seven-instrument experiment showed that optimised AMA settings did not transfer reliably across time and markets. The gold experiment reaches a similar conclusion within a single instrument and a single strategy family.

A technically convincing rule can generate an excellent learning-period result. Testing many indicator types, lags and Take Profit levels makes it even easier to discover an impressive historical configuration.

However, the more alternatives are tested, the greater the probability of selecting a combination that succeeded because it happened to fit the past.

Optimisation can identify the historical winner. It cannot guarantee the future winner.

Back to top

What Did 20 Years of Research Teach Us?

Twenty years of research required substantial time, effort and capital. Although the large-scale simulations were performed by computers, the development, testing and learning process was also supported by experience gained through live trading accounts rather than relying exclusively on hypothetical demo results.

The most important lesson was not the discovery of a perfect indicator or automated strategy. It was understanding why apparently successful solutions repeatedly fail when market conditions change.

Optimisation Explains the Past

Technical indicators can be optimised to produce impressive results over a selected historical period. A strategy calibrated to the previous two years may fit that particular dataset exceptionally well.

However, optimisation does not make the future resemble the past. Economic conditions, monetary policy, investor preferences, geopolitical events, liquidity and market sentiment continuously change.

Parameters that captured profitable movements in one period may:

• close the same movement too early in the next;
• enter after the opportunity has already passed;
• generate false signals in a different market regime;
• or produce losses instead of the previously observed profits.

This problem appeared regardless of the instrument, geographical market or indicator family being tested. Our research did not identify a fixed indicator-based strategy whose historical performance remained sufficiently reliable for us to entrust capital to it indefinitely.

Small Parameter Changes Can Produce Extreme Differences

During the simulations, we repeatedly observed that a very small adjustment could create an enormous difference in the final result.

Changing a Take Profit level by half a percentage point, altering an indicator period by one unit or modifying a minor entry condition could turn an apparently outstanding strategy into an unsuccessful one.

The explanation was often a small number of decisive trades. One setting happened to close a position near the top of a major movement, while another closed slightly earlier, remained exposed during the reversal or missed the movement completely.

The profitable setting did not necessarily demonstrate superior intelligence. It may simply have matched the historical path more favourably.

In another period, a different parameter would have appeared to be the “perfect” choice. This sensitivity is a warning sign of overfitting rather than evidence of a robust trading advantage.

Automation Does Not Create Adaptability

The promise of automated trading is attractive: a robot continuously monitors the market, follows predefined rules and generates income without requiring the trader’s attention.

Automation can execute rules quickly and consistently, but it cannot make rigid rules appropriate for every environment. A system optimised for yesterday’s market can continue applying the same logic even after the conditions that made it successful have disappeared.

A trading robot does not eliminate uncertainty. It merely automates the consequences of its assumptions—including the incorrect ones.

For that reason, we do not consider unsupervised automation a substitute for judgement and active risk management.

Automation Is Valuable as Decision Support

Although we would not hand complete control of the portfolio to an automated system, automation remains useful.

A decision-support system can:

• monitor open positions continuously;
• calculate the portfolio’s volume-weighted average opening price;
• measure long, short and net exposure;
• track current profits, losses and available capital;
• monitor selected indicators and price levels;
• detect when positions offset one another;
• and send an alert when a predefined condition is reached.

For example, the system can notify the trader when a position falls by 1%, rises by 0.5% or when the combined result of several positions approaches break-even.

The computer identifies the event and presents the relevant information. The trader evaluates the current situation and makes the final decision.

Artificial Intelligence Has the Same Historical Limitation

Artificial intelligence and machine-learning systems can process far more historical data than a human and identify relationships that would otherwise remain unnoticed. They are valuable analytical and decision-support tools.

Their predictions, however, are still derived primarily from patterns contained in the data available to them. When market behaviour changes or an unprecedented event occurs, the historical relationships may no longer provide a reliable guide.

A model can process thirty years of data in a short time, but this does not mean that it has experienced the next market regime. Learning a genuinely new pattern requires relevant new observations. Rare events may provide too few examples, and by the time sufficient evidence has accumulated, the original opportunity may already have disappeared.

Human judgement is not infallible either. Its potential advantage is the ability to interpret context, construct explanations and reconsider assumptions after limited evidence. AI can assist this process, but it should not be treated as an oracle that removes uncertainty from the market.

Protecting Capital Comes First

The final lesson returns to the mathematics of recovery. A 30% loss requires a 42.86% gain merely to restore the original capital. A 50% loss requires a 100% gain.

Avoiding severe drawdowns is therefore more important than producing an impressive backtest or maximising the profit of an individual trade.

Our Central Conclusion

After twenty years of research, we reached five fundamental conclusions:

• Indicators fail in the long run and in sideways movement
• Lack of consistency → technical analysis is unreliable
• Trading cannot be automated
• Useful only as decision-support tools
• AI also relies on historical data, struggles with unprecedented events, learns slowly
• A wrong signal always causes greater losses than the gains from correct ones.

Back to top

The Verdict – What Works?

After twenty years of research, millions of simulations and experience gained through live trading, our conclusion is clear:

Market psychology combined with active position management proved more robust than any indicator-driven strategy tested in isolation.

Technical indicators can describe past and present price behaviour. They can identify trends, momentum, volatility and frequently observed price levels. What they cannot determine consistently is whether the relationship they identified will remain valid after the next market event.

The objective is therefore not to discover a perfect indicator. It is to understand the forces influencing market participants and manage the portfolio as those forces change.

Market Psychology and Position Management

Our approach combines two essential components:

• market psychology, which interprets the collective expectations, emotions and reactions behind price movements;
• position management, which controls exposure and responds when actual market behaviour differs from the original expectation.

Market psychology helps us assess why buyers or sellers currently dominate and how participants may interpret new information. Position management acknowledges that this assessment can still be wrong.

The two components must operate together to provide predictable and reliable returns.

The Proper Role of Technical Analysis

Technical analysis remains useful, but primarily as a supporting tool.

It can help identify:

• potential support and resistance areas;
• changes in momentum;
• the strength or weakness of an existing trend;
• unusual volatility;
• and price levels widely observed by other traders.

These observations can inform a decision, but they should not be mistaken for fixed laws governing the market.

Why the 20-Period Moving Average Sometimes Matters

The 20-period moving average is frequently monitored by traders. As the price approaches it, many market participants may react at approximately the same time.

Some may close profitable positions. Others may open new trades, move their protective orders or wait for confirmation of a breakout. This concentration of attention can temporarily turn the moving average into an area of support or resistance.

Its effect does not come from the number 20 possessing any inherent predictive power. Its practical relevance comes largely from its popularity.

The 20-period moving average is not a magical price level. It can matter because many traders believe that it matters.

This is partly a self-reinforcing psychological effect. If enough participants expect a reaction near the same level, their collective decisions can help create that reaction.

Why the Same Level Eventually Fails

The 20-period moving average does not always stop the price. Historical charts contain many examples in which it acts as support or resistance and many others in which the price crosses it without hesitation.

When a sufficiently strong psychological wave develops, the orders associated with the moving average are overwhelmed. This may be caused by unexpected news, a major change in expectations, fear, enthusiasm or a rapid reassessment of risk.

The technical level has not malfunctioned. The balance between buyers and sellers has simply become strong enough to break through it.

Technical analysis can show that the price is approaching a widely observed level. It cannot reliably determine in advance whether the next psychological wave will stop there or break through it.

That question requires an interpretation of the broader situation—and even then, the outcome remains uncertain.

Chart Patterns Face the Same Problem

The same principle applies to chart formations.

A W-shaped formation, flag or other recognisable pattern can be followed by the movement predicted by technical analysis. A visually similar formation can also fail, reverse or develop into something entirely different.

It is easy to collect successful historical examples and present them as evidence. However, an objective analysis must also include every comparable formation that produced a false signal.

The pattern itself is therefore insufficient. Its meaning depends on the surrounding market psychology, expectations, liquidity and events. Two formations may look almost identical while representing completely different situations.

Without this context, technical signals can appear random and unpredictable.

Why Psychology Outperforms Fixed Indicator Rules

Indicators calculate the same way regardless of what is happening in the world. This consistency is computationally useful but strategically restrictive.

A fixed rule cannot understand why the price is moving. It cannot distinguish automatically between:

• an ordinary short-term fluctuation;
• a reaction to unexpected information;
• a temporary emotional overreaction;
• and a lasting change in expectations.

Market psychology does not attempt to treat every technically similar situation as identical. It considers how participants are likely to interpret the current environment and observes how that interpretation is expressed through actual price behaviour.

This makes the process adaptable. The method itself does not depend on predicting which world event will happen next. It provides a framework for evaluating and responding to events after their significance begins to emerge.

Why Position Management Is Indispensable

No psychological or technical analysis can guarantee that the market will move in the expected direction. Position management is what prevents an incorrect expectation from automatically becoming an unacceptable portfolio loss.

It determines:

• how much capital is placed at risk;
• how exposure is built or reduced;
• whether additional positions should be opened;
• when profitable positions should be realised;
• how opposing positions can modify net exposure;
• and when the original interpretation must be reconsidered.

A strategy based entirely on entry signals assumes that successful prediction is the primary source of performance. Our conclusion is different: long-term survival depends at least as much on managing incorrect predictions as on identifying correct ones.

Responsive, Not Independent

World events will always affect financial markets. No legitimate trading method can make returns completely independent of wars, crises, monetary decisions or changes in investor behaviour.

A robust process can, however, avoid depending on one predetermined reaction to those events.

Instead of assuming that a particular announcement must move the market upward or downward, the trader observes how participants actually interpret it and adjusts the portfolio accordingly.

The goal is therefore not independence from world events, but adaptability to their consequences.

Our Final Conclusion

Our research did not identify a technical indicator that could produce consistently reliable long-term results through fixed rules alone. Optimised indicators repeatedly lost effectiveness when the market environment changed.

The combination of market psychology and active position management was more resilient because it did not require every future market situation to resemble the past.

Technical analysis still has a role. Indicators can organise information, highlight relevant levels and support the decision-making process. They should not be entrusted with the decision itself.

The resulting framework is:

1. Assess the prevailing market psychology;
2. Use technical and fundamental information as supporting evidence;
3. Open positions with controlled exposure;
4. Observe the market’s actual reaction;
5. Manage the entire portfolio as conditions change;
6. Protect capital when the original interpretation proves incorrect.

Back to top