?

Whitepaper

1. Our Guiding Principles

We focus on a clear goal - to earn money for clients, without taking up their time.

Our primary objective is clear-cut: to generate returns for our clients without consuming their valuable time. Our existing and forthcoming offerings are grounded in three fundamental principles:

1. Efficiency

In order to deliver substantial returns, we ensure our cryptocurrency trading automation technologies remain aligned with market dynamics. We consistently test and incorporate successful trading strategies into our artificial intelligence model. All our strategies are firmly rooted in empirical and academic research.

2. Transparency

Clients have unfettered access to information regarding the management of their funds. Details concerning transactions, commissions, profits, and balances are readily accessible through your personal &Volume account as well as your exchange account. Our service pricing is transparent, contingent on deposit volume, and devoid of hidden charges. Any minor alterations to the terms of service are communicated well in advance.

&Volume is structured to minimize and resolve conflicts of interest that may arise among us, clients, and exchanges.

3. Security

Client funds are securely held in their exchange accounts and remain accessible at all times. We strictly adhere to the core principle of non-involvement in the withdrawal or transfer of these funds. This principle guides our development of new products for traditional stock and foreign exchange markets.

We neither gather nor retain personal information. Our utilization of analytics tools, such as Google Analytics, is strictly for anonymous user activity analysis. This data aids us in enhancing interfaces, ensuring a user-friendly experience on both the website and &Volume personal account.

2. Methodology

In this section, we delve into the fundamental concepts of artificial intelligence, the technological underpinnings of &Volume, and evaluate its effectiveness and capabilities.

The term artificial intelligence (AI) encompasses endeavors to devise algorithms capable of emulating human-like actions (Winston, 1984). AI can manifest as software operating on computers, servers, or integrated into chips within smartphones, automobiles, and household appliances. Consequently, any algorithm where one action hinges on another can be construed as an instantiation of artificial intelligence.

Researchers delineate three categories of AI: weak (Narrow AI), strong (General AI), and super-intelligence (Super AI) (Patel, et al., 2020). Presently, all existing forms of AI, including voice assistants and text and graphical generative models such as GPT-4 and Midjourney, fall under the rubric of weak AI.

The distinction between weak and strong AI lies in the latter's self-awareness and approximation to human intelligence (Turing, 1950).

Nevertheless, the adoption of AI through deep machine learning is considered by some scholars as a pivotal step towards achieving strong AI (Daudov et al., 2020).

Neural networks, a subset of artificial intelligence, constitute a machine learning technique that mirrors the architecture of the human brain for processing intricate data. Comprising numerous interconnected neurons, they engage in the processing and transmission of information amongst themselves. Each neuron receives input, processes it, and transmits the outcome to the subsequent neuron (Picton and Picton, 1994).

Several categories of neural networks exist, including forward, recurrent, and convolutional (Tkáč and Verner, 2016). Each possesses distinct characteristics and finds application contingent on the specific task at hand.

Neural networks find utility in automating speech and image recognition, natural language processing, and predictive modeling, among other domains. They can be trained on extensive datasets for tasks such as classification, clustering, and prediction. In decision-making contexts, including financial markets, machine learning algorithms draw upon data derived from neural networks, which can learn autonomously or under supervision.

2.1 Foundation of the AI Model

In essence, at &Volume, we employ a blend of tools, comprising a data collection system, an array of neural networks for data processing, and an algorithmic decision-making system grounded in machine learning.

The data collection system channels information from diverse sources to the neural networks for processing. Subsequently, the processed data is relayed to the decision-making system for interpretation. Following this, the system determines whether to initiate a transaction, selects the position's direction, and identifies the optimal selling moment.

Figure 1 illustrates the overarching model utilized by &Volume to prognosticate trade profitability. The initial step involves aggregating all requisite data to train and assess the predictive model. This data may be subjected to processing, transformation, or purification to eliminate extraneous noise, retaining only pertinent information. The processed data is then employed for model training. At the validation phase, hyperparameters undergo optimization. Ultimately, the tuned model proceeds through a testing phase to appraise the outcomes.

Figure 1. General AI model &Volume (2023).

For trading determinations, &Volume relies on machine learning algorithms overseen by a mentor, alongside algorithms featuring a classical reinforcement learning loop (Figure 2). In mentor-guided algorithms, the role of the mentor is assumed by ML programmers and professionals with expertise in both stock and crypto market trading, who assess the system's decision-making and rectify as needed.

Figure 2. Classical reinforcement learning loop model (2023).

Thus, our AI model assumes a hybrid nature, amalgamating distinct components into a unified system. As posited by Nazari et al. (2022), hybrid AI models, formed through the integration of various machine learning techniques, offer advantages over homogeneous counterparts. Each individual model within the hybrid framework bolsters the overall system's performance.

This hybrid AI model adeptly foresees daily price trends and executes profitable transactions within a 24-hour timeframe (Patel et al., 2020). Nevertheless, conducting transactions over a lengthier trading day horizon translates to heightened profitability rates. Hence, our primary strategy currently centers on managing finances in day trading, rather than the medium to long term.

Over extended periods, such as seven days, the model's efficacy diminishes noticeably, alongside its profitability ratio (Patel et al., 2020).

A pivotal benefit of an AI model lies in its absence of emotions. Even accomplished traders grapple with substantial losses attributable to impulsive actions driven by unbridled emotions (Astor et al., 2015). Emotions can precipitate impromptu decisions. Thoughtful planning and disciplined strategy can serve as safeguards against this, and AI mitigates the risk of emotional decision-making, thereby averting major losses.

Detailed comparative test data across various timeframes in relation to alternative investment instruments are expounded upon in subsequent sections.

2.2 Basic Trading Techniques.

Earlier, we outlined the fundamental operational framework of our AI model. Now, let's delve into the specific methodologies employed at &Volume, particularly in the context of day trading.

Basic Trading Techniques

Earlier, we outlined the fundamental operational framework of our AI model. Now, let's delve into the specific methodologies employed at &Volume, particularly in the context of day trading.

Trading Range

Trading range, also known as mean reversion, operates on the premise that prices have a tendency to revert to their mean (Wang et al, 2016). Our approach draws from the strategy expounded in the academic work of Leoung & Lee (2015). We pinpoint instances where an asset's price deviates significantly from its historical average, often due to overbuying or overselling. We then initiate trades with the expectation that, within a specified timeframe, the price will return to its average from several cycles prior.

In deploying a mean reversion strategy, we employ classical technical indicators such as moving averages and Bollinger Bands (Bollinger, 1992), alongside other intricate factors. Moving averages gauge an asset's historical price average, while Bollinger Bands, leveraging volatility metrics like standard deviation, assist in discerning significant deviations from the average market value (Bollinger, 2002). To effectively implement this strategy, we found it imperative to adapt and augment the AI model with more sophisticated conditions, as an excessive reliance on technical indicators would result in diminished profitability when trading crypto pairs.

The mean reversion strategy can be applied across various timeframes, encompassing longer-term ones (Bollinger, 2002). Nonetheless, owing to specific factors and constraints inherent to crypto exchanges, we exclusively employ this strategy for short-term intraday trading. At &Volume, this strategy is grounded in statistical analysis and the identification of abnormal price movements. Upon detecting a price deviation, the AI system generates signals to initiate trades in the opposite direction of the deviation (Figure 3).

Figure 3. An illustration of initiating and concluding a sell transaction utilizing the mean reversion strategy within a 30-minute timeframe (2023).

In mean reversion trades, our AI evaluates the probability of a price reversal and establishes appropriate stop-loss levels to mitigate potential losses if mean reversion fails to materialize. Our AI model integrates statistical analysis, risk management methodologies, and ongoing scrutiny of market dynamics.

Volume Weighted Average Price (VWAP).

The volume-weighted average price strategy involves breaking down a substantial order and strategically placing smaller segments into the market based on historical volume profiles for specific positions. The objective is to execute the order in proximity to the Volume Weighted Average Price (VWAP) determined by a general (Figure 4) or specialized formula.

Figure 4. General formula for calculating prices based on weighted average volume (2023).

Time-Weighted Average Price (TWAP).

With the time-weighted average price strategy, we divide large trades into intervals and dispatch smaller, AI-determined order segments to the market at evenly spaced time intervals between the commencement and conclusion times. Our aim is to fulfill the order at the average price between these time boundaries, thereby minimizing the market impact of high-volume entry into a market with limited trading volume.

Percentage of Volume (POV)

Until the trade order is entirely filled, our AI system continually dispatches partial orders based on the defined participation rate and market trading volume. This volume percentage strategy, also known as a step strategy, involves our AI system placing orders at an automatically determined percentage of market volumes. When the price of the cryptocurrency reaches a predetermined level, the AI adjusts this participation level accordingly.

In addition to these methodologies, we incorporate several other algorithmic strategies with the overarching aim of discerning the actions of other market participants. These algorithms within our AI model enable us to discern the actions of systems akin to ours on either the buy or sell side of a large order. This discernment aids us in identifying opportunities involving large orders, enabling us to execute orders at a more advantageous price point. Such maneuvers are referred to as AI advance development (Hens et al., 2018). We strictly adhere to the rules and regulations set forth by the Financial Industry Regulatory Authority (FINRA), given that forward action practices may be subject to legal scrutiny depending on the circumstances and are heavily regulated (FINRA, 2013; FINMA, 2023; FCA, 2023, ESMA, 2023).

2.3 Risk management

In Relation to Trading

The &Volume application employs a combination of techniques because, in most cases, relying solely on a single approach may not yield a profitable trade. Depending on the market conditions, the AI swiftly adjusts its approach in milliseconds. Even when a strategy is implemented competently at the right time, it does not guarantee the absence of unprofitable transactions. To mitigate losses, we continuously evaluate various strategies using historical data.

Furthermore, our "Children" system is operational for the first time. It diligently tracks all transactions and maintains an event log. In the event of an emergency, the system promptly notifies the user. Simultaneously, our AI system support specialist receives a documented account of the situation, providing real-time responses and furnishing data for further analysis to facilitate system enhancements.

In Relation to Your Personal Account

We request users to provide API keys for access to their exchange accounts with restricted permissions, preventing conversion or withdrawal of funds. Consequently, if a user independently misplaces their key, and the risk system is not at fault, the ensuing consequences are minimized.

API keys are securely stored in an isolated vault, accessible exclusively to the &Volume security officer. Access protocols are subject to stringent alterations during each approved cycle, adhering to internal security protocols.

Despite API keys not being of significant interest to potential attackers, as they do not grant the ability to transfer funds, we conduct thorough audio penetration tests.

If needed, specific guidelines on the secure storage of API keys can be provided upon request through our feedback form.

We do not gather or retain personal user data.

3. Profitability

The rationale behind focusing the service on trading cryptocurrency pairs lies in our belief that this constitutes one of the most effective realms of financial management. When weighing risk against return, cryptocurrency trading, with certain caveats, appears more appealing compared to conventional investment vehicles such as bank deposits, savings accounts, trust management, and the like.

To gauge benchmark efficiency, let's delve into trading statistics using a month-long example (Figure 5).

Figure 5. Chart depicting trading activity and account balance over the course of a calendar month (2023).

The vertical axis represents the account balance, while the horizontal axis denotes the number of transactions. Commencing with a stable coin balance of USDT 100, equivalent to $100, our system executed a total of 681 trades within the month. Of these, 299 trades yielded losses, while 382 were profitable. The most adverse outcome amounted to -$6.91, while the most favorable outcome stood at +$3.3. Throughout the month, the balance exhibited uneven growth, punctuated by intermittent drawdowns. These drawdowns are a consequence of the specificities inherent to our trading strategy and the functioning of the AI model. The cumulative weight of profitable trades gradually elevated profitability, culminating in a month-end balance of $152.02, or a 52% increase. Over a span of up to 12 months, the average efficiency approximates to ~360% annually.

This metric was selected as a benchmark for comparative assessment against traditional investment instruments.

3.1 Comparison with UK Banks

In the UK, the average interest rate for deposits typically ranges from 1% to 5% annually. However, most options do not entail monthly interest accrual, meaning there is no compounding, and payments are only provided at the end of the term. The most favorable interest rates are usually offered on fixed deposits that do not allow for partial withdrawals or additional deposits. This type of investment does not necessitate prior experience, and all deposits are safeguarded by the government's FSCS scheme, providing coverage up to £85,000 (FSCS, 2023). Consequently, the risk of investment loss is very low, making it suitable for safeguarding significant savings over the long term.

Nonetheless, this form of investment is susceptible to a high risk of losses due to inflation. For instance, in 2022, depositors experienced substantial monetary losses, with UK inflation reaching 11% (NY Times, 2022). This figure is more than twice the highest interest rate offered on a deposit.

3.2 Comparison with EU Banks

In European countries (EU plus Switzerland), the average interest rate for deposits typically ranges from 0.1% to 4%. In most cases, interest is compounded monthly, and there's an option for capitalization. It's worth noting that within the EU, there isn't currently a unified deposit insurance system; each country has its own policy. For instance, in Germany, the state insures deposits up to 100,000 euros under the DSGV scheme (Europa, 2023).

Similar to other forms of investment, this type is also susceptible to a high risk of loss due to inflation. For example, in 2022, inflation in European countries was at 9% (Inflation Tool, 2023). In 2023, it's anticipated to be 5% (Inflation Tool, 2023).

3.3 Comparison with US Banks

In the present economic climate, the interest rates on deposits in US banks typically span from 0.1% to 5%. Notably, in 2023, most US banks are not incentivizing residents to deposit money, as they offer rates that are notably below the inflation rate. It's crucial to highlight that all deposits up to $250,000 in US banks are insured by the government agency FDIC (FDIC, 2023).

In 2022, inflation in the US surpassed 8%, resulting in a devaluation of deposits (Statista, 2023).

3.3 Comparison with Japanese Banks

In Japan, the average bank interest rate typically hovers around 0.001%, with exceedingly rare exceptions where it might fluctuate around 1%.

Notably, in 2022, inflation in Japan reached a record high of 2.5% (Inflation Tool, 2023).

3.5 Comparison with Russian Banks

In Russia, the average annual inflation rate for 2022 was approximately 12%, with some months seeing it rise as high as 17% (CBR, 2023). Concurrently, the average bank deposit rate ranges from 6% to 12% annually, typically involving capitalization of the amount and monthly interest disbursements. In essence, deposits have the potential to mitigate the risks associated with inflation.

3.6 Comparison with Neobanks

The majority of neobanks provide flexible terms. For instance, when utilizing Revolut and Wise, your account balance can accrue interest daily at a rate of 4% to 5% per annum. In most cases, an investment fund oversees these funds. For instance, in Wise, Revolut, and Monzo, BlackRock manages the funds, which are invested in a variety of assets including funds, bonds, and shares.

It's important to note that there is no provision for fund insurance, meaning investments carry the full risk, albeit with a relatively low interest rate per annum.

3.7 Comparison with Investment Funds

Investing in investment funds offers substantially higher returns compared to bank deposits. The average yield on funds in 2022 ranged from 8% to 25%. However, it's essential to note that these funds are not insured, thus, the risk of loss remains a tangible possibility. Engaging in investments necessitates a degree of knowledge and experience in market trading and global economics.

No periodic payments are made throughout the year; fund values fluctuate based on the prevailing market conditions.

3.8 Conclusion on Profitability

Banks:

Investing in bank deposits is the least risky way to save money, requiring no additional knowledge. In most countries, deposits up to a certain amount are insured by the state in the event of bank failure. However, considering inflation, this method at best preserves the amount or incurs a slight loss, contingent on the level of inflation in a given period. Typically, funds in the deposit are not readily accessible, and premature closure may result in interest recalculation at a minimum rate.

Neobanks:

Utilizing neobanks as an investment tool can yield bank interest rates or slightly higher. In most cases, funds are accessible, and interest accrues on a daily or monthly basis. No specialized knowledge is necessary. However, it's imperative to note that funds are not insured, entailing a risk of investment loss.

Investment Funds:

Investment funds typically allocate clients' capital based on the level of risk, into funds, bonds, indices, and stocks. On average, returns on such investments surpass bank interest rates. However, this type of investment requires a solid understanding of market dynamics and the broader economy. Funds are not insured, and the risk of investment loss ranges from moderate to high.

&Volume

Employing the riskiest form of investment management - cryptocurrency day trading - &Volume provides clients full control over their deposit. Transactions are executed around the clock, with results visible daily. The risk of losses is cyclical, generally higher than with investment funds, and contingent entirely on market conditions. No specialized knowledge is required. Despite the elevated risk, the returns from such investments far surpass bank interest on deposits and significantly outstrip the returns from investment funds.

4. Recommendations

Recommendations

  • When making any investment decision, we recommend that you carefully consider all factors and understand your investment objective.

  • Most experts advise not to invest all your savings in one instrument, but to divide your savings into several instruments so that most of the funds are most reliably protected (SEC, 2023).

  • Spreading your money across different types of investments, such as international stocks and bonds, can reduce your risks. In a properly diversified mix of investments, if any particular investment or market performs poorly, the performance of other investments can help maintain overall returns and mitigate the impact of losses (FCA, 2023).

Key conditions for financial success according to the SEC:

Make a financial plan. Pay off all high interest debts. Start saving and investing as soon as you pay off your debts.

Before deciding to use any investment instrument, we recommend that you familiarize yourself with two useful sources:

Risk Warning.

&Volume is an investment instrument with increased risk.

Investments always involve risk. Your capital may increase or decrease, and you may not always be able to recoup the entire amount you invest. Past performance is not a guarantee of future results.

  • You may lose all the money you invest.

    The state of most crypto assets can be very unstable: their value falls as quickly as it rises. You should be prepared to lose all the money you invest in crypto assets and cryptocurrency trading.

  • The crypto-asset market is practically unregulated. There is a risk of losing money or any crypto assets purchased due to risks such as cyber attacks, financial crimes and exchange failure.

  • Don't expect that your money is insured if something goes wrong. Cryptocurrency and cryptocurrency trading is not a regulated activity and is not insured by any government.

  • Don't put all your eggs in one basket. Putting all your money into one type of investment is risky. Spreading your money across different investments makes you less dependent on which one will do well

  • A good rule of thumb is to not invest more than 10% of your money in high-risk investments.

Project Roadmap

The original idea of the project was to provide automated trading services in traditional stock and foreign exchange markets. However, entry into the traditional market is blocked by a strong regulatory barrier, which requires enormous initial investment to overcome. Therefore, we began the development of the project from the cryptocurrency market, as it is less regulated and open to experimentation.

  • Project Launch Stage

    At the first stage, we launched the main service - automated management of trading in crypto-pairs. At the moment, this is our primary product, with a very low barrier to entry. After load testing on a limited audience of up to 5,000 people, the project will be launched for a wider audience.

  • Enable Pro Mode

    Initially, no user settings for our system are provided. However, at the second stage, we will grant access to some settings. To change system settings without compromising efficiency, the user must have certain knowledge in trading. We'll also expand the number of supported crypto exchanges.

    For users who trade cryptocurrency on their own, we are preparing to release our own trading terminal with a built-in connection to our AI system.

  • Expansion of Investment Methods.

    In addition to automated trading, we plan to add tools for long-term investment in crypto, automated purchase of crypto assets in long-term mode, as well as automated balancing of the investment portfolio.

    To manage the balancing of crypto-portfolios in the medium and long-term modes, we are developing an automated sentiment analysis system based on open language models.

  • Algorithmic Trading in the Stock and Foreign Exchange Markets.

    We plan to accumulate enough liquidity to independently or through an umbrella broker begin to provide automated trading services for short-, medium-, and long-term investing.

Conclusion

&Volume is developing effective investment management methods and offers a product that helps users maximize their profits, taking into account their individual goals, financial situation, and risk tolerance.

We have a clear vision of how to further develop our product and in which direction.