Money Machine

Trading

Types of trading

Discretionary

Discretionary trading is trading manually after evaluating market dynamics and utilizing one's skill and knowledge to assess the market, price action, and place high-probability trades. One needs to have a well-defined edge and be able to execute it properly. An edge can be knowing the key levels where price tends to reverse, reading market conditions such as bear or bull and the zones that price may travel to next, understanding options hedging by dealers and gamma exposure, reading order flow to predict the very next move in price, or trading correlated pairs and their discrepancies. There are many different types of edges, more than can be mentioned here. Discretionary trading allows for a more human and nuanced assessment of the day, the context, and the situation, but can also make room for more human error. Trading is typically performed in real time while on the charts, but some trades may be placed and then left to play out. The approach can include day trading, swing trading, and even long-term investing. The edge needs to be simple, mechanical, and repeatable, so that setups can be found multiple times a day most days. While this type of trading is performed manually, execution should be as systematic as possible to preserve probabilistic advantage.

Often, the hardest part of discretionary trading is not finding an edge or even the ability to execute it at the right time and with precision, but the psychological aspects of trading in this manner. Human emotions come into play with regard to managing losing trades, proper risk management, position sizing, resisting the gambling and dopamine attractions that draw traders in, and overall capital management. Common mistakes include over-trading, not taking the time to analyze the market, revenge trading (chasing a loss), altering the systematic elements of an edge or not following rules, not sticking to a reasonable trading plan, and lacking patience in execution. The trader tries to bend the market to their will rather than letting the market come to them. A person's mood can be highly impacted by losses and poor outcomes, and the negative feeling can spiral, causing even more erratic and problematic behaviors, such as in gambling. The trader needs to learn how to have an even mindset, have resilience, and be able to control emotions, stick to a predetermined plan with mathematical advantage, and seek support when needed. The psychological maturity required for consistently profitable discretionary trading can take years to develop and is a primary obstacle to success in this approach.

Algorithmic

Algorithmic trading involves the use of coding and computers to develop programs that trade the markets based on a predetermined set of criteria. Once the algorithm is designed, a bot executes the program on various platforms in order to place trades and exit out of them based on the parameters of the strategy. While some algorithms can trade multiple markets effectively, some are specific to certain assets. Ideally, one might use several or many algorithms together to help support the robustness of the overall approach. These different algorithms can balance one another, shoring up the others' weaknesses, and helping to diversify a portfolio not simply by asset-class diversification but also by approach and timeframe. In a crude sense, an algorithm defines entry criteria, exit criteria, position sizing, and risk management for each trade. Each aspect may involve filters and parameters to define it. For example, the entry criteria may have one to three filters or parameters that determine its execution, and similarly for the exit criteria. Simple is better, since markets are hard to predict and market activity will wind through many different conditions. The more complexity, the more room for error, and the harder it becomes to measurably predict success. But more sophisticated modern algorithms don't always follow the typical pattern. Some trade continuously, adjusting position sizing dynamically, re-allocating automatically between assets and strategies, and utilizing hedging and rotation risk strategies rather than stop-losses or liquidation. Algorithms can be designed on a wide variety of both new and historical edges, backtested across years of market data, incubated, and ultimately tested in live markets, both with simulated funds and real money, in order to determine robustness and effectiveness. Several complementary algorithms working together often produce better results than one running on its own.

Coding up an algorithm is not difficult. Often, the hardest part of algorithmic trading is avoiding common mistakes and the refinement of implementation. Very often, a backtest will produce a positive result, but the strategy will fail in live markets, and this has to do with many factors. Backtests may not reliably mimic current or modern market conditions and dynamics, since markets, policies, and how financial business is conducted change over time. The data used for testing often has some problems with it, even if obtained from a reputable source and even when utilizing AI. The most advanced professionals take great care to clean their data in numerous ways to ensure accurate results in testing. Overfitting, survivorship bias, lack of diverse regime inclusion, and exclusion of fees and slippage are common errors that produce inaccurate testing results. The execution of code with regard to risk management, diversification, and position sizing is also an art, not just a science. Algorithmic trading is more systematic in its use of rigid parameter sets for decision-making, but it still involves a human element of discretion in how the code is programmed, combined, and applied. There is still art to this type of trading.

Like discretionary trading, algorithmic trading typically takes years to learn and implement properly. It helps to have mentors well-versed in the live-market use of code, not just in research. This type of trading is generally used for swing trading and long-term holding of positions, and less for day trading. Some use it in specific situations for intraday applications, but generally algorithms work better for trades between 3 and 250 days. While algorithmic trading can produce decent results over a longer period of time if learned and executed properly, the learning curve is steep, and systems need to be monitored and managed. It is not a “set it and forget it” type of trading.

Both types of trading, discretionary and algorithmic, have their advantages and disadvantages. One is not inherently better. Traders tend to pick what suits their personality and lifestyle best. Algorithmic trading can reduce some human emotion and error, but there are many errors that can be introduced into algorithmic trading as well. Discretionary trading probably requires less analytical talent and can be effectively deployed intraday, but approximately 95% of discretionary day traders are unsuccessful, so the odds are against finding an edge and being able to implement it successfully and consistently. Algorithmic trading can more easily blend itself into long-term investing, so the disciplined trader may be able to rely on the advantages of time and compounding, as funds and strategies can be shifted and overlap with long-term investment approaches. For the cautious trader, this can offer some safety and better risk management, but it takes more time to play out. Some say that day trading can be used as a short-term income engine for longer-term algorithmic strategies, and in that way these approaches can be combined, but it assumes consistency and profitability in the discretionary trading, which is difficult to achieve. Both discretionary and algorithmic trading are used by hedge funds to compound money faster and generate outsized returns. The larger Wall Street institutions tend to rely more on algorithmic approaches, since risk management, capital preservation, reporting, and a systemized approach to managing large amounts of capital in the billions can be easier. But the use of algorithms by large institutions is one reason why some say it is difficult to find an edge and inefficiencies to exploit using code. There are edges in both types of trading. Discretionary traders can be more nimble and flexible, as can algorithmic traders managing smaller portfolios who don't answer to investor clients or are comfortable with more aggressive approaches. Some hedge funds and bank trading desks utilize a blend of discretionary and algorithmic trading.

Asset Classes

While discretionary traders often focus and specialize in one asset class or strategy, such as options, stocks, futures, gold, currency pairs, crude oil, or trading stock indices, algorithmic traders can take advantage of broad diversification across strategies and asset classes, all working together, since the complexity of monitoring conditions and execution is relegated to the computers. Discretionary traders do better to focus on and specialize in one approach and one to five assets; algorithmic traders do better running highly diversified strategies.

There are many assets to trade, from crypto, to individual stocks, to ETFs that can include baskets of stocks or certain sectors, to currency pairs, commodities, futures, options, and more. Choosing an asset to trade sometimes depends on the part of the world you live in, the access and regulations of the financial system in your region, and when you're available to trade. Currency pairs require a little more understanding of complex macroeconomic factors. Crude oil and gold require an in-depth understanding of those specific assets. Stocks can be useful, as there are many to choose from, but also present complexity in that variety of choice relative to indices. Stocks can also gap up or down. Futures can be fairly straightforward, but have rolling contracts and certain peculiarities. Crypto is often the most volatile and unpredictable. Choose an asset you can study and become proficient in, and that fits your lifestyle. Stocks generally require that you trade your own money. In other assets, such as futures, it can be easier to secure funding to trade other than your own capital.