Algorithmic Trading: How It Works, Types, Benefits and Risks
Last Updated on 1 month ago by Antony C.
I recently tried algorithmic trading, and the results were quite impressive. But the experience also taught me something more important: the algorithm is only as good as the settings behind it.
BEST FOR: Retail Investors Looking For All-In-One Trading App to Invest in SG, US, HK and China (H-Share).
securely through Moomoo’s website
A wrong number, an extra zero or one incorrect option can turn a promising strategy into an expensive mistake. Algorithmic trading can help an experienced investor become more efficient, but it can also help someone who does not understand trading lose money faster.
KEY TAKEAWAYS
21
Brokerage Reviewed & Analyzed
Driven by data, run by investors with real-world experience investing with our own money.
We’re supported by readers who buy via links on our site. While this may influence which products we write, it will not influence our opinions and evaluation. Learn more.
What Is Algorithmic Trading?
Algorithmic trading uses computer instructions to automatically place and manage trades based on a set of rules.
These rules may use information such as:
- The price of an investment
- Trading volume
- Time of day
- Technical indicators
- The size of your current position
- How much the price has moved
For example, an algorithm could be told:
Buy 100 shares when the price falls to $9.80, then sell them when the price reaches $10.10.
If those conditions are met, the system places the orders automatically. You do not need to sit in front of the screen and click the buy or sell button yourself.
However, algorithmic trading is not necessarily artificial intelligence.
Many algorithms simply follow fixed instructions. They do not learn, think or understand why the market is moving. AI trading may use more advanced models to analyse data, but normal algorithmic trading does not require AI.
It is also not the same as high-frequency trading. High-frequency trading is a specialised form of algorithmic trading used by professional firms with extremely fast systems. A retail investor running a simple grid strategy is doing something very different.
How Does Algorithmic Trading Work?
Although there can be a lot happening behind the screen, the basic process is quite simple.
- You create the rules.
You decide what must happen before the system buys, sells or changes an order. - The algorithm monitors the market.
It watches the selected price, volume, indicator or other data. - The conditions are triggered.
When the market matches your rules, the algorithm prepares the trade. - The order is placed.
The system sends the buy or sell order to your brokerage. - The algorithm continues managing the trade.
Depending on your settings, it may place another order, take profit, limit the position or stop trading.
Imagine that a share has been moving between $9.50 and $10.50. You set an algorithm to buy 100 shares at $9.70 and sell them at $10.10.
The system will follow those instructions without becoming tired, impatient or emotional.
But it will also follow them if you accidentally enter 1,000 shares instead of 100.
That is the part people sometimes overlook. The algorithm does exactly what you tell it to do. It does not know what you meant to do.
Common Types of Algorithmic Trading
There are many algorithmic trading strategies, but beginners only need to understand the main ideas.
| Type | What It Does | Main Risk |
|---|---|---|
| Trend-following | Buys or sells based on the direction of the market | The trend may suddenly reverse |
| Mean reversion | Assumes the price will eventually return towards its average | The price may continue moving away |
| Grid trading | Places buy and sell orders at different price levels | A strong one-way movement can create a large position or loss |
| Arbitrage | Tries to profit from price differences between related markets | The price difference may disappear before the trade is completed |
| Execution algorithms | Manage how an existing order is placed | The final execution price may still be worse than expected |
1. Trend-Following
A trend-following algorithm tries to identify whether the market is moving up or down.
It may buy after the price moves above a certain level or sell after the price falls below one.
This can work during a clear and lasting trend. However, the algorithm may make repeated losing trades when prices keep changing direction.
2. Mean Reversion
Mean reversion is based on the idea that prices sometimes move too far from their normal range and eventually return.
The algorithm may buy after a sharp fall and sell after the price recovers.
The danger is that the lower price may not be temporary. A company may have released bad news, or the entire market may be entering a longer decline.
3. Grid Trading
Grid trading places buy and sell orders at different price levels within a range.
For example, the system may buy each time the price falls by a certain amount and sell when it rises again.
This can produce frequent small profits when the market moves up and down within a stable range. However, it becomes much more dangerous when the price keeps falling or rising in one direction.
The number of grids, shares per grid and maximum position all matter. One wrong setting can cause the strategy to use much more money than expected.
4. Arbitrage
Arbitrage algorithms look for small price differences between related investments or markets.
Professional firms may use them to compare an ETF with its underlying assets or the same investment across different exchanges.
These opportunities are usually small and disappear quickly. Retail investors may also face slower data, higher costs and slower execution than professional firms.
5. Execution Algorithms
Some algorithms do not decide whether an investment is good. They only help manage how an order is executed.
For example, a Time-Weighted Average Price algorithm, or TWAP, may divide a large order into smaller trades over a set period.
A Volume-Weighted Average Price algorithm, or VWAP, tries to trade in line with the market’s trading volume.
These tools are often used to reduce the market impact of a large order. They do not guarantee a better price or a profitable investment. The SEC’s report on algorithmic trading explains how these execution algorithms are commonly used by brokers and institutional investors.
Benefits and Risks of Algorithmic Trading
Algorithmic trading can improve how a strategy is carried out, but every benefit comes with a trade-off.
| Potential Benefit | Important Risk |
|---|---|
| Executes trades quickly | Mistakes are also executed quickly |
| Follows rules consistently | The rules themselves may be wrong |
| Reduces emotional decisions | Cannot judge situations outside its programming |
| Monitors the market continuously | Technology can fail |
| Manages repeated trades | Frequent trading increases costs |
| Tests ideas using historical data | Backtests may give misleading results |
Biggest Benefit: It Can Reduce Emotional Trading
One benefit I like is that the algorithm does not panic when prices fall or become greedy when prices rise.
It simply follows the rules.
This can reduce emotional decisions, which are common in day trading and other short-term strategies.
However, removing emotion does not automatically improve the strategy. A calm computer can still make a bad trade.
Biggest Risk: Small Setup Mistakes Can Become Expensive
The biggest risk may not be the algorithm itself. It may be the person setting it up.
Common mistakes include:
- Selecting buy instead of sell
- Entering the wrong stock or contract
- Adding an extra zero to the quantity
- Using the wrong trigger price
- Setting the maximum position too high
- Choosing the wrong trading period
- Forgetting about different currencies
- Leaving the strategy running when market conditions change
This is why professional firms take testing and monitoring seriously. FINRA’s guidance on algorithmic trading highlights code testing, system validation, monitoring and the ability to disable an algorithm quickly.
A real-world example shows how serious technical mistakes can become. In 2012, a system update caused a broker-dealer to send millions of incorrect orders into the market. According to the SEC, the error caused more than US$460 million in losses.
A retail investor will probably never operate at that scale. But the lesson is the same: automation can multiply a mistake before you have time to react.
Common Risk: Backtests Can Look Better Than Real Trading
Most algorithmic trading tools allow you to test a strategy using historical market data.
This is called backtesting.
A good backtest may look impressive, but it does not prove that the same strategy will work in the future.
The result may not fully include:
- Brokerage commissions
- Minimum trading fees
- Bid-ask spreads
- Slippage
- Currency conversion
- Delayed market data
- Partial or failed orders
There is also the risk of overfitting. This happens when you keep changing the settings until they work extremely well for one historical period.
The strategy may appear brilliant because it was designed around what already happened. When the market changes, the same settings may stop working.
Research on the probability of backtest overfitting shows why repeatedly testing different settings can produce strategies that look profitable but fail outside the original data.
Hidden Risk: Frequent Trading Creates More Costs
A strategy may show a gross profit but still lose money after all costs are included.
For Singapore investors, costs may include:
- Brokerage commissions
- Platform fees
- Exchange and clearing fees
- GST
- Currency conversion
- Market data subscriptions
- Bid-ask spreads
- Financing costs when margin is used
These costs matter even more when the algorithm makes many small trades.
A HKD30 buying cost and another HKD30 selling cost may not sound like much on their own. But repeated across many trades, these costs can take away a large part of the strategy’s profit.
Why Market Conditions Matter
An algorithm may perform well because its strategy suits the current market.
That does not mean it will work permanently.
| Market Condition | Strategy That May Fit | What Can Go Wrong |
|---|---|---|
| Sideways market | Grid or mean-reversion | The price breaks out of the range |
| Strong trend | Trend-following | The trend suddenly reverses |
| High volatility | Short-term trading strategies | Large price gaps and poor execution |
| Low liquidity | Slower execution strategies | Wide spreads and incomplete orders |
| Major news event | Event-driven strategies | The market reacts differently from expected |
This is especially important for grid trading.
A grid strategy may look very effective when a share repeatedly moves between two price levels. It buys low, sells slightly higher and repeats the process.
But what happens if the company releases bad news and the price falls below the range?
The algorithm may continue buying because that is what it was programmed to do. Instead of collecting small profits, it may build a growing position in a falling investment.
The strategy did not suddenly become broken. The market condition changed, but the algorithm’s assumptions did not.
One thing I found useful about algorithmic trading was how clearly it showed the market’s small price fluctuations. You begin to notice how often prices move between different levels during the day.
However, understanding those fluctuations is not the same as being able to profit from them safely. The market still decides whether the strategy has the right conditions to work.
Is Algorithmic Trading Suitable for Beginners?
Algorithmic trading may be available to beginners, but that does not mean it is a good place to begin.
Before using it, you should understand:
- What you are trading
- Why the algorithm is buying and selling
- Which market condition the strategy depends on
- When the strategy is likely to fail
- How much money it can use
- How large the position can become
- How fees affect the result
- How the strategy can be stopped
If you do not understand the trade, automating it does not make it safer.
This is why I do not see algorithmic trading as a new or smarter way to invest. I see it as a tool that should only be used after you understand what you are getting into.
It can help someone who understands trading become more efficient. It can also help someone who does not understand trading lose money faster.
Is Algorithmic Trading Useful for Long-Term Investors?
Active algorithmic trading is generally unnecessary for dividend investors and long-term investors.
If your goal is to buy quality stocks, ETFs or REITs and hold them for years, you do not need a system making frequent trades throughout the day.
My usual approach is based more on dividend investing, regular investing and holding for the long term. I do not want investing to feel like another full-time job.
Simple automation can still be useful for:
- Regular monthly investing
- Rebalancing a portfolio
- Setting a limit price
- Dividing a large purchase into smaller orders
But this is different from using an algorithm to generate frequent short-term profits.
The tool should fit your strategy. You should not change your entire investing approach simply because a brokerage offers an interesting new feature.
Final Thoughts
Algorithmic trading has impressed me, but it has also made me more careful.
It can execute a strategy quickly, consistently and without emotion. But it cannot tell you whether the strategy is sensible, whether your settings are correct or whether the market has changed.
The algorithm is only the tool. The knowledge, settings, risk limits and final responsibility still belong to you.
BEST FOR: Beginner Retail Investors and Professional Traders Looking For All-In-One Trading App to Invest in SG, US, HK and China.
securely through Moomoo’s website
Frequently Asked Questions
Disclaimer: All views expressed in the article are independent opinion of the author, based on my own trading and investing experience. Neither the companies mentioned or its affiliates shall be liable for the content of the information provided. The information was accurate to the best knowledge of the author. This advertisement has not been reviewed by the Monetary Authority of Singapore. * T&C Applies
Join 900+ BUDDIES who are growing their wealth with our weekly Income Newsletter
Antony C. is a Singaporean dividend investor focused on building passive income through REITs, ETFs, and Dividend Stocks. With 15+ years of experience investing in Singapore, Hong Kong, and China markets, he founded IncomeBuddies.com to share practical wealth-building strategies tested in his own portfolio since 2008. His expertise has been featured in Yahoo Finance, Nasdaq, and NFAA, and he’s the published book author of "Start Small, Dream Big".