An algorithmic trading strategy, also known as algo-trading, is the practice of employing programmed computerized algorithms to execute trading orders with speed and efficiency that is usually beyond human capacity. It involves pre-defined instructions based on variables such as time, price, volume, and other quantitative aspects of trading. This method eliminates the impact of emotional and psychological aspects of trading, resulting in a strategy that is more systematic and objective.

• Speed and Efficiency: Algorithmic strategies can execute trades at a speed and frequency that is beyond human ability. The algorithms can scan and analyze the markets, and execute trades based on the pre-determined criteria within fractions of a second.
• Disciplined and Emotion-free Trading: Algo-trading involves strict adherence to the pre-programmed strategies regardless of fear, greed, or any other emotional considerations that may influence human decision-making in trading.
• Reduced Transaction Costs: Through quick order execution at the best possible prices and the automated checking of market conditions, the transaction costs are often reduced.
• Backtesting Capability: Algorithmic trading strategies can be backtested on historical and real-time data to verify their viability before any actual trading.
• Scalability: The trading algorithms can be applied to different scales of trading, ranging from small individual trades to large-scale institutional trades.

Implementing an algorithmic trading strategy requires a meticulous approach and involves several steps. Initially, one must define their strategy and codify it into a computer program. Afterward, the trader must backtest the strategy using historical data, optimize it, and finally implement the strategy in live markets.

Choosing a broker who provides the necessary platform and tools for algorithmic trading is another crucial aspect. Algorithmic trading also requires continuous monitoring, fine-tuning, adjusting to market changes, and regular auditing to ensure optimal performance.

Algorithmic trading offers a plethora of advantages starting with:

• Precision and Speed: Owing to precise algorithms and high-speed computers, trades are executed at the best possible prices.
• Reduction of Human Error: Since trades are executed automatically based on coded instructions, the probability of human error is significantly reduced.
• 24-hour Trading: Algorithmic trading can function round-the-clock, ensuring that no potential trading opportunity is missed.
• Backtest Capabilities: Algorithms are backtested on historical data to determine their effectiveness, safeguarding money before application in live trading.
• Consistency: Traders can be consistent with their trading approach since the algorithm strictly follows its pre-programmed rules, removing the impact of emotional or instinctive trading decisions.

However, regardless of numerous benefits, some demerits of algorithmic trading include:

• System Failure: Since algorithmic trading relies heavily on technology and internet connectivity, any malfunction, slowdown, or loss of connection could result in loss of trading opportunities or managing risks.
• Over-optimization: Although backtesting is valuable, traders run the risk of relying too heavily on the results and creating an algorithm overly suited to past performance. This concept, called over-optimization, doesn't guarantee future outcomes.
• Market Manipulation: Sophisticated techniques known as “order sniffing” algorithms may be used by unscrupulous traders to exploit the automated algorithms and manipulate the market for their advantage.
• Communication Delays: Due to the time it takes to route an order from one's computer to the exchange's server, a lag can cause the prices to change substantially.

In conclusion, algorithmic trading strategies can offer a systematic, disciplined, and efficient way of trading that can maximize profits and minimize losses if implemented correctly and prudently. Primarily, the success of algo-trading hinges on the quality and robustness of the underlying algorithm, full understanding of its functioning, and effective risk management.