Investing is like detective work! A complete breakdown of expert trading systems, from win rate and expected value to conditional orders

Many people think "automated trading" means writing a complex program and letting the computer trade on its own. But that’s not the case. Investing is like detective work—don’t act without evidence!

We have invited our guest, Gary Sir, to provide a detailed explanation:How to gradually transform intuition-based trading into a data-driven, rule-based, and verified process, assisted by tools. By leveraging win rate, expected value, backtesting, and conditional orders, we turn "I feel" into a verifiable trading system in four steps.

Starting from his own experience in investigative work: detectives don’t act immediately just because they "feel" someone is suspicious; they first gather evidence. Investing is the same—you shouldn’t buy immediately just because you heard a rumor or saw a golden cross.From market analysis, win rate, and expected value, to MACD and quantitative backtesting, and then to Bull/Bear alerts, limit orders upon price touch, and trailing stop-limit orders,they all revolve around one central theme:Transforming trading from "I feel" to "What evidence do I have to support this?"。

1. Investing and detective work are actually very similar: First gather evidence, then decide on action

Given Gary’s background in investigative work, this mindset can be directly applied to investing. The hardest part of an investigation isn’t the final arrest, but the preceding steps of tracking, gathering evidence, and verification. Trading is no different. Seeing a stock appear poised to rise doesn’t mean you should buy immediately; hearing others mention a stock’s potential shouldn’t serve as a direct rationale for trading. He breaks down the entire process into:：Data Collection → Forecasting → Signal Confirmation → Action. First, obtain market data, then formulate hypotheses; next, observe whether new signals support the initial judgment, and only then decide whether to place an order.。This is the key difference from the traditional approach of "buying as soon as you hear a tip" and represents a more systematic trading method.

2. You can still lose money even if you predict the direction correctly; the real difference lies in "how you execute."

When people trade, they often flock to the hottest and largest-cap stocks in the market,Why do some people make money while others lose on the same stock?

The answer lies not in stock selection itself, but in execution.Even if you correctly anticipate a "rise then fall" or "fall then rise" pattern before the market opens, it doesn't guarantee a profit. This is because a series of questions remain:At what entry point should you get in? At what drop level does it indicate your judgment was wrong? At what price level should you close your position? What position size should you take? If profits are gained but then prices pull back, should you exit or stay? If "you got the direction right but still lost money," the reason is often not the direction itself, but not knowing how to exit. Therefore, the primary value of automation is not necessarily more accurate forecasting, but rather codifying certain operational rules that are otherwise prone to emotional influence.

3. Win rate alone is not enough; you must also consider "expected value" before trading.

Let's use card counting in Blackjack as an analogy.What truly matters isn't just knowing that a particular trade has "a chance to win," but rather calculating: What is the probability of winning? What is the average profit per win? What is the average loss per loss? Combining these figures yields the Expected Value (EV) in trading.Simplified, if a strategy has a high win rate but generates minimal profits per win while incurring significant losses per loss, the long-term results may still be suboptimal.Conversely, a strategy with a modest win rate can still achieve a positive expected value if the average gain from winning trades significantly exceeds the average loss.This concept also extends to options. It is crucial to remember that, beyond directionality, factors such as Implied Volatility (IV) also fluctuate. Therefore, simply "getting the direction right" does not guarantee a trade has a positive expected value.

4. The Kelly Criterion: A high win rate does not justify going all-in.

Once you know the win rate and expected value, the next step is:How much capital should be allocated? The core principle of the Kelly Criterion is to determine position sizing based on the win rate and the risk-reward ratio, rather than using the same stake for every trade.If the strategic edge is strong, consider increasing your exposure; if the edge is weak, position sizes should naturally decrease, or you might even choose not to trade at all.

However, the Kelly Criterion presupposes that your estimates of the win rate and risk-reward ratio are reasonably reliable. If the input data is biased, the calculated bet size will also be distorted. Moreover, the full Kelly criterion can be quite aggressive. Therefore, what is truly worth retaining is not the idea of "calculating a Kelly percentage and betting accordingly," but a more important concept:Position sizing should be tied to strategy edge, risk, and data reliability, rather than increasing bets based on gut feeling.。

5. Why do some people profit while others lose using MACD and SAR? The problem may not lie with the indicators themselves.

Many investors use technical indicators such as MACD, SAR, and moving averages. Some present "very attractive" historical examples: for instance, the market surging after a golden cross appears, making it seem simple to just trade following the signals. However, problems emerge when you look further ahead.The same MACD can yield completely different results across different years and timeframes. Why are the default parameters 12, 26, and 9? On what timeframe was it originally used? If it works on daily charts, does it still work on 5-minute charts? Has the win rate changed as market conditions shifted?If these questions are left unresearched and one simply uses MACD because "others say it works," they are effectively treating a tool as the answer to trading decisions. Therefore, the real focus is not whether MACD is good or bad, but rather:Any indicator should be backtested first before deciding how to use it.

6. What is algorithmic trading? It is not simply buying automatically when an indicator generates a signal.

Algo Trading is not merely coding MACD into a program to buy on a golden cross and sell on a death cross.True algorithmic trading involves first codifying trading logic into rules, then acquiring data for computation to verify whether the strategy holds a sufficient edge under specific conditions, and executing only when those conditions are met.。For example, one can start fromStart with price, volume, and time to calculate momentum, then incorporate conditions such as support/resistance levels and breakouts.But when you continue to ask: How strong does the momentum need to be? What magnitude of breakout is considered valid? In which market environments should it be applied? The question returns to the most fundamental process of quantitative research:Data → Hypothesis → Testing → Optimization.It is not about finding a set of magic parameters and using them forever, but rather continuously verifying whether the original rules remain effective.

7. The Biggest Change Brought by AI: You Can Start "Speaking" Your Strategy Even Without Knowing How to Code

Previously, one of the barriers to entry for quantitative trading was programming. In my early years, I had to use tools like MultiCharts to code my trading logic and then test it against historical data. But now, AI has significantly simplified this step. For example, you can firstdescribe your strategy in natural language: what conditions trigger entry; what conditions trigger exit; which indicators to use; what timeframes to apply; where to set stop-losses; and then let AI help organize this into strategy logic or even code.I have turned strategies into simplified indicators with red and green signals. On the surface, users only see color changes, but the actual calculation rules are embedded in the background. The greatest value of this approach is not that you "don't need to learn trading," but that it shifts the time previously spent on coding to more important matters:Is the strategy logic actually sound?

8. Once You Have a Strategy, Don't Rush to Use It; The First Step Is Backtesting

Implementing a strategy without historical data support is like taking action in an investigation without any evidence.Therefore, whether AI helps you generate dual moving averages, MACD, SAR, or other new indicators, the next step should always be backtesting.In the NiuNiu Quantitative System, you can view historical backtesting results, including equity curves and related reports.Additionally, while higher simulated returns were recorded during specific historical periods, the strategy still requires continuous refinement and improvement.

The real question backtesting should answer is not "How much profit was made in the past?" but rather:In which years was it effective? In which years did it incur losses? What is the maximum drawdown? Is the number of trades statistically significant? Does the edge persist after accounting for commissions and slippage? Do results become completely invalid under different parameters? If a slight parameter change turns huge profits into huge losses, be wary of overfitting.。Historical backtesting is merely a validation tool and does not guarantee that similar returns can be replicated in the future.

9. Say Goodbye to Screen Watching: How Can Alerts, Limit Orders, and Trailing Stops Help?

After completing the research, we move on to another key point in this live stream topic:Automating execution.First, you can set upprice alerts. For example, if a stock only warrants renewed attention when it reaches a certain price,there is no need to monitor the market from open to close. You can set an alert first and act only when the condition is met.。

Furthermore, you can usestop-limit orders. Suppose the currentprice has not yet reached your target range; you can set a trigger price first. The system will submit the pre-set limit order only after the market meets the condition.This way, youwon't have to stay glued to the quote screen waiting for that moment because of work, childcare, or other commitments.

The other isa conditional order of the trailing stop-loss/trailing take-profit type. For example, when a stock keeps rising, you don't necessarily have to sell immediately at your original fixed profit target; instead,let the exit price move upward with the market; the exit condition is triggered only when the price falls from its high by a pre-set number of points or percentage。

It addresses precisely the most common problem faced by many investors:The price hit the target, but I wanted to squeeze out a bit more profit; in the end, I didn't exit, and all gains were wiped out.。However, triggering a condition does not guarantee execution at the specified price. Slippage may still occur during fast markets, gap openings, or periods of insufficient liquidity.

10. What automation truly solves is not "no need to watch the market," but rather enforcement of discipline

Discipline is the most important aspect of investing.Even if you have a strategy with a decent historical win rate, if you fail to cut losses when you should, or change the rules due to greed when you should take profits, the actual results you achieve can still be completely different from the original system. Therefore, "saying goodbye to screen-watching" does not mean turning off your phone and ignoring everything entirely. The true meaning is:Pre-define your entry prices, exit conditions, and other criteria that would otherwise require constant manual monitoring by embedding them into rules or conditional orders as much as possible.。

Investors remain responsible for strategy formulation, while tools assist with execution.Additionally, he structures his personal trading schedule to concentrate research and trading activities within specific time windows rather than monitoring the market all day. However, this is merely his personal trading habit and not a universal "golden hour" suitable for everyone.When data analysis, backtesting, position management, and conditional orders are integrated, the true value of automation is not about "lazy investing," but rather about minimizing the interference of real-time emotions on your original trading plan.。

395K Views