Recently, I’ve been chatting with a lot of traders and found that everyone really has a love-hate relationship with the MACD indicator. To be fair, MACD isn’t hard to understand—its three core components, the fast line, the slow line, and the histogram, are enough to get it. But if you want to truly use it well, the key lies in MACD parameter optimization.

Let’s start with the standard 12-26-9 set of parameters. Honestly, there’s a reason why this set of data has become the default on many trading platforms. EMA(12) captures short-term momentum, EMA(26) looks at long-term trends, and then the signal line EMA(9) filters out noise. This combination is indeed stable. Also, because everyone uses the same parameters, a certain “consensus effect” forms in the market. At key moments, important signals often draw more attention, which in turn increases the reference value of the signals.

But that doesn’t mean 12-26-9 is perfect for everyone. Especially in highly volatile markets like cryptocurrencies, or if you’re an ultra-short-term trader, this set may seem too smooth and not responsive enough.

I’ve tried many combinations myself. The 5-35-5 set reacts particularly fast, allowing you to capture the upswing and downswing points more accurately, but the trade-off is that there’s also a lot more noise and frequent false signals. The 8-17-9 is suitable for markets with larger volatility. The 19-39-9 leans toward medium- to long-term cycles and can effectively filter out most noise. The 24-52-18 is for long-term investors—it reacts slowly, but trends are clearly reflected.

There’s a trade-off you need to understand here: the higher the sensitivity, the sooner the signals arrive, but the more noise there is, and signal failure becomes more frequent too. The lower the sensitivity makes trend judgment more reliable, but you may miss some opportunities.

Last year, I backtested the daily price action of Bitcoin in the first half of 2025, comparing 12-26-9 with 5-35-5. Over the first half of the year, 12-26-9 produced 7 obvious signals, including 2 successful times when a “golden cross” led to an upswing, and 5 failures. The 5-35-5 set produced 13 signals—5 of them were followed by clear gains or losses, while the rest ended in failure.

The key was the rally starting on April 10. Both sets captured it. But the death cross from 5-35-5 came earlier, and it ate up some of the profit. This is the double-edged nature of high-sensitivity parameters.

When it comes to MACD parameter optimization, I have to pour some cold water on it: there is no absolute best set of parameters. Many traders adjust parameters and feel the results are good, then become obsessed with finding the “perfect parameters,” only to fall into the trap of overfitting instead. You look at historical candlestick charts and deliberately adjust parameters to make the backtest data look good—that’s like writing an exam by cheating with the answer key, and it has no way of matching real trading in the market.

My suggestion is this: beginners should first use the default 12-26-9 to observe, then adjust according to their own trading style once they’re used to it. Short-term traders can try 5-35-5 or 8-17-9, but they must backtest and verify. After you choose a set of parameters, don’t switch it every few days unless it truly performs poorly. Some people also watch two MACD setups at the same time to filter out noise—this can work too, but it increases the number of signals and also raises the requirements for judgment.

Finally, a reminder: MACD is just a tool. No matter how you optimize the parameters, it’s not a magic bullet. The key is to integrate it into your entire trading system, and use it together with other indicators and market structure to make decisions. After you adjust parameters, you must review and backtest to confirm whether it truly fits your logic—rather than blindly following other people’s settings. That’s how you can genuinely use MACD as your weapon, not as a stumbling block.