Recently, a friend asked me how to use the KD indicator to determine entry points, and I realized that many people’s understanding of this classic indicator is still only at the surface level. Rather than saying it’s “KD,” it’s more accurate to say that many people confuse the difference between KD and KDJ. Today, let’s talk about the logic behind it.

When it comes to the KD indicator, first you need to understand how it’s “forged.” It consists of three components—RSV, the K value, and the D value—progressing step by step. RSV stands for the immature stochastic value. Put simply, it shows where today’s closing price sits within the high-low range over the past few days. The calculation is: (today’s closing price − the lowest price in the most recent n days) divided by (the highest price in the most recent n days − the lowest price in the most recent n days), then multiplied by 100. The default n is 9. So if today’s closing price is at the highest point in the past 9 days, RSV is 100; conversely, if it’s at the lowest point, RSV is 0.

Next is the K value, which is the line in the KD indicator that truly reacts quickly. It takes today’s RSV and yesterday’s K value and computes a weighted average using this formula: (yesterday’s K value × 2/3) + (today’s RSV × 1/3). This way, it preserves the sensitivity of RSV while effectively filtering out some noise. Then the D value is obtained by smoothing the K value again using the same approach: (yesterday’s D value × 2/3) + (today’s K value × 1/3). As a result, the D value becomes the most stable line within the KD indicator.

Some trading software displays an additional line, usually in purple or yellow—that’s the J value. The J value is calculated very simply: 3K − 2D. Its role is to amplify the degree of divergence between K and D. You can think of it as an extreme-signal prompt within the KDJ system. When J is above 100, it indicates extremely overbought; when it’s below 100, it indicates extremely oversold. So in essence, KD and KDJ are the same thing—just presented differently.

When I use the KD indicator myself, I generally don’t pay much attention to the J value. The reason is simple: the J value reacts too aggressively. Although it can quickly capture turning points, it also produces many false signals, which makes it easy for people to overtrade. In most cases, the plain KD indicator is enough.

As for the default parameters 9, 3, 3—this set of numbers definitely isn’t chosen at random. The 9 represents using the past 9 K-bars as the calculation period. In traditional finance, 9 days is roughly equal to two weeks of trading days. This length is just right to capture short-term fluctuations without being too slow to respond. The following two 3s are the smoothing counts for the K value and the D value. The first 3 applies a 3-day moving average to the raw RSV, filtering out sudden spikes and sharp drops on the day. The second 3 then applies another 3-day moving average to the K value to obtain the D value. With this step-by-step filtering, the support and resistance signals become especially precise.

Why can 9, 3, 3 become mainstream? One reason is that it performs well in range-bound market conditions across stock markets, the foreign exchange market, and the crypto market. It can help you judge trends, and through golden crosses and death crosses, it provides straightforward entry timing. Another reason is that everyone uses the same set of parameters, which invisibly forms a collective consensus. When market participants all look at the same KD signal, the signal’s effectiveness becomes stronger.

Of course, KD indicator parameters are not fixed. If you do short-term intraday trading, you can reduce n to 5—changing it to (5,3,3). This will make golden crosses and death crosses appear more frequently, but the number of false signals will also increase. It’s best to pair it with other technical analysis tools to filter out noise. Conversely, if you want to do swing trading and don’t want to be stopped out by daily up-and-down fluctuations, you can increase n to 18—changing it to (18,3,3). The K and D curves will become smoother, and crosses will only appear when big trend reversals truly happen.

My experience is that different timeframes do require different KD parameters. On small timeframes like 5-minute or 15-minute charts, where random volatility is especially intense, I usually set the parameters to (14,3,3) to filter out noise. For hourly and daily charts, the default (9,3,3) feels quite comfortable, and the signals are more reliable for reference. If you look at weekly or monthly charts, (9,3,3) is also applicable. Even though there are fewer signals, each one is powerful—especially for long-term positioning.

A common misconception is that the more finely tuned the KD indicator is, the more accurate it must be. But that’s not true. Adjusting parameters is meant to match your trading strategy, not to predict the future. If you use very short parameters—for example, (3,2,2)—you’ll see countless cross signals. In the end, that only leads to overtrading and frequent stop-outs.

For most investors, the default KD parameters 9,3,3 are already sufficient for everyday needs. Traditionally, when KD exceeds 80 it’s considered overbought, and when it’s below 20 it’s considered oversold. However, in different markets, these thresholds may need to be adjusted slightly. The most important thing is not to chase novelty by using non-mainstream parameters, because that may cause you to lose the effectiveness that comes from market consensus.

Understanding the calculation logic behind KD and KDJ helps you find your own trading rhythm in choppy markets. Especially with the classic 9,3,3 parameters, which fit most markets, the collective consensus effect can also make support and resistance signals more precise. Finally, a reminder: these are just technical analysis tools—how you use them still depends on your own risk tolerance and trading strategy.