Best indicators for day trading: what each does and where it fails

Contents

1. Is there a best indicator for day trading?
2. What do the most common day trading indicators measure?
3. Why does the same indicator work on one chart and fail on another?
4. Trend or mean reversion: which indicators combine well?
5. How do you test an indicator's settings before using it?
6. What can an indicator backtest tell you, and what can't it?
7. Frequently asked questions
8. References

Is there a best indicator for day trading?

There is no best indicator for day trading. Moving averages and VWAP describe trend and fair value. RSI describes momentum. ATR and Bollinger Bands describe volatility. Whether any of them helps depends on the market, the bar size, the exits and the costs. The useful answer comes from testing a complete rule on data it has not seen.

A technical indicator is a calculation on past prices, and sometimes volume, that turns a chart into a line or a number. Day trading is opening and closing positions within the same trading session, so no position is carried from one session into the next. Each indicator answers one narrow question about recent bars. Examples are "Is price rising?" and "Is the market unusually quiet?"

Asking for the best indicator is like asking whether a hammer or a screwdriver is the best tool. The answer depends on the job. A hammer drives nails well and does nothing for screws. In the same way, an indicator that helps a trend-following rule on 15-minute bars can hurt a mean-reversion rule on 1-minute bars.

What do the most common day trading indicators measure?

The five indicators most day traders start with measure four different things: direction, session value, momentum and volatility. The table shows what each one measures, its common default setting, and where it usually breaks down.

Indicator	What it measures	Common default	Where it fails
Moving average	Average price of the last N bars; shows direction	20 bars, simple or exponential	Lags at turning points; gives many false crosses in sideways markets
VWAP	Average traded price of the session so far, weighted by volume	Resets at each session open	Trend days, when price stays on one side all session; depends on the session start you choose
RSI	Size of recent up moves compared with down moves, on a 0–100 scale	14 bars, levels 70 and 30	Stays above 70 or below 30 for long stretches in strong trends
ATR	Average size of each bar's true range, gaps included	14 bars	Gives no direction; reacts late after a sudden jump in volatility
Bollinger Bands	Moving average with bands a set number of standard deviations away	20 bars, 2 standard deviations	A band touch on its own is neither a buy nor a sell signal

Moving averages. A moving average is the average closing price over the last N bars, recalculated on every new bar. The simple version (SMA) weights every bar equally. The exponential version (EMA) gives more weight to recent bars, so it turns sooner. It also gets whipsawed more often, which means it gives more signals that reverse quickly.

VWAP. VWAP (volume-weighted average price) is the average price of all trades in the session so far, where each price counts in proportion to the volume traded at it. The line depends on where the session starts. TAIEX futures have a day session from 08:45 to 13:45 Taiwan time and an after-hours session from 15:00 to 05:00 the next day. Exchanges can change session hours, so check the TAIFEX contract page before setting up a test. A VWAP that starts at 08:45 is a different line from one that starts at 15:00 the afternoon before. Pick one start, and use the same one in the backtest and in live trading.

RSI. RSI (Relative Strength Index) compares the average size of recent up closes with recent down closes and scales the result from 0 to 100. In a ranging market, a reading above 70 often comes before a pullback. In a strong trend, the same reading often shows strength, and price keeps rising. That is why "RSI above 70 means sell" fails on trend days.

ATR. ATR (Average True Range) is the average of each bar's true range. The true range is the largest of three distances: high to low, previous close to high, and previous close to low. ATR is most useful for sizing stops and targets and for skipping quiet periods. On 5-minute bars it usually rises around the open and the close. A stop set at a fixed multiple of ATR is therefore wider at the open than at midday.

Bollinger Bands. Bollinger Bands are a moving average with an upper and a lower band placed a number of standard deviations away. Standard deviation measures how spread out recent closes are. Two standard deviations does not mean price stays inside the bands a fixed share of the time. Intraday prices jump, and volatility changes during the day. A close outside the band can start a breakout or mark the end of a move. The band alone does not tell you which.

Why does the same indicator work on one chart and fail on another?

The same indicator gives different results because four things change around it: the market, the timeframe, the exits and the costs.

Factor	What changes
Market	Some contracts trend more within a session. Others swing back and forth more. Tick size and liquidity also differ.
Timeframe	Shorter bars give more signals, smaller moves per trade and more noise.
Exits	The same entry can show a profit with a trailing stop and a loss with a fixed target, or the other way round.
Costs	Commission and slippage are paid on every trade, however small the move.

Costs matter most for day traders, because intraday moves are small compared with the fixed cost of each trade.

Put costs into the test from the first run. The guide to trading costs in backtests shows how to compare the average trade with the round-trip cost.

Trend or mean reversion: which indicators combine well?

Trend and mean-reversion indicators answer opposite questions. Combine them by giving each indicator a different job, not by stacking indicators that say the same thing. A trend indicator signals when price keeps moving in one direction. A mean-reversion indicator signals when price has moved far from an average and may come back toward it.

A market regime is the kind of behaviour the market shows for a period, such as trending, ranging, quiet or volatile.

Job in the rule	Question it answers	Typical indicators
Regime filter	Is the market trending or ranging? Is it moving enough to trade?	Moving average slope, price above or below VWAP, ATR level
Entry timing	When should I enter within that regime?	RSI pullback, Bollinger Band touch, moving average cross
Exit and size	How far away is "wrong"? How large is the position?	Stop at a multiple of ATR, VWAP as a target

A complete rule built this way could look like the following. Trade long only when price is above the session VWAP and a 50-bar EMA is rising. Enter when a short RSI dips below 30 and then closes back above it. Place the stop 1.5 ATR below the entry. Close any open position 10 minutes before the session ends. Each indicator has one job, and the rule says exactly what happens at every step.

Avoid pairing indicators that measure the same thing. RSI and the Stochastic oscillator both describe where price sits within its recent move. Using both adds parameters but little new information, and every extra parameter is one more setting that can be fitted to noise.

How do you test an indicator's settings before using it?

Test the area around a setting and how it behaves on unseen data, not just the single best number. A parameter is a setting you choose, such as the RSI length or the entry level. Out-of-sample data is history that was not used to choose the rule or its parameters.

1. Write the full rule. Define the entry, exit, stop, position size, trading session and end-of-session exit.
2. Set costs and execution. Include commission and slippage. Fill on the next bar after a signal at the bar close.
3. Hold back recent data. Set aside the latest part of the history, for example the last 30%, and do not look at it yet.
4. Run a grid on the rest. Test a range of values, such as RSI lengths from 7 to 21 and entry levels from 20 to 35.
5. Choose from a stable area. Pick a value in the middle of a region where neighbouring settings give similar results, not the single peak.
6. Run the held-back data once. Compare the average trade and the drawdown with the in-sample result.
7. Repeat over time. Use walk-forward analysis to repeat the optimize-then-test cycle across many windows.
8. Write down how many variants you tried. The more you tried, the more the best one is flattered by chance.

In AlphaPrime (Valynox), strategy generation lets you set parameter ranges, entry and exit conditions and search constraints. Parameter ranges like the ones in step 4 are set as part of the research setup.

Check	Stronger sign	Weaker sign
Parameter neighbourhood	Nearby settings give similar results	Only one setting works
Trade count	Dozens of trades or more in each year tested	A handful of trades in total
Out-of-sample	Keeps a reasonable share of the in-sample average trade	Turns to a loss, or loses most of the edge
Years and sides	Most years contribute; long and short are both reasonable	One year or one side carries the whole result
Cost stress	Still positive with double the slippage	Fails with one extra tick of slippage

What can an indicator backtest tell you, and what can't it?

An indicator backtest shows how one specific rule behaved on past data under your assumptions. It is evidence, not a forecast, and not proof that the rule will work.

First, indicators add no new information. They rearrange the same prices and volume you already see on the chart. Every one of them looks backward, so every one lags.

Second, searching creates false winners. Picture 200 people who each flip a coin 10 times. On average, about 11 of them will get 8 or more heads. None of them is a skilled flipper. Testing many indicators and settings works the same way: some will look strong by luck. Sullivan, Timmermann and White (1999) showed how to adjust technical trading rule results for the number of rules searched. Rules that looked strong over earlier decades did not keep that strength in their later out-of-sample period. White (2000) describes the method behind this adjustment. Bailey and co-authors (2014) show that the best backtest from a large search looks better than it will perform, even when no variant has a real edge. The article on overfitting in strategy generation covers how to keep this in check.

Third, markets change. Volume patterns within the day, tick sizes and session hours shift over the years. A rule tuned to one period meets different conditions later.

Fourth, the result is only as good as its assumptions. Fills, slippage, contract rolls and missing bars all change the numbers.

What this means if you use AlphaPrime: build indicator rules with parameter ranges and realistic costs, then send candidates through strategy validation. Validation offers Basic, Standard and Extensive levels, walk-forward analysis and your own thresholds. Treat the strategies that pass as candidates for further review, not as proven strategies.

Frequently asked questions

What is the most accurate indicator for day trading?

No indicator has a fixed accuracy. How often a signal is right depends on the market, the bar size, the exit and the costs. A rule can be right often and still lose money if its losing trades are larger than its winning trades. Judge a complete rule by its average trade after costs on unseen data.

What RSI setting is best for day trading?

The common default is 14 bars with levels at 70 and 30. Shorter lengths reach extreme readings more often and give more signals. No setting is best in general. Test a range of lengths and levels, and choose a value from an area where neighbouring settings give similar results.

Should I use VWAP or a moving average for day trading?

They answer different questions. VWAP starts fresh each session and weights prices by volume, so it shows the average price traded today. A moving average rolls across sessions and shows recent direction. Many intraday rules use VWAP as a value level and a moving average as a trend filter.

How many indicators should a day trading strategy use?

Usually one indicator for each job: a filter, an entry trigger and an exit or stop. That is often two or three in total. Each extra indicator adds parameters that can be fitted to noise. Add one only when it answers a question the others do not.

Can I trust a backtest of an indicator strategy?

You can trust it as evidence about the past, under the costs and fill assumptions you set. It becomes more convincing when nearby settings also work and when the rule holds up on held-back data and in walk-forward tests. It never shows what will happen next.

References

• Sullivan, R., Timmermann, A., & White, H. (1999). Data-Snooping, Technical Trading Rule Performance, and the Bootstrap. The Journal of Finance, 54(5), 1647–1691.
• White, H. (2000). A Reality Check for Data Snooping. Econometrica, 68(5), 1097–1126.
• Bailey, D. H., Borwein, J., López de Prado, M., & Zhu, Q. J. (2014). Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance. Notices of the American Mathematical Society, 61(5), 458–471.

What AlphaPrime does here

AlphaPrime's strategy generation turns indicator ideas into testable rules with parameter ranges and entry and exit conditions. Strategy validation checks them with walk-forward analysis and your own thresholds.

See plans and pricing