RSI 30/70: backtest results

The textbook RSI strategy: buy when the 14-period RSI says oversold (below 30), sell when it says overbought (above 70). We ran it on real daily data for 19 stocks. It beat simply buying and holding on 6 of 19 stocks.

Results by stock

Best: Tesla on +390.0%. One set of rules, fixed in advance, run on every stock: no tuning to make the numbers look better. A strategy with few trades or a big drop deserves less trust than its return suggests.

The exact rules

Paste those words into QuantProof and it runs the same test, shows every trade on the chart with entry, stop loss and take profit, and compares it with buy & hold. Change a number and run it again in seconds.

Free TradingView indicator (Pine Script v5)

Copy this into TradingView's Pine Editor and click Add to chart. It marks the buy and sell signals and draws the levels.

//@version=5
// Indicator generated by QuantProof from your rules.
// Built for AAPL on the daily (1D) chart. Config 8c6032ba37.
//   Buy when:  rsi(14) < 30
//   Sell when: rsi(14) > 70
//
// Execution, identical in the QuantProof backtest, this script and the alerts:
//   rules are read on the bar close, the order fills at the next bar's open
//   5 bps slippage per side and 1 bps fee per side are applied to every fill
//   stop loss and take profit are measured from the filled entry price, stop first when a bar hits both
//   one setup at a time: no new entry until the open one closes
// Marks where your rules say buy and sell, and adds alerts you can turn on in TradingView.

indicator("RSI 30/70 signals", overlay=true, max_boxes_count=300, max_lines_count=300, max_labels_count=300)

// Prices are written the way the market quotes them: a $500 coin does not have a target of 566.130924.
qpPrice(x) => str.tostring(x, x >= 1000 ? "#.##" : x >= 100 ? "#.###" : x >= 1 ? "#.#####" : "#.########")

rsi14 = ta.rsi(close, 14)
entrySignal = (rsi14 < 30)
exitSignal = (rsi14 > 70)

// One setup at a time, in the states QuantProof uses:
//   flat -> armed (signal on the close) -> active (filled at the next open) -> closing (exit signal) -> flat.
// While a setup is armed, active or closing, new entry signals are ignored.
var string state = "flat"
var float entryPrice = na
var string exitReason = ""
slipUp = 1 + 0.0005
slipDown = 1 - 0.0005

// 1. Orders queued on the previous close fill at this bar's open.
filledNow = state == "armed"
exitedAtOpen = state == "closing"
if filledNow
    entryPrice := open * slipUp
    state := "active"
if exitedAtOpen
    state := "flat"

// 2. While the setup is active, the stop is checked before the target, as in the backtest.
stopLevel = na
targetLevel = na
stopHit = false
targetHit = false
hitNow = stopHit or targetHit
if hitNow
    exitReason := stopHit ? "stop loss" : "take profit"
    state := "flat"

// 3. Rules are read on this close; the order waits for the next open.
// A setup can arm on the same bar an earlier one closed, exactly as the backtest does.
armNow = state == "flat" and entrySignal
closeNow = state == "active" and exitSignal
if armNow
    state := "armed"
if closeNow
    exitReason := "exit rule"
    state := "closing"

exitPrice = stopHit ? math.min(open, stopLevel) * slipDown : targetHit ? math.max(open, targetLevel) * slipDown : open * slipDown
closedNow = hitNow or exitedAtOpen
if closedNow
    label.new(bar_index, low, exitReason + " @ " + qpPrice(exitPrice), style=label.style_label_up, color=color.new(color.red, 25), textcolor=color.white, size=size.small)

// Levels stay drawn through the bar where the trade closes.
wasInTrade = state == "active" or state == "closing" or closedNow
levelBase = entryPrice

// The setup drawn as a position: target zone above the entry, stop zone below it, and the entry
// itself as a dashed line between them.
var box tpBox = na
var box slBox = na
var line entryLine = na
if filledNow
    if not na(targetLevel)
        tpBox := box.new(bar_index, math.max(targetLevel, entryPrice), bar_index + 1, math.min(targetLevel, entryPrice), border_color=color.new(color.teal, 55), bgcolor=color.new(color.teal, 88))
    if not na(stopLevel)
        slBox := box.new(bar_index, math.max(entryPrice, stopLevel), bar_index + 1, math.min(entryPrice, stopLevel), border_color=color.new(color.gray, 55), bgcolor=color.new(color.gray, 82))
    entryLine := line.new(bar_index, entryPrice, bar_index + 1, entryPrice, color=color.new(color.silver, 10), style=line.style_dashed, width=1)
if wasInTrade
    if not na(tpBox)
        box.set_right(tpBox, bar_index + 1)
    if not na(slBox)
        box.set_right(slBox, bar_index + 1)
    if not na(entryLine)
        line.set_x2(entryLine, bar_index + 1)
plot(rsi14, "RSI(14)", display=display.data_window)
plotshape(armNow, "Setup armed (fills next open)", shape.triangleup, location.belowbar, color.new(color.teal, 40), size=size.tiny)
plotshape(filledNow, "Entry filled", shape.labelup, location.belowbar, color.new(color.teal, 0), text="BUY", textcolor=color.white, size=size.small)
plotshape(closedNow, "Exit", shape.labeldown, location.abovebar, color.new(color.red, 0), text="SELL", textcolor=color.white, size=size.small)
alertcondition(armNow, "Setup armed", "RSI 30/70: {{ticker}} setup armed on the close, fills at the next open")
alertcondition(filledNow, "Entry filled", "RSI 30/70: {{ticker}} entry filled at the open")
alertcondition(closedNow, "Setup closed", "RSI 30/70: {{ticker}} setup closed")

Test your own version

• Different stock, crypto or forex pair? Change the ticker in the description.
• Add a stop loss or take profit and see if it helps.
• Saw a strategy on YouTube? Paste the link and get the rules, the backtest and the indicator.

Other strategies we tested

• Golden cross (50/200 SMA) backtest
• 9/21 EMA crossover backtest
• 20/50 EMA crossover backtest
• RSI dip in an uptrend backtest
• 20-day breakout backtest
• Price above 200-day SMA backtest
• 50-day SMA trend with 10% stop backtest