Backtest Centre

Buy a stock the day it hits its highest price in a year, but only if unusually many shares are changing hands. The idea is that a stock breaking into new territory on strong interest tends to keep going, because nobody who owns it is sitting on a loss waiting to sell.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps. A realistic simulation won 46% of trades with +0.85% average profit — before any costs, which would likely erase that.

Find a stock in a long-term uptrend that has just had a short bad patch, and buy it on the first day it turns back up. You are trying to get a discount on something that is still fundamentally rising.

Verdict. This holds up. After the signal, stocks beat comparable situations by +0.10% over the following month. Simulated with a real stop-loss and profit target, it won 47% of trades, averaging +8.76% on winners against -6.30% on losers — an average of +0.85% per trade over about 14 days. Note you lose more often than you win: the profit comes entirely from winners being bigger than losers, so cutting winners short breaks it.

Buy stocks that have been strong for months and are pushing to a new short-term high. The bet is simply that things already going up keep going up.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps. A realistic simulation won 42% of trades with +0.23% average profit — before any costs, which would likely erase that.

Buy when a stock in a long-term uptrend gets heavily sold off in the short term. You are betting the panic is overdone and it snaps back.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps. A realistic simulation won 41% of trades with -0.20% average profit — before any costs, which would likely erase that.

MACD is a popular momentum indicator built from two moving averages. When its fast line crosses above its slow line, many traders read it as a buy. We are testing whether that actually predicts anything.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps. A realistic simulation won 45% of trades with +0.36% average profit — before any costs, which would likely erase that.

Sometimes a stock opens far higher than it closed the day before - usually after news or earnings. This tests whether a jump of more than 10% overnight is followed by unusually big price swings, in either direction. It is a claim about how WILD the stock gets, not about which way it goes.

Verdict. This reliably tells you the stock is about to move a lot — roughly 2.7 times its normal 20-day swing. It does NOT tell you which way. That makes it useful for deciding position size, widening stops, or options strategies that profit from movement itself — and useless as a straight buy or sell trigger.

Same overnight jump, but now asking which way it goes next. Do you buy the excitement and ride it higher, or is the jump the top, meaning you should sell into it? These are opposite trades, so getting this right matters.

Verdict. The signal is real but points the wrong way. Following it as normally taught loses money (-3.98% versus comparable conditions). The tradeable version is to do the opposite of what the textbook says.

A stock opens far BELOW where it closed - usually bad news. The instinct is to get out. This tests whether that instinct is right, or whether the drop is an overreaction that recovers.

Verdict. The signal is real but points the wrong way. Following it as normally taught loses money (-1.75% versus comparable conditions). The tradeable version is to do the opposite of what the textbook says.

The most famous buy signal in charting. When the 50-day average price crosses above the 200-day average, traders call it a "golden cross" and treat it as the moment a downtrend turns into an uptrend. We are checking whether that is actually true, or whether the stock had already gone up long before the signal appeared.

Verdict. The signal is real but points the wrong way. Following it as normally taught loses money (-1.03% versus comparable conditions). The tradeable version is to do the opposite of what the textbook says.

The opposite signal: the 50-day average drops below the 200-day average. Widely treated as a warning to sell. We are checking whether stocks really do keep falling afterwards, or whether the worst is already over by then.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps.

RSI measures how overbought a stock is on a 0-100 scale. The idea here is that a golden cross is more trustworthy when the stock has NOT already run up hard (low RSI), and less trustworthy when it is already stretched (high RSI). We are testing whether that filter actually helps.

Verdict. This does not work. Stocks did roughly the same thing after the signal as they did without it, so acting on it is guessing with extra steps.