RSI(2) is a short-term rule for buying a dip. RSI is a score from 0 to 100 that drops when a stock has been falling, and the 2 means it only looks at the last two days. The rule buys at the close when that score drops under 10, but only while the price is above its 200-day average, so it buys pullbacks in a stock that is otherwise going up. It sells when the price closes back above its 5-day average. There is no stop loss, and it never sells short.
It's popular because it sounds safe. You only own the stock for a few days at a time, right after a drop, and you're in cash the rest of the time.
Being in cash is also why it fell behind. It made money on four of the five stocks and lost a little on Microsoft. All five went up a lot over these dates, and the rule missed most of that rise while it waited for the next dip.
In the table, the biggest drop is the most the account fell from a high point before it recovered.
| Stock | With the strategy | Just holding | Trades | Biggest drop | Backtest |
|---|---|---|---|---|---|
| SPY | $12,525 (+25.3%) | $14,692 (+46.9%) | 30 | -4.7% | See the backtest |
| QQQ | $10,743 (+7.4%) | $16,717 (+67.2%) | 27 | -11.2% | See the backtest |
| IWM | $13,207 (+32.1%) | $13,521 (+35.2%) | 26, plus 1 still open | -6.3% | See the backtest |
| AAPL | $10,007 (+0.1%) | $19,649 (+96.5%) | 22 | -7.5% | See the backtest |
| MSFT | $9,824 (-1.8%) | $12,168 (+21.7%) | 13 | -9.5% | See the backtest |
How we tested
We ran the textbook rules on SPY, QQQ, IWM, Apple and Microsoft from Mar 26, 2024 to the end of last month, in the same Backtest tab you can use. The rule needs 200 days of prices before its average exists, which is why the test starts later than our data does.
Every test starts with $10,000 and puts the whole balance into each trade, with no borrowed money. Costs are the tester's defaults for stocks: half a cent a share in commission and a cent of slippage on every fill. Slippage is the gap between the price you wanted and the price you got.
"Just holding" buys the same stock on the first day of the test and keeps it to the end.
Questions
Does RSI(2) beat buying and holding?
No. It didn't beat just holding on any of the five stocks after costs, even though it made money on four of them.
Why use the 200-day average with RSI(2)?
Without it the rule buys every stock that falls hard, including ones in a long slide. The average keeps it to dips in a stock that is otherwise going up.
Can I test it on a different stock?
Yes, with a paid plan. Press "Try it yourself" to open the Backtest tab with RSI(2) already loaded, then change the stock. A backtest of RSI(2) needs more history than a free account can reach.