Every month we pick one market and open it up: build a rule that buys the dips, we backtest it on real monthly prices against plain DCA, and the whole leaderboard is public — including how many entries failed to beat simply buying every month.
Two boards, two assets: one global, one Indian. Build a rule in the Strategy Lab and press Submit to contest — there is no form to fill in.
Loading…
Three steps, no payment, and your name never appears.
Open the Strategy Lab on this month’s asset. It starts on 40 years of real month-end prices, in your own currency.
Choose how far it has to fall before you stop buying and move to cash, and how far before you buy back in. Or let Find the best search thousands of combinations for you.
Press Submit to contest in the Lab; you can submit before a contest opens. Every morning at 07:00 IST we re-run every entry on the latest prices and rank them, under a random alias — never your name or email.
Dollar cost averaging — DCA — means putting the same amount of money in on the same day every month, whatever the price is. It takes the decision of when to buy off the table, which is the decision most people get wrong. In India the same thing is called an SIP.
On five criteria, each weighted for that contest and shown on its card: returns (annual return), risk to return (annual return per unit of worst fall), alpha (return over plain investing), beta (how much it moves with the market; lower is better) and effort (how many times the rule switches in and out; fewer is better). Each entry is scored by where it ranks among the contest’s entries on each one, and the weights combine them into a score out of 100. Only entries from a confirmed email address can win.
Sometimes, on some assets, over some periods — and far less often than the internet suggests. A rule that sells after a fall and buys back lower has to earn back what it gave up sitting in cash, and on a market that mostly goes up that is a high bar. That is the entire point of publishing the failure rate next to the winners: you can see how thin the odds are before you act on anybody’s backtest, including ours.
Every entry runs on the same monthly contribution, the same cash return and the same selling costs, over the last 10 and 5 years of real month-end prices, so the only thing that differs between two entries is the rule. Returns are money-weighted (XIRR), and ranked before tax — so nobody wins or loses on a tax bracket. The tax that rule would have cost is shown beside it as its own number, estimated under US rules on the global board and Indian rules on the India board.
No. A free account is enough to build a rule, submit it and be ranked. Entries appear under a random alias that you cannot edit and we do not derive from your email, so nothing on the leaderboard identifies you.
Free premium — the exact term is on this month’s contest card above. Premium is what shows you every entry’s exact thresholds and opens any of them in the Lab to change and re-test.
No. Everything here is a backtest on past prices, published for education. A rule that won the last ten years is not a prediction about the next ten, and past results are not indicative of future returns. Nothing on this page is a recommendation to buy or sell anything.
Backtests use month-end prices and assume every contribution lands on the same day each month. Index levels are price returns unless the entry names an ETF, so dividends are excluded where the underlying data excludes them. Entries are ranked on money-weighted return over a plain monthly investment in the same asset. One entry per account per asset counts towards a rank.