Crypto weekly contrarian
The weekly contrarian buys the coins that fell the most over the past week, among the ones that trade the most. Kość, Sakowski and Ślepaczuk published it in Physica A in 2019. Backtoast ran it. This version failed our test, losing 63.4% a year over 4,405 trades.
| Strategy | Trades | A year | Test |
|---|---|---|---|
| Weekly contrarian | 4,405 | −63.4% | Failed |
Backtoast
You pay $99 for two full tests of your own strategy, and a pass is an offer of 30% of the profit Backtoast makes trading it with its own money.
Crypto weekly contrarian
The book is the weakest quarter of a liquid list, equal weight, held for a week and rebuilt every Monday. It's long only, a bet that last week's losers bounce.
The paper sorted a top-100 list by market value. The run kept every pair that averaged at least $10 million a day in trading over two weeks, so the list grew and shrank around 100, and stablecoins stayed in, as the paper's rule was written. It paid Binance's usual fees, which are lower than the paper's 1% a trade.
Crypto cross sectional momentum is the opposite bet, buying the coins that rose most over three weeks.
When it works
Weekly reversal does best in range-bound markets, when coins swing back and forth around a level, and worst in a long fall, when the weakest coins keep falling. A trend filter on bitcoin, so the book sits out the weeks bitcoin is in a downtrend, is the variant worth trying, and a draft can test that before you spend a full test.
Origins
Kość, Sakowski and Ślepaczuk tested May 2014 to October 2017. The contrarian book's yearly return per unit of risk was 2.1, against 2.7 for holding the top 100 in equal weights and 0.2 for momentum. The run is from late 2021 to September 2026.
The run is Backtoast's, from 29 October 2026, after costs.
99% of trading strategies fail. Test yours.
$99 for two full tests. If one passes, we offer you 30% of the profit we make trading it.
Run the test