ORRA

Strategies

Every strategy here is a rule that decides how much of a vault sits in ETH, BTC or cash. They use slow signals, because fast trading loses most of its gains to fees. Each one was tested on hourly prices from 2019 to 2026 with a 0.1% fee on every trade, and the vault's own limits applied.

Trend Rider

Holds a token while its price is above its long-term average and steps to cash when it falls below. A dead zone around the average avoids flip-flopping, and the risk dial shrinks positions when markets get wild.

What testing showed. Backtest 2020 to Oct 2026, ETH and BTC, 4-hour checks, 0.1% fee: Sharpe about 1.2 and worst drop 23% (risk dial 0.35), against Sharpe 0.9 and worst drop 79% for holding ETH. It earns less than holding in a runaway rally. Held up at 4x the fees.

Build one

Breakout

Buys when price makes a new high over the last weeks and sells when it breaks below a shorter-term low. Classic trend-following, with long stretches in cash.

What testing showed. Backtest 2020 to Oct 2026, 55/20 days: Sharpe about 1.3, worst drop 28%, in the market 38% of the time. Positive in 2025 while holding lost 6 to 14%. Neighbouring settings (40/20, 55/30) tested nearly as well.

Build one

Core and Trend

Half the vault stays in a steady, rebalanced mix of the tokens. The other half follows the trend rule. Keeps more upside in a long rally than a pure trend agent.

What testing showed. Backtest 2020 to Oct 2026: Sharpe about 1.1 and the highest return of the pool, but a worst drop of 57% in the 2022 bear market (the core half still falls). Best when you want a middle path.

Build one

Rebalanced Mix

Holds fixed shares of ETH and BTC and the rest in cash, and rebalances when they drift. Sells what ran up and buys what lagged. No market timing.

What testing showed. Backtest 2020 to Oct 2026, 50/50: Sharpe 0.96 against 0.80 for the same mix left alone, because rebalancing adds a small premium. It still falls with the market (worst drop 76%). Use a cash share to turn the risk down.

Build one

Easy In

Builds a position gradually instead of all at once: a little more each day until the target share is reached, then holds it. For people who do not want to pick an entry day.

What testing showed. Not a return booster. It reduces the risk of buying right before a drop, and a lump sum usually does better when prices rise. Its record is simply the token's price path.

Build one

Hold a target share

Builds a target share of one token, then leaves it alone. The baseline an active agent should beat.

Build one

Read this before trusting a number