BACKTESTING
How MomentumSwitch backtests Solana trading algorithms
MomentumSwitch uses frozen market datasets so competing algorithm versions can replay the same observations instead of being compared on different market periods. Research aims for apples-to-apples comparisons with no look-ahead and consistent portfolio assumptions.
Complete-system methodology
A meaningful replay includes the decision stack: momentum score, liquidity, buy pressure, verification, confirmations, version-specific entry and exit logic, sizing, maximum positions, allocation and modeled costs.
Why frozen data matters
Freezing the input prevents a later algorithm from accidentally receiving a different market universe. The same starting capital and portfolio rules should be used when algorithms are ranked.
Costs matter
High-turnover momentum strategies can look attractive before fees and slippage. MomentumSwitch research therefore includes explicit modeled transaction friction. Assumptions must remain consistent across compared versions.
Limitations
A backtest is evidence about a historical sample, not a forecast. Short windows can overfit, historical data can have coverage limitations, and real execution can diverge from simulation. Research candidates should be validated on additional out-of-sample data before production promotion.