Skip to content
TradingTune

Sequential Improvements

Coordinate descent: optimize one parameter at a time.

Quality
Low to Medium
Speed
Moderate
Budget
O(N x k)

How it works

Sequential Improvements changes one thing at a time. It walks through your enabled parameters in turn, gives the current one a fresh value, and leaves the rest where they are.

That makes it easy to follow and cheap to run: every cycle isolates the effect of a single parameter.

The trade-off is that it is blind to interactions. If two parameters only pay off when they move together, tuning them one at a time can miss the joint sweet spot entirely.

Under the hood

This is classic coordinate descent. Cycle i perturbs parameter i modulo the parameter count with a random in-range value while carrying forward the others, so it sweeps each coordinate independently in round-robin.

Because it never varies two coordinates at once, it cannot see coupling. That is the core limitation: the search direction is always axis-aligned.

The extension deliberately never auto-selects it. It is a useful manual probe when you have good reason to believe your parameters are independent, but most realistic strategies have at least some coupling.

Per cycle
1 parameter
Order
Round-robin
Others
Held fixed

When to use it

Reach for it when

  • Strategies whose parameters are genuinely independent
  • A quick, easy-to-read probe of one parameter's effect

Reach for something else when

  • Anything with parameter coupling (it misses joint optima)
  • When you want the best result rather than the simplest search

Want to see optimization in action? Browse our real strategy backtests vs buy and hold on NVDA, TSLA, and AAPL.

New to this? Read how to run your first optimization or learn how to choose the right optimization method for your strategy.