Optionize

Approach

Nothing is traded that has not been measured.

Measurement first

Measured, not assumed

The ordering of the three systems is the whole argument. Recording exists so that evaluation has something truthful to run against, and evaluation exists so that execution is given parameters that came from measurement rather than from judgement.

It is easy to build a strategy that describes the past. It is harder to build one whose parameters were chosen without knowing the answer, and harder still to keep the discipline of discarding the ones that do not survive that test. The apparatus here exists to make that discipline cheap enough to actually apply.

This is why the site reports no performance. A number without the method that produced it is not evidence, and the method is the part worth describing.

Execution cost

In short-dated options, cost decides viability

Short-dated options are traded in small premiums. The spread a position crosses on the way in and out is not a rounding error against the position, it is frequently comparable to the edge the position was opened to capture.

A strategy evaluated at the midpoint of the spread is therefore not being evaluated at all. The question is not whether the logic would have been right, it is whether the logic would have been right after paying to express it — twice, once to open and once to close.

That is why execution cost is carried inside the replay rather than deducted from a result afterwards. A cost applied at the end can be tuned; a cost applied at the point of each simulated fill cannot.

Risk posture

Capped downside, no leverage claims

Risk is defined structurally rather than managed reactively. Positions are opened with a known worst case, which means the downside of a position is a property of the structure chosen rather than of a stop being reached in time.

There is no naked short exposure. This is a constraint on what may be traded at all, not a target to be relaxed when conditions look favourable.

The conservative posture is deliberate and it has a cost: it excludes strategies whose expected return depends on accepting undefined downside. That trade is made knowingly.