How should we handling imprecise warmup period results?

[DaveSkender]

Some indicators, like ADX, EMA, or others that use smoothing techniques require a runway of about 150-250 periods before the algorithm produces better precision. Early periods during this initialization window can have an error variance of 5% or more. Indicators that require a runway have this mentioned in the documentation + history thresholds were implemented to warn users; however, we've left the imprecise values in the results.

Question: How would you like the library to handle this initialization period?

Contribute ideas and vote for your preferred option.

[DaveSkender]

Do not design in any handling of imprecise values. I do not even want the current history thresholds; keep thresholds at the minimum required to calculate, but do not include consideration for the initialization runway. NOTE: users may not be aware that they are using imprecise results with this option.

[DaveSkender]

Use history thresholds only to require enough history to cover the initialization runway, but keep the imprecise results and let me handle it on my own. NOTE: users may not be aware that they are using imprecise results with this option. This is the current state.

[DaveSkender]

Exclude imprecise results based on a general analysis of where there is more precision. Use a history threshold exception to alert users when insufficient history is provided. NOTE: This will help users from having to determine if results are good and bad; however, it may exclude imprecise results that may be "good enough" for some use cases.

[DaveSkender]

"Exclude" in this context would mean to provide a null period result instead of the imprecise value. We'd still return the same size result set as the history provided, for consistency.

[DaveSkender]

User directed pruning to make this a user option by providing an extension method that will either prune the recommended warmup periods or allow users to specify the number of initial periods to prune.

// normal, without pruning
var results = history.GetEma(20);

// with recommended pruning
var results = history.GetEma(20)
  .PruneWarmupPeriods();

// with specific pruning
var results = history.GetEma(20)
  .PruneWarmupPeriods(150);