Choosing Equity Curve Granularity and Download Formats

Summary

The document raises design questions for a web chart showing the equity performance of foreign-exchange trading strategies. It asks whether equity should be recorded on every market tick or at a coarser interval, noting that frequent snapshots can create large datasets and a noisy display. It also considers whether download size could be reduced by smoothing or simplifying the plotted line instead of sending many observations.

The sole answer emphasizes that the best export format depends on who will use the data. Assuming the audience consists mainly of people, rather than software consuming the data, it suggests CSV as a reasonable format. The response does not settle the sampling frequency, describe a smoothing method, or compare transfer and fidelity trade-offs. It offers a limited format recommendation, while leaving the central questions of chart granularity and data reduction open.

Key ideas

• Recording equity on every tick can create large datasets and a noisy chart.
• The appropriate observation frequency depends on the chart’s users and purpose.
• The answer suggests CSV as a reasonable download format when people are the main audience.
• The document does not recommend a specific sampling interval or explain how to simplify a curve.

Tags

# Equity Chart - design and granularity


# Equity Chart - design and granularity












I am looking to build a web based Equity chart to display performance of FX trading strategies.

I would like to hear opinions and advice on a few areas that I am unsure about.

Granularity

Equity can - and typically does - change every tick. Should I therefore save equity every tick? If I do I am likely to be saving a lot of data! And then the display of this data will also be a challenge - as there would be a lot of noise.

If I am to save a snapshot every moment in time, what would be a recommend timeframe? Every minute?

Optimizing for download

As the amount of data in the equity chart could be quite large, what are some recommend approaches to optimize for download? Would it be advisable to somehow smooth the equity curve and download just a vector line rather than downloading a csv/json with many thousands of datapoints?

Thanks for any feedback - its really appreciated.

## Answer by madilyn (score 1)

https://quant.stackexchange.com/a/15201

Granularity





Optimizing for download

It depends on your users! Without any further information, my best guess is that most of your users are human users and not machines/programs, so CSV is probably the most reasonable.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.