Kindly help me to understand this flow.

There's a known bug with ratio metrics right now that we're trying to figure out. We use the Delta method to estimate the variance of the metric. In some cases, this estimation process can result in a negative number, which breaks our statistics engine

Hi @future-teacher-7046 I also see even I don't applied the ratio metrics still the metric result start with negative number. Ideally it should start from 0%. Its also the part of same bugs?

That graph shows the results at the end of each day, not the start. I guess we could add an extra data point at the beginning to force it to start at zero.

But the experiment start on same nov 6 so ideally it should be at zero.

What the graph is saying is that after that first day of data on Nov 6 came in, the variation was losing by 30%

Ok! then I think this would be greate

I guess we could add an extra data point at the beginning to force it to start at zero.

I'm not sure what you're referring to. Do you have a screenshot?

@future-teacher-7046 I'll start another thread for above question. But I just want to know till when the bug related to ratio metric will resolved.