Hi, I realized that the API credits consumed while using the tool “Download factors” is not realistic, and the final consumption is greater than announced at start. I am not sure, but it looks like the generation of the file at the end of the process is consuming new API credits.
The estimate is based on the current universe size. If the universe counts vary over time and were higher in the early years, then the estimate can by too low.
I just did a test where I use the Russell1000 universe which has consistent counts over the 20 year period. The estimate was 155 credits and the actual usage was 156.
Then I did the same test with a smallcap universe which has considerably more stocks in the early years. The estimate was 181 credits and the actual usage was 227.
Do the counts in your universe vary over time? If you want to check that, just run a Screen backtest over the same period as you data download period. Use your universe and no rules in the Screen. The chart with #Pos is the universe counts over time.
