It is my understanding that all prices in the P123 database are split adjusted.
Is there any way to access point-in-time prices, i.e. non-split adjusted before any split adjustment has been applied?
Thank you.
I suspect you can use FHist() with close(), I am not sure though.
The “price” command will give you unadjusted prices. So will FHist, as you suggest. So will changing the date of observation. For example, in the screener, you can change the “as of” date. In backtests, the prices are always “as of” the date of ranking, unadjusted for “future” splits and dividends, no matter what command you use. In the “issues” tab for any stock, you can access unadjusted historical prices by toggling the “adjustment” box over the chart.
In a ranking I have been using a Stock Factor simply called “PRICE” as a node.
When backtesting using this ranking node, am I getting the correct point-in-time prices (the correct way to backtest) or are the prices affected by splits?
Yes, you’re getting the correct point-in-time prices if you’re using the “price” command, as I said. But you do realize that you’re ranking stocks on the basis of their price, right? I’m not sure why you’d want to do that. Is a stock priced at $50/share more or less likely to do better than a stock priced at $3/share?
The backtest tells the story.
I may be misunderstanding something, but FHist seems to be returning split adjusted prices.
Take EDU as an example. It had a 1/10 split on 4/8/22. Using the screener w/ today’s date (5/1/22);
ShowVar(@Close0,Close(0)) = 12.5500
These prices should be prior to the split date and they look split adjusted. I’m I reading it wrong?
ShowVar(@Close25,Close(25)) = 13.3000
ShowVar(@HClose5W,FHist(“Close(0)”,5)) = 12.6000
ShowVar(@HPrice5W,FHist(“Price”,5)) = 12.6000
Walter
Not sure what the question is. Everything on a past date is shown as it was then. You don’t really need to think about it. Price > 3 or Close(0) > 3 is exactly what it says: filter for stocks trading above $3 on whatever date you are looking at.
What for? can you elaborate on the use case?
Do not use FHist for anything other that ratios (like PE, PS, etc) unless you understand what’s going on. The documentation says not to use for any split sensitive data (ratios are not affected by splits bc they are ratios) . This is why: our data internally, prices and financials like EPS, is fully adjusted for splits and dividends for efficiency. We only reverse out future splits and dividends only when necessary, for example if you want the price with an as-of date in the past like in the example I gave above. Fhist does not do any adjustments after the value is computed in the past because it does not know what is computing. This does not matter for a PE in the past, but it matters for the example that Walter gave.