I use CSI for stocks and ETFs and yes you can get adjusted prices for splits, dividends, and other corporate actions. The only issue is that you can't have both (splits and dividends) in the price series along with adjusted prices without paying extra. Not a big deal if you are doing your own adjustments but it is a bit irritating.
Folks - any idea where I can cheaply purchase a time series (say quarterly or even annual) of index level PE ratios for the major equity markets that we trade?
@globalarbtrader Hi Rob, In AFTS Strategy 2 and strategy 3, you use the actual price of the contract to calculate position size and the back-adjusted price to calculate volatility. This is fine, but while calculating performance metrics, the returns on roll days would be incorrect. So you ignore the roll issue for performance metrics for strategy 2 & 3? Thanks!
I don't know about PE, but if you're fine with Shiller's CAPE you can get historical data for major countries for free.
I found a small configuration error today: The roll offset of micro gold and silver at IB should be -35 instead of -30. I currently own gold contracts and IB informed me that they will be closed out today due to their physical delivery policy. My system planned on rolling tomorrow.
"you use the actual price of the contract to calculate position size and the back-adjusted price to calculate volatility." actually I do this with everything, not just strats 2&3. It is the correct way to do it. This is fine, but while calculating performance metrics, the returns on roll days would be incorrect.", well no, because you use the back adjusted price to calculate p&l Rob
I have really enjoyed reading some of your books, so I took the time to create a user, just to provide a reply. I've been looking into this earlier. You're correct in Andreas Clenow using Norgate. At least he mentions it in his book (and the data he provides also confirms it). You can easily extract historic data from Norgate using python (I'd be able to provide you with a script for this, if you want). They provide historic constituents and the eod data for both current and previous constituents (also those that are now delisted). The only caveat I found is that they provide adjusted close, but no data on splits. Regarding the Windows issue. I ran their software via VMWare on my linux box, without to much fuzz. A good alternative, that I've heard good things about is the Sharadar dataset: https://data.nasdaq.com/databases/SFA Regarding https://github.com/fja05680/sp500 Be aware that you'll spend a lot of time mapping the tickers of inactive (and potentially delisted) companies to the correct data, given that tickers may be reused. You'd also need the data of the delisted tickers.
Out of curiosity I played around with different volatility models for volatility targeting and all the other places where volatility is an input: Neither Ornstein-Uhlenbeck nor eGarch make much of a difference. All stats we care about are more or less the same. But they come at quite a hefty price computation wise. I guess mixed ewma squeezes all the information out of backward looking data that is there to have.