If I am backtesting a daytrading strategy (closes all positions at eod) on an ETF or stock which has undergone a number of reverse splits (ie TVIX), and the data provider I am using doesn't use the actual traded price of the day, will it skew the results? If you take a look at TVIX historical data on Yahoo Finance it reports it trading at over $10,000/share in 2016, which of course wasn't reality. If I'm using fixed dollar position size, ie $50,000/position, then in theory the percent change differences for intraday moves should still be the same? What data provider can I use that data will backtest properly for splits/reverse splits?
You provided an example of properly back-adjusted pricing (from Yahoo) but then you’re asking which data provider does the same and also properly back-adjusts the pricing. So the answer is the same you’ve already given - Yahoo. The $10000+ price is properly back adjusted, so how can you properly adjust again something that’s already properly adjusted? This is the only price you need and any other adjustments would be incorrect.
Unless you’d want an unadjusted data, but then for what purpose? Having the old/original/unadjusted data would be useless and would need to be back-adjusted just like Yahoo did, which is what Yahoo does and provides, and it’s the only way to use such data.
You're right, I'm probably using the terms incorrectly. What I believe I need for accuracy is the actual trading price on the day - the unadjusted data. This article was useful and gets into it: https://decodingmarkets.com/reverse-stock-split-good-or-bad/
The article you provided states the same thing I stated, meaning that you need the adjusted Yahoo data. Your article states: “ Most charting programs show the back-adjusted price which, of course, is how it should be. The back-adjusted price chart is the important one. If you trade off the unadjusted price chart you might make the mistake of thinking a stock has gone up (or down) thousands of percent when in fact it hasn’t actually moved at all. “ So you do need the adjusted data and that’s what Yahoo provides, so you have everything you need. Your article also mentions that unadjusted data can be useful for analyzing past events and understanding why an adjustment happened, but that’s secondary and actually irrelevant in case of TVIX since it reverse-splits simply because the price gets too low to be tradable by serious traders. No one would want to trade TVIX at $0.05 or at $0.0005 per share...
Attached is unadjusted TVIX data from quotemedia.com (probably ok for always exiting by end-of-day). The last column (Adj Close) should be ignored. If this isn't good enough, you can try data sources listed in https://quantpedia.com/links-tools/?category=historical-data
If you had the unadjusted data, and you had some sort of data column that contained the split dates then you could do what you want. For instance IQFeed doesn’t split adjust their intraday day data. The community wanted it that way but believe they provide either the dates for splits or they provide a multiplier to reflect splits as they go back. That method adds a level of complexity any charting software isn’t going to have built in. You’d have to go custom but that’s another complexity. Plenty of software out there to get going. The idea of creating a custom backtester isn’t for everyone, including myself.
Don't forget defunct stocks too. Reverse splits are quite common on failing companies, so you would want to include those stocks that have delisted. Also many companies move to OTC as well. You might not want to trade them whilst they're OTC. I use Norgate Data which provides this capabiulity, but you asked about daytrading (I am assuming some sort of intraday system). Norgate is only daily data, but would be useful to help you find such reverse split events. Also note that sometimes (but not always) the ticker changes to have a "D" at the end for some period of time. I'm not really sure how that mechanism works, and I think it might only be Nasdaq stocks only too that do that.
%% That; + get a 2nd data proVider, eVen if you find one better than YHOO. Also reverse splits are much different form other splits; for example C, reverse split to hide its a penny stock[under $50= under $5] So be sure to moniter you profit + loss + make adjustments from that LOL. IN other words how many ETFs/stocks trade @$10,000 Same thing, somewhat with TZA, but they pay a diVidend...............................................................................................
Hi Kris, our platform can get you that. We are also able to produce output for all stocks that had splits in a given time period