The revisions presented last week are on track for implementation this weekend, and that means that some screens that utilize volume related fields or the standard deviation field (StDv) will require some changes. We’ve received some positive feedback on the upgrade path, and we are working hard to ensure a smooth transition.
If you view a screen which you have flagged as a ‘Favorite’ you might now see a notice regarding a necessary screen revision. If you click the link, you can review the proposed changes and have the opportunity to inform us of potential problems before implementation this weekend. Obviously, we would suggest you take a look if you rely on the screen. If you do not see any type of notice this means that either the screen utilizes none of the affected fields, or you only recently flagged it as a ‘Favorite’. If the screen uses any of the affected fields, you should see some type of notice.
If you study the revisions you will see come commonality:
1. Comparisons of one volume field to another require no adjustment because the revisions will affect both sides of the equations the same. ex. [Volume] > [Vol-5d].
2. Comparisons to numeric values are converted to use the letter suffixes. ex. [$vol-21] > 2000 becomes [$vol-21] > 2m.
3. Fields used in equations or uv’s are adjusted in a default manner that divides by 1000. This has the effect of converting the number to what it is now before the changes take place. We obviously could have made this default change everywhere, but changes made in this manner are harder to read, therefore we tried to minimize their use. ex. [Volume] > 1.25 * [Vol-5d] becomes ([Volume]/1000) > 1.25 * ([Vol-5d]/1000).
Again, please let us know if you see any issues as we want the transition to be as smooth as possible. At this point we are planning to make this transition on Saturday, Jan 24.
Thanks for your continued support of our ETFScreen.com website.
I love these changes. Thank you.
Now that you have REITS and CEFS you might try and get Yield!
That is a big part of this world.
Best, RG
Thanks for the comments, and we hope to have yield statistics soon.
If you have other suggestions please let us know.
The site is now updating data every 10 minutes? Nice!
I’m glad you noticed and commented on it. I think the more frequent updates are appreciated by many of our regular users including myself. We all are accustomed to instantaneous data these days and the 30 minute updates were far from instantaneous. We are still utilizing 15 minute delayed data, so with 10 minute updates our data ranges from 15 to 25 minutes after actual market time. We still have room to improve if we can work out the economics. Thanks again. – Hugh
Regarding data, there is a good chunk of VQT ETF price data missing in all the screens, from 5/17/2011 to 10/4/2011. Can you fix this, please?
Unfortunately what you are seeing is the result of low trading volume during that time period when VQT was not trading often enough to generate valid data. When there are numerous days with no trades then the price stays static but the market is continuing to move. Then, when the price does move it often catches up with the market in a one day move that is not at all reflective of anything related to that symbol on that date. Our policy is to exclude such symbols from our screener, but we recognize that such an exclusion creates it’s own set of issues.
Hello Hugh,
I would like to be able to weight the choices in the backtesting screen inversely according to volatility. For example, if I choose 5 ETFs I do not want to weight them equally, but inversely according to volatility and then rebalance every month.
I know it is not possible yet. When will it be possible?
Thanks for the suggestion. We would like to support multiple weighting strategies and have several in mind. Do you have a particular way of weighting by volatility you are looking at, or just looking for some general method that puts more weight on the first symbol and subsequently less on the rest? One other option could be to specify the weights by position. How do you see this working?
Right now, I was thinking of weighting in inverse proportion of volatility.
For example, if I have 3 ETFs and the backtest for one month says the number one has a daily percentage change of 1%, and the second and third 2% each. In that case, we would put twice as much money in the first one as in each of the other two (let’s say we have 10000$, we put 5000 in number one and 2500 in each of the others).
The rationale is that every percentage movement would risk the same amount of money.
Of course, if we could choose among several weighting methods in the backtest would be great, including the ones you suggest.
Thanks. let me put some thought into it and we’ll see what we can do.