New feature allows you to create your own "top-down" series

Dear All,

A new powerful tool is now live that lets you create a custom series from universe aggregates. With this tool you can create breadth indicators, re-create our SP500 earnings series, or focus in on certain sector trends.

At the moment the only other place where you will be able to use these series is in the MARKET->MACRO CHARTS , but soon you will be able to reference them from systems rules.

In short…

  • The tool looks like a screen but instead of screening each rule returns a value
  • The series is created from the value of the very last rule

Try clicking on “Totals” in some of the pre-built examples.

The other main difference is that the tool has two references. One labelled “Series” is for the aggregate functions that are available like UnivCnt, UnivSum, etc. The other labelled “Stock” is for what you can use inside the parameter “criteria” for these universe functions (ETF based custom series will be coming soon).

For example to count the number of stocks in a universe trading above the 200 day moving average you enter:

UnivCnt(“Close(0)>SMA(200)”)

Please see our description in the Help here . Documentation is skimpy at the moment. We will add more in the next weeks. Hopefully it is intuitive enough to get something done

Cheers

Wow Marco! The mind is already running wild with applications.

This will be incredibly powerful. “Wow” is the only thing I can think of at the moment.

Marco

Will it be possible to plot / reference the equity curve of a Sim / Port / Book?

Okay…so I got using this and am now late for an appointment. All I can say is that this might be the most powerful addition to P123 in the last 5 years. Its mind-blowing really…and fast! You needed an entire morning with ClariFI to run the same tests you can do here in 10 minutes. Seriously…I don’t know what major investors are watching this - but ClariFI got bought out by CIQ for $300 million… and it is a lumbering bloated pig compared to what is being built here…

Now I am really late and have to run. But WOW!

Fantastic!!! a really great addition.

Could you add MEDIAN as well please? It would make it easier to avoid extreme values affecting the averages, when plotting aggregate ratios.

Thanks a lot for this addition!

series, tools

I chose universe nasdaq 100 at the top

added 1 rule

UnivCnt(“Close(0)>SMA(10)”)

date 8/7/14

hit total and got 7063; doing something wrong?

Calif, you didn’t save the series. PLease save it so we can check. Thanks

I retried and it works now, thanks.

Marco,

This is a great new feature.

There seem to be errors in the data for P/B, P/E EBIT/EV. See the public chart from Rellis labeled “Price to Book for the S&P500”. The big steps don’t make sense.

Rick

Marco,

Will we be able to use the new functions UnivAvg(), UnivCnt(), UnivSum(), etc. as buy and sell rules in Sims & Ports like:

100 * (UnivCnt(“close(0)>sma(200)”)/UnivCnt(“true”))>50

I tried it but it doesn’t work now. I got the error: [color=red]Error near ‘UnivCnt’: Invalid command ‘UnivCnt’[/color]

Denny :sunglasses:

I’m getting the same error as Denny

Relllis,

WYNN’s Pr2BookQ on 11/18 went from N/A to 17K+ which throws off the average. I think that is the correct value for that ratio since their BookVal was close to 0.

We should probably add UnivMedian to make this type of series meaningful.

Or you can add your own sanity check to eliminate from the average the outliers. For example in “criteria” for UnivAvg use Pr2BookQ<20

Yes, please add UnivMedian, it would be very helpful in many cases like this one.

Very cool Marco!

I will love the see a contributing factor (weight) that can be negative(!) for each component.

An optimisation mechanism will be cool too.

Marco & All:

Is it possible to create a weekly series from daily data? If not, would you consider adding this functionality in the future. Weekly data (with its inherent filtering of daily fluxuations) is sometimes valuable. Thanks.

Bill

dmicsa, you referring to the series tool? what weights?

strader1, maybe. Inside the “formula” you have free reign. You have a specific example?

Can we please get more than 15 years of data for the series tool. For example, % of SP500 stocks > 200 day average. It is very limited with only 15 years/2 recessions. I would like to test some market timing strategies with a longer period of time.

thanks, Debbie

Marco & All:

 Yes, I would like to be able to calculate / use Rsi(x, y, #Series) based on weekly data.  Can I use "top-down" series to do this for say SPX? If so how?  Thanks.

Bill

Marco, now you sum with factor 1 on all oscillators. It will be nice to can alter the weight of any oscillator we use. Some oscillators works well reversed and are not symmetrical.

A parameter defaulted to one on all oscillators will make it behave like now.

I know we can duplicate oscillators and they will contribute with integer multiple but still the one that contributes in a negative way are impossible to quantify.

The overal oscillator:

OverallOscilator = Sum(Wi * Vi) / Sum(Wi); where Vi = {0, 1}

now is done like this:

OverallOscilator = Sum(Vi) / n; where Vi = {0, 1}

Marco & Team,

It looks like Christmas is coming early this year. Thank you for this extraordinarily valuable upgrade to the platform. These new functions save so much time over the alternative (labouriously copy-pasting into a sheet / running series operations in sheet) that I am just about ready to shed a tear…

Best regards and keep up the great work,

D