Portfolio123-com-s-stock-ranking-systems-for-subscribers

Add your vote for your own request.

For example, here are the characteristics of one of my strategies in Portfolio123 using their ranking systems


And this is the result of live trading in 2024 using this strategy


But I miss their capabilities in terms of technical indicators and other things that are in WL

I don't understand the use case for this. Portfolio123 already has a fairly sophisticated strategy development environment and they have links to 1 or more brokerages that enable you to trade the results of your strategies..
What would this enhancement enable us to do if we did have access to Portfolio123 platform?

If the request is to add access to all of their point-in-time fundametal data, that would be awesome. What is the cost to have access to all of their fundamental data via API?

QUOTE:
I don't understand the use case for this.

I don't either. How a stock should rank for your strategy is a function of your strategy. And they don't know what my strategies look like.

Now if they have some kind of ScoreCard criteria (or plug-in) I can install in WL to rank my stocks with my strategies, I would be interested. But I don't think that's what they are selling.

I "think" they are selling a ranking service that works with their strategies, not mine. So how can that ranking service be applied to my WL strategies? Will their ranking service even be compatible with my strategies?

QUOTE:
If the request is to add access to all of their point-in-time fundamental data, that would be awesome.

I'm not sure I would call it "fundamental data". I would prefer to call it "event/economic timing" data. Timing the trade is what matters most.

QUOTE:
I don't understand the use case for this.

I too couldn't understand. This feature request could've been more developed.

I would use it in blocks like this for example
- Simbol is Ranked Top 20% by P123 Ranking Systems(Basic: Sentiment, Stock Universe - S&P500 LargeCap) or
P123 Ranking Systems(Basic: Sentiment, Stock Universe - S&P500 LargeCap) is greater than 80

When building systems, P123 lacks many of the capabilities that WL has. For example, there are no stops based on ATR, no limit take profits, and many technical indicators are missing. Therefore, it would be great to combine the capabilities of P123 and WL.

QUOTE:
Symbol is Ranked Top 20% by P123 Ranking Systems(Basic: Sentiment, Stock Universe - S&P500 LargeCap)

But you can do that now with WL. The only thing missing is "Basic: Sentiment". So is this feature request about bringing a "Basic: Sentiment" indicator into WL? (... which you can then use as a block.) Exactly what's in Basic: Sentiment?

There are already several sentiment data streams available in WL. Are you using them today? Are you using IndexLab as well?

I think what's really missing in WL are integrated strategy examples that combine sentiment data (including the timing part) with IndexLab data, and price-and-volume action. What we have now are baby examples that do a little of this, and a little of that, but nothing is really integrated together in harmony.

And I can build integrated strategies with WL using C# (It takes some work.), but doing so with blocks sounds scary. If you can't code in C# (Why can't you?), then you really need a partner that can (which is a good solution) for developing integrated strategies.

---
Portfolio123 employs high-level blocks to build strategies. I suppose WL could develop some high-level blocks to do the same thing. (So you're suggesting there's a market for this.) The down side is that you have little fine-level control. But if you don't want to code in C#, then you probably don't care about fine-level control.

WL could also develop some meta-indicators that have non-static member methods that do fancy things (like multi-variant regression and robust thresholding). It's a high-level evolution of the C# indicator classes.

This is what it consists of Core: Sentiment-
100%Core: Sentiment
-45%Estimate Revision
--33.33%EPS Estimate Revision CYƒ(x)
--33.33%EPS Estimate Revision CQƒ(x)
--33.33%EPS Estimate Variability CQƒ(x)
-25%Surprise
--65%Surprise%Q1
--35%Surprise%Q2
-15%Recommendations
--75%Changes in Recommendationƒ(x)
--25%AvgRec
-15%Short Interest
--100%Short Interestƒ(x)

This ranking system is similar to the ranking system from Zacks
But I gave Sentiment as an example, there are many other good ranking systems, for example Small and Micro Cap Focus -
100%Small and Micro Cap Focus
4%consistent EPS growthƒ(x)
2%EPS current quarter growthƒ(x)
4%EPSExclXorGr%PYQ
3%FCF%AssetsQ
2%FCFGr%TTM
3%FCF PEGƒ(x)
4%forward earnings yieldƒ(x)
4%forward EBITDA/EVƒ(x)
3%forward revenue yieldƒ(x)
4%gross plant to sales compared to industry medianƒ(x)
6%GrossProfit%AssetsA
3%industry RSIƒ(x)
2%inventory changeƒ(x)
3%median ROEƒ(x)
8%MktCap
6%OpIncGr%PYQ
1%PEGST near 1ƒ(x)
1%ROA%Q
6%SalesGr%PYQ
6%Sentiment: P123 basic
4%share turnover, 1 yrƒ(x)
7%share turnover, 3 monthsƒ(x)
2%subindustry momentumƒ(x)
3%unlevered free cash flow to EVƒ(x)
2%velocityƒ(x)
4%volumeƒ(x)
3%volume increaseƒ(x)

The question is, is it possible to consume this sentiment score on a historical basis for a stock from their web site/api?

Sounds like there's a startup opportunity for building high-level sentiment indicators using WL tools and data. Someone should get started.

The same goes for ranking potential stock performance. Is someone here working on this? Perhaps one of the IndexLab users is already doing this. Please step forward. (How many WL users are employing IndexLab now? Maybe we need a video to generate more IndexLab interest.)

---
Back in the Fidelity WL Pro days I built a strategy that compared industry performance (Fidelity subscribes to Standard and Pour's sector and industry historical price data) for the stock I was trading. Surprisingly, it wasn't all that effective. A given stock will follow its industry price stream, but it rarely outperforms it.

---"The question is, is it possible to consume this sentiment score on a historical basis for a stock from their web site/api?"--
https://www.portfolio123.com/research/dataminer-api

For those of you unfamiliar with IndexLab, you could take the breakdown shown in Post #9 and create a "custom index" that would perform the same ranking. But there are two problems:

1) Some of the time series mentioned in Post #9 are not easily obtainable. For example, Industry RSI. To get that, one would need the Standard & Poor time series for that particular industry, which you are probably not subscribed to. Of course, you could substitute an equivalent ETF for that particular industry instead.

2) The weighting breakdown they are employing in Post #9 is right for their strategy trading against their dataset. But it's not really right for yours. So you have a multi-variant regression problem to rebalance all the weights so they fit your specific strategy and dataset. If you want to skip the statistical analysis on the regression terms, the Math.NET numerical analysis package can fit it. But if you want to know which regression terms are most significant (i.e. compute Ps with contrasting F-tests), then you need to run it on your stat package employing stepwise regression like a macro economist would do.

Of course, employing a poorly balanced ranking system is better than not having one at all. We all know that. :-)

---
Somebody is going to ask, so let's address the question now. Could one take all the time series mentioned in Post #9 and feed it into NeuroLab so it would return a ranking indicator? Maybe. Let me know if you get that working. :-)

I thought that IndexLab can make an "index" (for example, sentiment) only for some "Universe", and not for a separate stock relative to this universe. And the ranking systems in P123 give a rating of a separate stock within the universe.

Agreed. But some of the time series in Post #9 will be the same for most stocks. You can make a common index for those using IndexLab, then combine it with the stock-dependent time series cases to get your specific stock rankings. Not a problem.

In WL, you might be better off keeping the common indexes separate from the stock-dependent ones and just ANDing these conditions (rankings) together in the final test. Ideally, you want the common conditions AND the stock-specific conditions to both be met before you pull the trigger.

So what about the question raised in Post #10? Will the Portfolio123 API provide historical sentiment data that can be backtested with WL? Or is this feature request just "hypothetical"?

Yes, it does. It is a bit tricky, but it works. I work on an interesting Solution.

QUOTE:
It is a bit tricky, but it works.

So are your strategies more productive now with their sentiment data, or do you know yet?

Is it really worth subscribing to their service just for the sentiment data?

Somehow I think their other data feeds (ranking systems) would be too dependent on one's strategy and dataset to be useful.

QUOTE:
So are your strategies more productive now

I don't know yet. Need a few more days to get this fully working.

QUOTE:
Somehow I think their other data feeds (ranking systems) would be too dependent on one's strategy and dataset to be useful.

It is possible to send series from WL to P123. The plan is to send some strategy results to P123 and use these in advanced rankings on their side.
Then us the ranking results from P123 to create dynamic datasets on WL's side.

I just finished the beta version of a new Portfolio23 extension.

See this post https://wealth-lab.com/Discussion/WANTED-Beta-Tester-for-Portfolio123-extension-12272 for details.