- ago
I just finished the Beta version of the new Portfolio123 extension.

From the help File:
The Portfolio123 extension creates a two-way connection between Wealth-Lab and Portfolio123(https://www.portfolio123.com/index.jsp?apc=DrKoch).

This extension supports the following actions:
* Download historical ranking results that were calculated with a Portfolio123 Ranking System on a specific Universe with a selectable rebalance frequency.
* Create dynamic DataSets from a downloaded ranking, for example the top 50 stocks in each quarter.
* Access fundamental data (contained in the downloaded ranking)
* Create an index indicator in Wealth-Lab and send it as a *Data Series* to Portfolio123
* Calculate an indicator for all symbols in a DataSet and send it as a Stock Factor to Portfolio123

Requirements
* A subscription / membership at Portfolio123.com. (Starts at $25 for first month)
* A historical data provider like Norgate Data(https://norgatedata.com/) that can provide historical prices for the tickers used in your Portfolio123 rankings.

Beta testers write a mail to: rene dot koch at finantic dot de

You'll receive a zip file with the extension.

The extension will run for free until it is finally published.

A beta tester is asked to report:
* experiences
* problems/bugs
* suggestions, comments.
3
640
25 Replies

Reply

Bookmark

Sort
- ago
#1
S&P500 LargeCap (IVV) is missing when selecting Universe.
0
- ago
#2
The combobox for universes shows just a few examples. (The list from here: https://www.portfolio123.com/app/opener/UNIV?cat=-2).

The textbox is writable, you may enter the name of any (existing) Portfolio123 universe.
1
- ago
#3
Also, you may enter your favourite universes in the file

<WL-Installation-Folder>/Portfolio123/KnownUniverses.xml

The universes-ComboBox gets its contents from that file.

(please send this file back to me after you entered interesting universes)
1
- ago
#4
Can universes be made of their own screens, i.e. contain several filters, including those based on different ranking systems? Or can only standard Universes be used? I don't want to waste quotas for experiments once again))
0
- ago
#5
As far as I understand things:

The connection between Portfolio123 and WealthLab (which is restricted by the available API methods) works as follows:

1.) Create a Universe of stock tickers (or ETF tickers). It is possible to use any combination of Fundamentals and Rules to build such a Universe.

2.) Create a Ranking System. It is possible to use any combination of Fundamentals or other Stock Factors and Data Series to define such a ranking.

3.) The extension downloads ticker and ranking information for the complete (ranked) universe. (Once for each Ranking/universe combination)

4.) The extension constructs dynamic/rebalanced DataSets from the Universe based on the ranking information. (It is possible to build many different DataSets from one ranking/universe)

Example: With a Rebalance Frequency of "One Quarter/Three Months" such a DataSet could contain
the Top 50 Stocks in every quarter for the last 10 years.

It is possible to run a "Buy And Hold" strategy on such a dynamic DataSet.
The result should be very close to a Screener Backtest on Portfolio123.

(But: A Screen/Screener is a different animal and not usable with the Portfolio123 extension.)
1
- ago
#6
QUOTE:
2.) Create a Ranking System. It is possible to use any combination of Fundamentals or other Stock Factors and Data Series to define such a ranking.

So you're saying I can supply my favorite stocks in step 1) and the Portfolio123 server will figure out a ranking system to find more of the same in step 2)? What do I have to give it in step 2) so it can work this magic?
0
- ago
#7
QUOTE:
So you're saying...

Let me clarify:

Step one is done on Portfolio123.com: Select a starting universe, for example all US Stocks with enough liquidity and a price above $3. You might apply further filters/conditions to create a nice large selection of stocks (or ETFs) called a "Universe".

Step two is also done on Portfolio123.com: You choose some fundamental data and conditions and other "Stock Factors" to build a multi-factor, hierarchical ranking system.

Both steps together result in a "Ranked Univers" and this thing (a large set of tickers, fundamentals and ranks) can be downloaded to Wealth-Lab.

On the Wealth-Lab side of things it is now possible to create a dynamic DataSet based on the "Ranked Universe" which contains something like the Top 100 Stocks rebalanced every quarter (or every month). With all fundamentals that took part in the ranking formula also available as "Event Data" and/or fundamental indicators.

And here starts the interesting story:
Will a trading strategy benefit from such a DataSet?
What fundamentals will improve our strategies?
0
- ago
#8
Q: Is there a free trial period for a Portfolio123 membership that lets me test this new extension?

A: The minimum required membership costs $25 for the first month (Screener membership with 5 years of backtest data).

Q: Is there a special price for Wealth-Lab users?

A: Not yet. But the Portfolio123 people promised to lower the price by 15% if enough people (15 or more) buy some paid membership through this link: https://www.portfolio123.com/index.jsp?apc=DrKoch

This link also enables a 35 days trial (instead of 25 days) but such a trial allows no API access, so the Portfoilo1213 extension can't be used with such a trial.
0
- ago
#9
DrKoch, thank you very much for making this extension.
I have the first results. In P123 I have a subscription "Screener". In P123 I created a universe that gives the following result when tested:

In WL I made a strategy based on this universe



I made the strategy quickly and will continue to refine it.
0
Glitch8
 ( 5.77% )
- ago
#10
DrKoch, looks like you’re in contact with P123? Any word on them promoting your excellent extension to their customers?
0
- ago
#11
QUOTE:
I made the strategy quickly


An APR of 37% with a MaxxDD of -16% is just crazy.

Are you sure you didn't heavily over-optimize?
0
- ago
#12
5 years is of course not enough for testing, so it is possible that there is some curve fitting here
0
- ago
#13
DrKoch, I sent you the "Universe" and the ranking system to your email. If you have a subscription for more than 5 years, you can check it out.
0
- ago
#14
DrKoch, can you make it so that it would be possible to make a dataset not only from Universes but also from Screens? Screens can have filters like "Rating("Core: Value")>20", but Universes cannot have such filters. This would greatly improve the capabilities of this extension.
0
- ago
#15
QUOTE:
... but also from Screens?


Sorry, but no. The Portfolio123 API allows the download of historical ranking data only.
There is no download for Screens.

Try to reach your goals with conditions for a Universe plus factors for your Ranking.

Please note: All factors in a ranking are also available on the Wealth-Lab side. So it would be possible to add logic to a Wealth-Lab strategy that uses these factors.
0
ww58
- ago
#16
Has anyone gotten such good results like dmitry7?
0
- ago
#17
QUOTE:
Has anyone gotten such good results

Here is a very quick proof of concept:
I took the "Knife Juggler" strategy from WL's Sample Strategies with its original settings.
(Limit 2%, Timeout 2 bars, PT 5%, no optimization or simliar.)

I ran it for the last 10 years on three different Portfolios/DataSets:
* S&P 100 (WealthData)
* S&P 500 (WeathData)
* a Portfolio123 DataSet called All-Stars: Graham, Top 30 symbols, quarterly rebalanced

Here are the resulting equity curves:

(This is the Compare Tool, part of finantic.ScoreCard extension)
1
- ago
#18
Here the performance metrics of these three runs:


I adjusted position sizes to arrive at roughly the same MaxDrawDown.

I think these results speak for themself.

There was no optimization involved. All settings were "invented" a long time ago.
0
- ago
#19
Conclusion:
1.) Choice of portfolio/DataSet plays a significant role for performance
2.) Portfolio123 delivers excellent possibilities for selecting a profitable portfolio
3.) The new Portfolio123 extension makes these advantages available for Wealth-lab users

My Personal Summary: The combination of a ranked and rebalanced Portfolio (aka dynamic WL DataSet) and a decent Trading Strategy comes close to the Holy Grail.

2
ww58
- ago
#20
>comes close to the Holy Grail
It seems so. On one hand, it's too good to be true, on the other, it all makes sense-we combine fundamental and technical analysis to gain the advantages of both. I’ll definitely give it a try. As I understand, since dataset is dynamic it’s better to use the "Exit orphan positions at Market" option here, right?

An idea for Glitch's video series: to test a metastrategy on a such dataset.

@DrKoch, does the extension cache the downloaded datasets to prevent credits from being spent again?
0
Glitch8
 ( 5.77% )
- ago
#21
Good idea, bookmarked!
1
- ago
#22
QUOTE:
does the extension cache the downloaded datasets to prevent credits from being spent again?


Yes. The extension loads a complete historical ranking (ranking system + universe, 10 years back with a backtest membership, selectable rebalance frequency) just once and stores it on the local disk.

From this ranking it is possible to create an unlimited number of dynamic datasets with:
* Number of symbols (example: Top30)
* The ranking metric (any of the fundamentals and stock factors used in the ranking)
* the ranking direction (Best / Worst)

If you have a quarterly rebalanced ranking, and after three months it is time for an update, this will cost just two credits, because the delta (last quarter) is downloaded only.
1
- ago
#23
QUOTE:
Conclusion:
1.) Choice of portfolio/DataSet plays a significant role for performance

This is absolutely true. I always cherry pick my datasets for this reason using the Symbol Rankings tool. For ranking symbols, my favorite ScoreCard metrics are SharpeRatio, WinRate, ATR, and ProfitPerBar.

I have not found an ideal ScoreCard metric that takes everything into account. And I have devised my own ScoreCard metrics in an attempt to do so with marginal success. I'm afraid it's going to take a stat analysis to find the best composite ScoreCard metric to get this ranking business working right. :(

By the way, I'm a trader, not an investor. So what the fundamentals are doing isn't that relevant for my trading. I do check to see that the fundamentals aren't bad (trailing PEG ratio < 2) before pulling the trigger, but that's all. I think Jim Cramer is right; don't try to turn an investment into a trade, and vice versa--they are different.
0
ww58
- ago
#24
QUOTE:
By the way, I'm a trader, not an investor. So what the fundamentals are doing isn't that relevant for my trading
I disagree, if you take any dip byer system it is very important to separate stocks that are sinking, from good companies that are currently undervalued. Of course, if your trades are shorter than a few days, fundamental data does not play a role here.
0
- ago
#25
QUOTE:
if you take any dip byer system it is very important to separate stocks that are sinking, from good companies that are currently undervalued

This would be a concern if you're trading micro-cap stocks--yes that small company is probably sinking. In contrast, if a mega-cap stock dips (Nvidia, Tesla), it's probably political. :(
0

Reply

Bookmark

Sort