Question Regarding P123's Privacy Policy

General
1

Hi everyone!
I’m looking at p123’s features and it seems like a good fit for automating the strategies I’ve been trading for quite a few years. Before doing so, I went through the privacy policy and have a few questions.
I’m aware that p123 requires generous permissions in order to place orders, and given the number of years that the platform has been around, I’m not too worried about this aspect. Having said that, my main concern is about the privacy of the strategies implemented on the platform. I know I can make them “private” instead of “public”, but what’s the legal situation with respect to p123’s team having access to them? Is this covered somewhere in the privacy policy?

Thank you very much in advance!
Rick

2

Here is our privacy policy: Privacy Policy. It doesn’t address the specific question you pose, which is addressed in section 8 here: Terms: “It’s the Company’s policy not to access Private Custom Systems internally or make their content available to any third parties, except in cases where access might help in identifying errors.”

3

Thank you very much @yuvaltaylor !
I don’t know how to interpret it correctly since I don’t have a law background. I appreciate that it’s their policy not to access user’s private data, but those terms sound rather vague and non-binding. Am I misinterpreting it?

4

I am not a lawyer, but I believe that terms of use are often recognized as contractual and can be legally enforced.

5

What’s the main concern here? Our developers accessing the system to debug something, or something else? Or is it the wording? Let us know what less vague language you think is appropriate.
Thanks

6

Sorry for the late reply!
Portfolio123 definitely is a reputable platform and has been around for many years, which puts my mind at ease. Still, given the nature of what we do, it would be nice to only provide access with explicit consent on a per-need basis (for bug fixes and that sort of stuff).
My primary concern is the privacy of the proprietary developed systems: rankings, screenings, portfolio rules, etc.

7

Here’s the problem with that. Sometimes we find a problem with a factor that may have caused a bad recommendation or produced the wrong result. We then write a SQL query and find the ids of systems that use that factor, and pull up the chart or look at the transactions looking for anomalies.

Many want us to be proactive, others want more privacy.

To support both privacy policies we would have to create a list of private systems that first need approval to be examined, wait for the approval, remember why we need to examine it, examine it, and let each owner know if there was a problem.

It’s simply not a viable solution. All I can say is there are hundreds of thousands of systems. It’s impossible to know which are real which are curve fitted. The biggest danger IMHO is someone you know that wants your IP.

But we’ll do anything for the right fee :slight_smile:

8

A little more re. this. We’ve heard this request about privacy before and there may be a handful of users that will pay extra for it. Not enough to justify the dev time. So what some users do is finance a project they really want , which we can then try to sell to others.

Like the Chaikin indicators… Marc Chaikin himself called me to add them and provided the algorithm. I forgot for what in exchange. I think we were just happy to oblige. Anyway, he went on to create Chaikin Analytics, eventually bought data directly from S&P and canceled P123, and sold the company to MarketWise for around $7.3M plus who knows what else in options. So we learned something about business :face_with_diagonal_mouth:

(not implying anything btw, he did nothing wrong, we simply did not ask for anything)

9

@marco, very nice anecdote hahaha

That makes sense, I’m sure you have a backlog of features that might be more relevant or add more value to the product. What I had in mind was rather simple, like a checkbox for “Allow access for debugging purposes” that defaults to true in the settings, and filtering by that variable while debugging.

10

Marco,

An observation that I think is factual: This never would have happened if Chaiking had to pass it by Yuval.

So the costs for a DataMiner download today are reduced compared to what happened with Chaikin, No?

You should welcome ideas for DataMiner downloads consider the reduced cost. Turn it into a cost/benefit analysts.

And in Chaikin’s case at least it might have been good that he was spared the debate in the forum and the final decision on his method’s worth by Yuval (or whomever is in charge of that now). Good to the tune of a few million dollars. Although, I suspect Chaikin might have had some fun with it.

BTW, Dan gave me a simple no on my request. But he listened and understood my request and was able to understand it, was able to figure out the cost (talked to the programmers) and came back with a respectful answer.

You should welcome all requests for DataMiner download and treat them as a simple cost/benefit analysis (including increasing memership). Treat us with the same respect that you gave Chaikin in other words.

For machine learning requests, if you are interested in brining AI/ML to P123 it might include having someone who is interested in marine learning and who will take the time to understand the request (again, like Dan in this instance).

Again, simply have someone who is interested in and capable of accurately assessing the cost/benefit analysis of a machine learning project (including any DataMiner downloads).

It has been profitable so far, No?

Jim

11

Jim, huh?

We’ve been building a AI/ML system for the past two years at a huge cost. We’re all in. We’re getting close to a release. And we will still support DM.

12

Marco,

Yuval does not not understand machine learning. Full stop Accurate.

My request for an up to date download is a request I have made for the last 10 years in the forum. It was made initially when I had an XGBoost model I wanted to run daily. I wanted to download the screen data but there was a 500 row limit for ranks.

I made the point then that rank work just fine for machine learning which gave birth to your profitable API and DataMiner.

But still no daily download that I could use to rebalance my XGBoost port. I ended up abandoning it, in part, because it was hard to rebalance. BTW, I use a lot more factors today and I would never attempt it today with the screener.

Now you are attracting machine learners and Jonpaul is in the same position I was 10 years ago. He is better position than I was in part because of the API and DataMiner that I told you would be useful for him.

Anyway, Dan initially took my request for more than 500 rows of downloads in the screen and helped me understand how I could do some of that with DataMiner.

He understood my request after 10 years. Daily would be nice but he said that would not be profitable to P123 at this time.

You have spent money on machine learning but the person who understands machine learning is completely sequestered from us.

There are some machine learing request that could be made in the forum and addressed in less than 10 years. Even if the answer is no after a cost/benefit basis, it would not be bad if someone on the forum understood machine learning.

Yuval who is the face of P123 does not understand machine learning and has had no use for anything machine learning in the past despite his recent conversion.

Generally, I cannot think of a single statical metric or machine learning request that P123 implemented starting with Parker’s request for the information ratio a long time ago.

Even if you are for machine learing it is NOT EVIDENT IN THE FORUM CONVERSATION. IT IS NOT EVEN GENERALLY UNDERSTOOD IN THE FORUM BY P123 STAFF. Dan being an exception. He took the time even though he may not use machine learning or the type of machine learning that I do.

Any feature requests for machine learning that were Yuval’s responsibility a product developer have been totally ignored (until Dan helped me). Prove me wrong. You are missing some easier opportunities to advance machine learing.

Jim