We’re floating an idea around the office, and we can’t seem to agree on much about it. I wonder if current users might weigh in.
The main idea is to offer a subscription-based service vaguely resembling the designer models but focused on stock picking. It would be kind of a cross between traditional newsletter subscription services such as those offered on Seeking Alpha—where you have no idea of the actual real-time performance of the service—and the kind of service offered by Collective2, where you can follow the real-time performance, but which is focused on technical analysis and charges an arm and a leg ($99 a month plus they keep 50% of the subscription fees). All portfolios would consist of manually chosen stocks.
Here’s what I think we agree on:
Stock pickers would, using Invest, keep track of every stock they buy or sell. The real performance and certain past holdings would be available for anyone to see. No subscriptions would be offered until the stock picker has six months’ worth of real-time performance. Subscribers would be able to follow the stock pickers’ picks automatically via Invest. There would be no charge for offering a subscription, though it would cost resource units. And, obviously, there would be no simulations.
Here are the big questions we can’t seem to find answers for:
Would anyone currently using P123 be interested in this? We’re wondering if our users are all like me—people who understand backtesting and data and accounting and performance measures but rarely analyze individual stocks—or whether there are users who prefer to choose stocks manually, and are interested in offering their expertise to paying followers.
Would stock pickers be required to write about each stock they buy and sell?
Would stock pickers be allowed to offer more than one service?
We’ll be offering strict accountability. Will that discourage people who are currently doing a newsletter thing? Maybe they don’t want to be accountable.
What should it be called?
What should be the liquidity minimum for offerings? Some of us favor a very low minimum, others want a high one.
Can we use the current designer model format as a basis for this or should it be completely different?
Should there be some sort of in-office vetting to make sure we’re offering only reputable services?
Is anyone interested enough to help us build a prototype?
Should the blogs be like chat rooms, in which subscribers post questions and the stock pickers answer, all public to subscribers? Or should the blogs be separate and questions be messaged to the stock picker in a separate space?
How would stock pickers market their service?
Is “stock picker” one word or two?
Just to make clear, this idea will probably not be implemented in the very near future: we have a lot of things on our docket!
I would probably recommend against too much human “analysis.”
I would go with something like Vanguard’s “Analyst Opinions.” These are all automated and include Zacks Opinion Research, Ford Equity Research, ValueEngine, Citigroup Investment Research, Refinitiv/Verus……many. The automation provides some text that, I think is just canned text for different ranks or levels.
Zacks for sure is just the Zacks Rank. Any others that I have looked at closely state they are quant models. There may be human involvement somewhere for a few.
Quant models. Isn’t that what we are supposed to do? We should finish doing that before going to discretionary stock picking.
Probably. But the alternative is to let Zacks keep the market for this.
BTW, Zacks has tried human analyst involvement. They have kept some and many have failed: people with degrees in finance and access to computer models.
I will go with two as I make it a policy to never disagree with a former editor on grammar;-)
I think Yuval is getting too specific with the questions. I would simply ask
A - Is there’s a desire from designers to offer models they control manually (as opposed to the current , fully automated Designer Models).
B - Is there’s a demand for said models
Thanks
PS. the closest thing to the above is a newsletter. Key differences would be that it’s a list of holdings and weights, with transaction history, performance tracking, etc
No. In my view this goes against the very core of automated strategies which we are fortunately able to develop on p123.
Not for me at least. I can’t see why a stockpicking portfolio should do betterr than an automated strategy. In my view there is a lot more psychology coming into the picture when stockpicking is applied which can be very dangerous.
I vote no also- prefer to use other sites like Value Line, Morningstar, etc. for individual stock analysis. Would prefer P123 to devote more attention to maintaining and testing the current software platform and database…
If you do what others does, where is the edge? Right now you are a unique plattform, I think it’s importat to remain like this. Then, I think you should keep on your track.
This negative feedback should have been expected. You guys obviously like the mechanical platform, are great for us, but are a very small universe (and I don’t think charging more is the answer). We need to spread our wings (diversify?) and give more reasons for people to use the site, even if they don’t pay us anything. That’s what these projects are for: Invest, broker integration, a planned blog, etc.
We can have this capability with very little work , or go all out with new interfaces. My gut (and your feedback) points to very little work within the Designer Models framework, like a checkbox to filter for “managed” or “automatic”. And some minor mods to Invest strategies so a designer can expose the manual strategy as a DM model.
With a proper ecosystem that keeps producers and consumers coming back, it might work. Just like DMs right now, it takes time. Would be easy to give up on DMs now after the initial excitement and flame out. I still think automated DMs will be something. It just needs track record.
I think that P123 could subsidize the site by offering an advisor platform. Don’t use the P123 site. This is for developers. Make a slick site with all kinds of thematic models. Allow the advisor to select models and weighting for each client account. Then on rebalance day, they simply need to verify the recommendations in the mother models. If they accept it, the changes cascade into each daughter account.
If you can give advisors a wide selection of models and semi-automate their process - I think you’d really have something to market. I can think of people that would probably help you with marketing as well. I would be interested in providing models - but you already have quite a few on the site you could repackage.
We do like the mechanical platform and I want to amend my “negative post.” I might be open to providing an automated solution using data and methods not available thru P123.
Call it: Big Jim’s Big ML. Yea, I kind of like the alteration.
Hmm, maybe I will consider other names but I would think about doing this in a serious manner.
How about Yuval’s “Ultimate Omega” assuming he still likes that algorithm.
Or Geog’s Ultimate Market Timer: using various time series analyses or forecasting methods. Maybe something that uses his own universes—as he has discussed on the forum.
Would David (Primus) have an automated solution—a regression or something—that could be made available if he did not have to share details.
Walter?
This seems to be the knee-jerk reaction of the community:
P123’s answer to this might be: "You are right. Help us with our automated solutions by making a few or yours available to the site.
You could also keep with the plan of allowing some discretionary solutions.
With this community, I think the automated solutions have the greatest chance of surviving—especially when the API becomes available. People could do a lot with an API and Python. And there would be no need to discuss the method or give away any of the “secret sauce.” Personally, I would let the results of Big Jim’s Big ML tell the story—people could guess at the algorithm all that they want.
Indeed, there would be no ability for anyone to see what is being done on a personal computer with Python. This would probably be important for anyone that actually had something that works. This could prove significant for getting really good ideas.
FWIW. I am for anything that helps the P123 business model.
I think a big problem with traffic is that the average retail investor (and let’s face it that’s where the traffic comes from) either doesn’t understand nor trust a model designed by someone they don’t know. It is much easier and for the lamen feels “safer” to invest into an ETF. With the continued evolution of ETF’s they are now growing to include quant picking, AI learning, strategy weighted rather than broad indexes of various types that used to be the norm. I see ETF’s continuing down that path in the future as well. ETF’s offer the one stop shop portfolio for low fees. That was a problem that existed for run of the mill retail investors that would just hand their money over to the bank. Now they can make a very easy purchase and feel like they’re in more control. Finding solutions to common problems is a money maker in it’s self.
So how do you connect the general public to the world of investing and not make it feel overwhelming, and feel safe and understood? I think the Invest idea is great but isn’t really a traffic generator. At best it will add a few additional subscriptions and that’s always good of course but you said you want traffic. So IMO what you need is something that changes information (that people care about) frequently. I think the blog is actually a good idea. Make sure people can post YouTube vids since that could be an incentive for blog writers to gain an audience and generate a bit of a side income through YT views.
Thinking out loud, I suppose i’m less interested in what stocks someone is picking vs what they have to say about their story. The sharing of “secrets”, “hints”, ideas, fwd projections, track records (that’s an interesting one), these are the kinds of things I think people would come back to update themselves on over time. Building an audience is the basis for how Youtube was designed and you can’t do that without continuous new content.
Right. That’s how I see it too, it opens the doors to external models. But a good name is needed for these “manual DMs”. Names like “newsletter” or “stock-picker” will not work at all.
Maybe an “external” attribute is all that is needed, which means the signals come from outside P123. Simple. Why make them into completely different species?
Just a reminder … Designer Models were started to cater to people that don’t want to or can’t create their own strategies. It has not taken off for a variety of reasons but we are working on fixing them. And the designers that stick around, and have the longest track records, will be rewarded eventually.
I like the idea and it cannot hurt and it might be educational. The vast majority of Designer models do not beat spy over 3 years. It would be interesting to see if someone manually picking or using a different mechanical model can actually beat SPY over 3 years. I personally would like to know how they did it. To think the only way to beat spy is with what we have is a bit naive. If someone has a better way let them prove it.
I very much agree with that and from my experience with investment Meetups, I would go even further and say many, if not most, individual investors have no interest in understanding the stock market or building models - they just want to make money.
To overcome any barriers to model acceptance, P123 could offer several models based on well-known strategies (Dual momentum, Magic Formula, Swensen, Permanent, All-Weather, 7Twelev, etc). If the models were free and included source, all the better. In addition, perhaps one free, limited Book would be nice, too.
And P123 really needs to market more. Over perhaps a dozen Meetups, I’ve have yet to meet anyone who know about us. It’s pretty amazing. They know about every two-bit, flea-bitten screening site on the web, but not P123!
Do you guys do any drip-marketing campaigning for those that visit? Is there any sales collateral for email campaigns?
I do not know much about this but I take Walters comments very seriously.
I do know this. It is very possible to get bombarded with Zacks emails. I DO NOT MIND. First, some are pretty good. And unsubscribe works perfectly.
From a different field: it used to be illegal to advertise drugs on TV. Now it is okay and okay to put some pretty weird adds on the TV even e.g., take this migraine medicine and you can play with your child in slow motion all day (or whatever).
Point is, anything goes as long as one can unsubscribe, I think.