P123 absolute time wasters, the screen picks keep changing, three years this crap has happened, I have made some $$, buy hey, had to keep changing strategies for that to happen, now I realise your platform is not fit for purpose and all looks good with compounding !!!, a handful succeed on your platform, far too many fail, read the forums all of them, and you run after AI, has Singapore got a edge, let me get that data, wake up and fix your existing platform faults, I shall cancel my sub this month
1 Like
Hi Euroventure.
Im not working for P123, just a happy user. Would you be open to discussing on skype or some other videochat tool about what you are running into?
Id love to change your mind.
Best,
Victor
6 Likes
I'm in the industry since a while and I fundamentally disagree with your statement. P123 is by far the best / only platform which lets you (you=retail) build robust quantitative systems on the same datasets used in every major buy- and sell-side firm globally. The learning curve is steep to figure out what actually can practically be implemented but this ultimately comes down to the user and is not a platform fault (garbage in = garbage out).
10 Likes
While I agree that P123 is not for everyone, it is an effective platform. But the difficulties of developing effective investing strategies are very significant. If it were easy, everyone would be above average.
Over the years, I've made hundreds of ranking systems, run thousands, maybe 10s of thousands of simulations to end up with a handful of productive strategies. This game is definitely not for everyone. I only stay in it b/c of love for the field.
I would encourage you to stay only if you feel you can commit the time and energy and attention to this difficult subject.
5 Likes
It is my theory that people who stick with P123 have an "anchor," A pretty unshakable belief in some idea. It does not hurt if there is a little dopamine release when you are proven right. Large excess returns are optional.
Example of an unshakable belief in ranking everything: Screening the Market: A Four-Step Method to Find, Analyze, Buy and Sell Stocks 1st Edition by Marc H. Gerstein
My anchor is the unshakable belief that it will all make sense when I get the math right (still working on that). Another anchor is the belief that one can find a list of unique factors that no one else has thought of that will give you a unique advantage. P123 accommodates all 3 anchor styles in spades (including Marc Gerstein's affinity for ranking everything). No better place than P123 to do any of those methods (and more).
Some do not have an anchor that gives a dopamine release on a regular basis with investing. Not necessarily an unhealthy thing.
I have been frustrated and was a signature away from turning all my money over to a professional. I did not for more than one reason. But I remember asking my wife what my new hobby would be.
Impossible to know if I would be sailing or rock climbing now if I had done it. But nothing wrong with index funds or a profession advisor and doing something else.
5 Likes
I have only been using P123 for real investing for a few months. I had strategies that worked to my satisfaction and that I used on other platforms before using P123. Since I have been using P123, I continue to follow these strategies successfully, but I can backtest them, manipulate them to my liking and possibly create new ones based on my previous knowledge and experiences. It seems to me that what matters here is: we have ideas, we test them, we modify them, we experiment again. It is not a magic formula. You have to work, get your hands dirty. But I find it to be an interesting tool for me.
1 Like
Hi Roger,
I'm sorry you got so frustrated that you had to do it publicly, after you canceled. I checked your past conversations with us and I don't see any mention of any problems.
What I do see is a conversation with Dan in 2023 where he recommends switching to screen for at least the top 5 stocks instead of top 2. P123 backtests should be viewed as nothing more that one possible outcome of many for a particular strategy. And, the more stocks you include, the more robust the backtest is (for example using the top quintile). Only using the top 2 is a recipe for curve-fitting, and any tiny data change can greatly change past results.
Also it appears that you were running the screens daily. This alone could explain why your "picks keep changing" since, historically, we only have weekly snapshots (doing daily backtests is on the roadmap). P123 is not geared for daily rebalances at the moment. Perhaps it never will.
Lastly, you only had the Screener membership which is underpowered to create your own strategy. You can only use pre-defined ranking systems, no portfolio strategies, and only have 5 year backtests. Screener membership, in my view, should only be used for idea generation, which is then followed up with discretionary research.
Let us know if we can assist in any way.
Regards
12 Likes
I'll give my two cents here as well. As others have pointed out, P123 has a steep learning curve, but if one is willing to put the work in, the results are worth it. Also, much of the low hanging fruit of simple screen factors etc are more or less arb'd away. It takes time to develop new factors that are unique, test them out of sample, which may or may not work out.
Not for everyone - but if you're willing to put the work in, it can work out well.
3 Likes
I didn't realize that the Screener tier only offers per-defined ranking systems. That's under-powered even for idea generation, IMO.
1 Like
You can't always expect extraordinary results from subpar effort.
1 Like
For only $25/month, it is impressive that they get access to FactSet data and can do 5 years of backtesting. For that price you can take time to learn the platform well until you are ready to upgrade for more customized ranking systems.
1 Like
I wouldn't recommend it.
1 Like
There's actually no need for complex learning; Dan's list of factors is fine on its own.
1 Like
New members could be encourage to do a number of ,backtests with data from 2010 - 2020 for a couple weeks and then pick their best model and run it 2020 till now. One shot on the 2020 till now test.
Their first experience would be a lesson in realistic expectations.
Maybe P123 could facilitate this in some way. Maybe give a small bonus discount for completing the challenge and if they can get greater than 20% annualized a little larger bonus to keep them interested in completing the challenge. Keep the bonus small enough that if they cheat with the idea of getting a discount on a long-term membership P123 still makes money on a long-term membership, I don't think it would have to be a large bonus to encourage a 2-week challenge that they might understand is a good idea anyway.
I think that's probably the main reason: you're fighting the most powerful enemy on the most difficult battlefield.
Of course, even in the microcap universe, stock selection does remain very very difficult.
2 Likes
If I were to characterize P123, I would say it’s an excellent tool for cross-sectional algorithms (whether using P123 classic or machine learning) and for swing trading or moderate time frames. However, it’s not necessarily the best choice for sequential analysis (e.g., technical mean reversion strategies) or day trading. No platform can excel at everything.
It’s impressive that Euroventure managed to get sequential analysis to work on the S&P 500 with P123. That’s quite an accomplishment—and certainly not an easy task.
2 Likes
Yeah, a 71% annualized return is pretty rotten. I see now why you are quitting. 
6 Likes
And 68% alpha
If that's realistic, they just beat Yuval a lot...
I think it means they have gained so much wealth that they can't see what P123 has to offer.
A better advertisement for P123 than anything else /s
2 Likes
I wonder how this screen or backtest will do over the next 5 years (without any modifications). Maybe it depends on whether there is a correction in the tech sector.
Many, including me, have a natural tendency to believe a backtested strategy will continue—without any decay--going forward or that someone presenting a backtested strategy to us had the foresight to put all their money into that strategy—assuming they put any money into it at all.
Sometimes—without cross-validation—there can be something to be learned, I guess.
P123 recently asked me to clear out some of my old sims and ports. A reasonable request as there was a lot of useless strategies taking up computer memory.
It forced me to look at a 5-stock backtest that earned 72% annualized. What is remarkable is that I funded it and it continued to earn at that rate for a while. I was on hog heaven, until the excess returns just stopped.
I was smart enough to get out of that strategy and rethink a lot of of things—including all of my ports have 15 stocks now, But I rethought much more than just the number of stocks in my ports.
The moral of the story is that when I was forced to look at that port before removing it (on auto) it it did not outperform the benchmark for a long period. It just had more volatility. It never came back. It WAS NOT just a shift in market regimes.
In my experience It is not unusual for a backtest to be around 71-72%, BTW. Maybe a practical limit on the amount we can fit or overfit (whichever the case may be). Or a combination of good fitting with a little overfitting in some cases. I am not trying to make the argument that there might not be some overfitting when creating what turns out to be good strategy based on a good process.
As an example here is what is probably and excellent strategy which was probably based on a good process. But the out-of-sample results are not as good as the backtest around 72% again. Around 72% +/- on a backtest with enough optimization is beginning to look like something that is not unusual to me:
In the financial markets, one has to understand what is meant by publication bias.
2 Likes