Wealth Lab just got an extension that allows to use Universes and P123 ranking systems. This allows to combine the fundamental analysis from P123 and the technical analysis from WL.
https://www.wealth-lab.com/Discussion/Build-1-of-Portfolio123-extension-released-12300
https://www.wealth-lab.com/Discussion/WANTED-Beta-Tester-for-Portfolio123-extension-12272
https://www.wealth-lab.com/Discussion/Portfolio123-com-s-stock-ranking-systems-for-subscribers-12156
HI! I am the author of the Portfoio123 extension for Wealth-lab.
Today, I see three typical use cases, listed in order of complexity:
Select a universe and ranking system from Portfoli123, build a dynamic DataSet (the WL name for Portfolio) for WL and run backtests of complex trading strategies on such a dynamic DataSet.
Calculate interesting/complex indicators, upload these indicators as Stock Factors to Portfolio123. Combine these Stock Factors with other factors to arrive at a tailored ranking. Proceed as in 1.)
Conduct a detailed study about the influence of various fundamentals on the profits of individual trades. This should result in a list of fundamentals/factors that have a high correlation with trade profits. Use this information to build a Universe/Ranking. Proceed as in 1.)
People have asked how typical results look like if I combine a ranked portfolio with a decent trading strategy in Wealth-Lab. I did a quick prrof of concept with a publicly available trading strategy called "Knife Juggler." I run the exact same strategy on three portfolios:
The three resulting equity curves, shown side-by side look like this:
I think it is worth noting that no optimizations or other fiddling with parameters was involved. All parameters used were "invented" a long time ago.
For those who like technical analysis.
It seems very good, right?
It is much better than almost all traditional technical indicators
The conclusion is that similar to media sentiment indicators, almost the best technical analysis is very good in theory, but useless or essentially useless after paying transaction costs.
The lack of universalism is concerning
I really don't want to spark a religious war here!
It is very clear that people intersted in the markets split in (at least) two groups:
Usually the traders (present at wealth.lab.com) have not much of an idea about fundamentals, there is a lack of good data providers and much more a lack of interest and understanding. We try to have many trades in a short time, we don't want to wait for the next earnings report in three months.
On the other hand investors seem to hate technical analysis, that often is close to voodoo (for them). After 20 years of research I can say: Most of the published material (books, journals, internet) is worthless.
But: There are some quantitative ideas/methods/algorithms (we call this Trading Systems) that have a reliable edge in the market.
And I firmly believe that a combination of profitable portfolios (based on ranking of fundamentals) with a clever Trading System will result in better performance than either alone.
But before we exchange beliefs here: Let's see what users of that new bridge between these worlds come up with. A solid result says way more than any discussion based on opinions or beliefs.
And Re your SSRN paper: It is very easy to show that something does NOT WORK. There are so many ways to do something wrong!
It is much more intersting (and harder to do) to show that something works.
These are not two worlds. There is only one world, the world of predictive power. As for technical analysis, there is a lack of indicators with out-of-sample predictive power beyond basic momentum-type indicators.
I have read basically every exploratory paper on the subject (technical analysis). This paper has been the most convincing of them all. But sadly, it's still not feasible.
It's not “different beliefs“ and there is no religious war here, there is only a war of science against superstition.
Edit: This is almost the second-best paper I read on this topic:
I created alphas for WorldQuant and they were agnostic as to what you could use to create an algo. Price and volume was standard for mean reversion signals. They had fundamentals and all kinds of alternative data sources. Here at P123 we do mix them and we don't wait for earnings reports. All kinds of signals that combine price and fundamentals. Not high turn like intra-day platforms but most systems here do use both.
In addition, individual backtests or trading records as you had claimed are far from being “solid results”; discounting and adjusting based on low perspective and low publication bias studies may be called “solid” (still not really), but does not depend on “results” LOL
There are many ways to do it wrong, which just means that when even the best way doesn't work, then you're much much more likely to be wrong. And while individual studies far outweigh the “solid results” of any individual's luck, choices, and biases, even the best studies still don't mean much. And, reading a lot of research and papers and rationally analyzing and discussing them accordingly is far from being anything “(merely) based on opinions and beliefs” LMFAO
The main useful scientific work is to falsify, to lower the ceiling, to recognise that things are harder, not to discover that something “has a magic solution”. It's just a brand image of scientists deceiving themselves and laymen SMFH
For example, relativity centres on an upper limit on the speed of light and an upper limit on the density of matter, and quantum mechanics centers on the lack of determinism at the microscopic level
A meaningless backtest curve chart is not "typical". Sorry.
quick prrof
And not any prrof or proof
a
I'm sorry, but you even completely ignored survival bias and selection bias
no optimizations or other fiddling with parameters was involved
“I could have been more biased, but I haven't been, so I'm quite right” Sorry, that's not how it happened
For example, the publicly available strategy involves significant overfitting until at least mid-2021 regarding beta timing in the US market. “No further overfitting as I often do” does not mean ‘no optimization and parameter tuning’. But that's only a small part of the bias
Overfitting + Survival bias + Selection bias + Others = Shit result, not so-called "solid results" LMFAO
Edit1:
Correct order:
Solid conclusion based on a lot of solid papers > solid papers controlling many biases > any paper > backtest without such a lot of bias > Single backtest with overfitting, survival bias, selection bias and other bias = merely based on opinions or beliefs, and even much worse
Not as you claimed:
Single backtest with overfitting, survival bias, selection bias and other bias > backtest without such a lot of bias > any paper > solid papers controlling many biases > solid conclusion based on a lot of solid papers = "merely based on opinions or beliefs"
No, that's obviously wrong. You are not “I'm the one who represents a combination of price signals and fundamental signals, and it's my preferred technical analysis that really represents quantitative” as you claim. The fact is that legit quantitative investors consistently combine the two and know correctly that technical analysis signals are essentially useless except for basic momentum-type signals.
I already over-catered to your emotions in my first reply, but in hindsight, I shouldn't have. Sorry, no sorry, you're not cut out for the stock market. That's only your opinion or belief.
Edit2: Of course, I would be very curious to see how receptive P123 is to free speech. Many self-proclaimed freedom lovers are actually very reluctant to accept criticism, even if it is not directed at them. “But, but, but we need civility and manners!”
Edit3: If you are trying to make a sales pitch for this random program (or platform), then unfortunately your "argument" proves that this shit is so harmful that it is not worth downloading for cashback, even though it even costs money irl
Edit 4: Also, timing in order to change exposure to a strategy with significant positive alpha is silly because the alpha part doesn't lend itself to timing.
