You massaged an algorithm until it produced a 24.85% annual return training on historical data.
Come back when you are ready to claim you have made ~24.85% per year with an algorithm you created 5-10 years ago.
Deny it, Downvote it: Destiny still arrives.
I fully understand the idea of historical algorithms being no true indicator of the future. With that said.......
If an algorithm consistently performs over 20+ years of data (through multiple black swan events, multiple major events), then why is it not safe to assume it likely will continue going forward? Wouldn't 20 years of "evidence" be a huge amount, such that future events likely wouldn't deviate much from that...?
My ML knowledge is somewhat rusty... does overfitting occur more often on models with many input parameters (ie.. neural networks).
His algorithm seems very simple, without really using ML at all, it's more of just a procedural 1-2-3 step thing, with no actual learning.
Can you explain how overfitting works into his algorithm?
I specifically asked how overfitting applies to a simple procedural technique, rather than a multi-dimensional method like a neural network.
Trying myself, and losing money, doesn't explain how overfitting applies to procedural steps (as I said, my ML is rusty)
For momentum, Jegadeesh-Titman paper is much of the foundation for these types of strategies, if you're interested. But as others have pointed out, the "smart money" saturated this trade decades ago.
> Implementing this on any kind of scale would be expensive to trade since it rebalance's daily.
On {day/week/month} n, determine which {n} product(s) of {product brand}'s {product type} of {underlying asset type} gave the {highest/lowest/some of each} return.
At n + {a number} {days/weeks/months} go {short/long/some of each} at {market price/limit price} at {market open/market close/time in day} the previously identified securities. Close your position on day n + {a number} afterward at {market price/limit price} at {market open/market close/time of day}.
This particular attempt seemed to have succeeded. But how many were tried that didn't? If you torture the data long enough it will confess.
People invent many algorithms all the time. You only hear about the ones which turned out exceptionally well.
In the extreme: If you take completely random price fluctuations and plot them on a graph then review them, they are no longer “random” - they become fixed because they’ve been recorded. It becomes possible to look for patterns, of which there will always be some (that’s the nature of randomness). It doesn’t matter the time scale. So your system that works over 20, 30, 50 years of past data is just fitting to the randomly-generated but not-actually-random once recorded set of data. From now into the future, the randomness will re-emerge and the system will fail.
Another way of thinking about it: imagine you click a button to generate a random time series over 20 years, then build a trading algorithm based on that single click. Great - it works! Now click again and regenerate the entire series and see if it still works :)
Obviously, sitting down and creating that "strategy" would be silly, nobody would think that would work. But if you're using ML or just test a million random strategies, you could end up with something along those lines, a strategy that is optimised for the particular paths stock prices took up to know.
The best strategy doesn't seem to me to be straightforward to calculate, even with perfect hindsight.
It might be to continually switch in the short-term most profitable asset, after taking into account the transaction costs and risks of influencing the market.
especially since market paradigms shift.
ie the last 20 years have been low interest rates, low inflation, with strong secular growth in tech and stagnation elsewhere, and globalization
this varies a lot from the 60s or 70s where there was massive inflation, high in interest, etc..
we may be entering a new paradigm with the changes in fiscal, monetary policy and globalism running its course
If anyone is interested in trading algorithms there are entire communities that filter out random backtests that havent been run live.
Check out QuantConnect