KukaXoco Finance - TRADING PHILOSOPHY

Many statistical studies confirm it: most uses of most popular investment strategies just don't work. So we don't use them. Some comments, and links to solid studies on these issues.


FUNDAMENTAL ANALYSIS

Between ETFs inflows/outflows, pension rebalancing, random (Trump) political distruptions, nonsense economics such as EBITDA and non-GAAP, socialist subsidies, yet-to-be-convicted corporate crimes, meme stock squeezes by gaming retail tradeers, and the Federal Reserve's 'cocaine' (rate cuts) or 'cold turkey' (rate hikes), fundamental analysis just is not that useful anymore, more for providing entertainment on CNBC and Bloomberg. Read Barron's for inspiration. Some papers:


TECHNICAL ANALYSIS

Most technical analysis techniques, especially moving averages, look good in articles and books on trading systems. But rigorous statistical analysis reveals most are not that profitable, as compared to just buying the SPY or GLD. Some articles:


CANDLESTICKS

Most candlestick patterns are not statistically profitable over the long-term. While there are a few patterns that are highly profitable, they occur do infrequently to be of little use beyond generating some fun, spending money. Some papers:


ARTIFICIAL INTELLIGENCE

Despite the buzz of chabot LLM AI models for trading, assuming that such models provide higher returns, that everyone has easy access to them means that such returns will quickly evaporate (as have other profitable systems, once popularized, soon lose all power - especially with the biggies on Wall Street using their AI to capture most of the profits).

In the past, most funds that have tried to trade stocks and ETFs using AI-generated signals have been unsuccessful, either losing money or underperforming the relevant benchmark index. Indeed, if modern AI is trained on all of the Internet filled with commentary from fund managers and pundits who can beat the indexes, how useful can they be?

At best, AI is good for increasing a non-AI system by a few points (which KukaXoco is more than glad to exploit), this benefit can often be obtained from traditional multivariate statistical techniques such as VARIMA, for which the fitted parameters are clear and understandable. Or similar benefits can be obtained from other statistical techniques, such as k-nearest-neighbors. All of which KukaXoco is more than glad to exploit.

And when the current AI bubble pops, and it will, AI in trading will be disparaged. And it will pop, given the extreme weight of AI in the S&P 500: measures of concentration for market cap, returns, earnings, and capex (Apollo Global, Sept 2025).


BUY-AND-HOLD AND COMPOUNDING

The classic strategy for many individuals and funds. Problem? Doesn't seem to work well in practice.


SPACS - SPECIAL ACQUISITION COMPANIES

Basically, these deals are "give us a lot of money, we will find something profitable to buy." NOT!.

NO ZOMBIE COMPANIES FOR LONG-TERM TRADING

There are a large number of companies, that across multi-year, multi-decade time periods, can't outperform the SPY. So why waste time in your life analyzing these companies - just buy the SPY. And many of these companies are so poorly managed that they can't beat buying-and-holding gold/GLD.