241031 Bad data & negative returns
A first for Two Crows Systematic in October 2024
Welcome to the 10th monthly performance report for Two Crows Systematic long-only Week#1 equity portfolios. This will be fairly brief as I’ve been
sick, and
debugging flaky data from my data feeds (again)
I’ve been largely out of action due to a flu, so am yet to pen my “omfg will you ‘AI investors’ please cut the bullshit” rant as promised previously, but if anything, this week’s events further support my theory: using an “AI investor service” (especially anything based on an LLM and/or with “Chat” and/or “GPT” in the name) as a tool to blindly make your investments, is going to fail. Spectacularly, embarrassingly, hilariously.
Why? While quickly validating the October results, I noticed my professional, paid data feed — who claim to serve customers including Moody’s, KPMG, OpenAI, and Citadel — include in their latest data feed
dividend data that, for a single stock, incorrectly contains multiple duplicate dividends scattered over several days in the same month where no such dividends occur,
have occassional back-adjusted stock prices that do not always correctly adjust for stock splits, so include random buggy “spike” data points easily visible to the human eye when plotted but not always easily programmably filterable (because sometimes, in the real world, stocks DO “flash-crash”)
newly added, incorrectly declared sector values so that logic for detecting and allocating a given stock to a particular sector, broke and needed repair.
This.. this is the kind of painful, tedious, laborious work that is required all-the-damned-time because the real world data is messy, broken and flawed, and the idea that some clowns whipped up an LLM to do their investing for them by simply pointing it at a pile of financial data, waving their hands wildly and shouting loudly “INVESTUM AUTO-MAGUM!!” makes me die laughing.
So yeah.. more on that soon.
October 2024 Week#1 Commentary
October marked a first for our Week#1 portfolios since inception, in that all four Week#1 portfolios suffered a negative return. Not ideal, nor the end of the world. Already, this month of November is looking quite promising, at its current trend. We shall see.
For now, the AUD portfolio is CRUSHING its benchmark by a factor of 5x, the JPY and USD both exceed their benchmarks (albeit barely), and our EUR portfolio continues its fairly horrid underperformance measured as TWRR of -6.5% YTD, trailing the Euronext benchmark with 3.7%.
Our Two Crows Systematic portfolios keep to trend, in that they currently rank well against HedgeNordic professional managers as has been the case YTD:
Week#1 Performance Tearsheets for October 30 2024
In addition to the performance figures stated above, PDF tearsheets for TCS Week#1 AUD, EUR, JPY and USD Model Portfolio performance figures and overview for the month of October 2024 follow:
AUD Portfolio & Benchmark:
EUR Portfolio & Benchmark:
JPY Portfolio & Benchmark:
USD Portfolio & Benchmark:
As always thanks for reading, more soon…