oetti-ds GmbH
Recommended Providers
Overall Rating
Profile Overview
About oetti-ds GmbH
Oetti-ds GmbH is a consulting company specializing in AI that advises its clients on the use of AI methods and designs, develops and operates corresponding AI models. clients include Mercedes Benz, Bausparkasse Schwäbisch Hall, Rewe-Digital, TÜV-Rheinland, GfK and Bertelsmann.
Locations
Germering
Römerschanzenstr. 1Expertise
- AI Company: 80%
- BI Company: 10%
- Consulting: 10%
Artificial Intelligence
- Machine Learning: 70%
- Pattern Prediction: 10%
- Image Processing: 5%
- Pattern Analysis & Recognition: 10%
- Speech & Text Processing: 5%
Portfolio
Key Clients
Mercedes Benz gfk Bausparkasse Schwäbisch-Hall Bertelsmann-Arvato Rewe Digital TÜV RheinlandCase Studies
Description
During my last Christmas vacations I built an AI based trading model and then tested it with real money starting from January 19th. The model generates a forecast of the market development, which I then apply with CFDs on the DAX or DOW Jones Index. The goal of the test was firstly to see if my machine learning model works and secondly to gain experience with CFD trading. Since it started successfully, I added more private (real) money after four months and thus quadrupled my total investment.
The results after 9 months: The model (strictly speaking, there are 4 models per market) works! In approx. 64% of cases my prognoses were correct and led to a profit. The statistical accuracy of the models was also in this order of magnitude, so that I can assume that it is not a matter of coincidence. The predominance of correct forecasts over incorrect forecasts has led to an overall performance of +46% (vs. +12% for the DAX). By the way, in the phase with negative performance (June) I experimented (too much) with new markets and models and was also unsettled / prone to errors due to my personal higher investment of money. In July and August I paused with investing, because I was first frustrated by the negative phase before and then also had to struggle with technical problems (with the yahoo api), not to forget some weeks of vacation.
Since September I am invested again, focused and on-track. What does that mean now? I will probably - with a little patience - become rich (--> exponential growth). My wife said at the beginning: "This can't work at all. Why doesn't anyone else do it already?" My interpretation: Even though it's simple machine learning and not “rocket science”, the approach does contain some pretty ingenious ideas in detail. The implementation of the investment strategy with CFD's combined with proper micro-timing, stop-loss and take-profit hedges is "solidly clever".
Apparently, of the approximately 1 million professionals employed in the investment industry there are just an incredibly amount of (overestimated) honks around, who do not achieve anything except purely criminal stuff like CumEx instruments.
It takes a “dahergelaufener” data scientist from Germering near Munich to show them how it's done!? I am obviously just a fucking good data scientist. Otherwise I wouldn't bet my pension on it. (Note: I have learned that modesty and subtle hints don't really work in LinkedIn). So where do I go from here? No, I will never invest anyone else's money and will refuse all requests in this regard (even from family members, closest friends, etc).
There are several reasons for this: I have only invested a fraction of my own money so far. That means that I still have a certain skepticism and therefore I will bet on an organic growth of my invested capital. I will invest my own money first before gambling with other people's money. Es wurde kein Alt-Text für dieses Bild angegeben. It is not a fully automated trading system controlled by artificial intelligence, but a recommendation system with still room for human (wrong) decisions. The four models sometimes contradict each other, the signals are not clear or not very strong. The decision whether to actually invest is still a human decision. The same is true for the exact timing of trading orders.
Do I invest at 9:46 or at 10:23? In my case, the human element has led to a greater number of wrong decisions and negative performance the moment I quadrupled my own stake and suddenly "played" for a lot more money. Not to think if I would trade with other people's money. The applied trading strategy with CFDs has limited scalability and is therefore sufficient for my own money, but not for an investment fund worth millions. This could be implemented with other instruments, but I just don't want to do so. Therefore, I will just continue to invest “lonely” and all by myself, watching how my model is slowly and steadily “printing money”. Sorry, my friends.
P.S.: If one of the honks from the investment industry would come to me to "buy" the model, I would not want to exclude this totaly. But, Dear Honk, please don't waste my time if you are not basically willing to invest a higher 7-digit amount for it.
https://www.linkedin.com/pulse/ai-based-trading-strategy-46-ytd-michael-oettinger%3FtrackingId=feJVdihOQZOzK2N0q3VhEw%253D%253D/?trackingId=feJVdihOQZOzK2N0q3VhEw%3D%3D