Neat test!! I can see how it would work when purchasing several stocks but I don't see it working for 1 stock because there's always factors that are beyond our control, like somebody signing or not signing a contract with a company. Signing contracts with bigger companies boost stock prices.
You're right, tmscase, but there is still the question of which methodology you will apply your computer testing to. It can produce positive results with just about any reasonable methodology you choose. I used to get excited until I saw so many systems testing positive. I've researched many of them and have written software to test them, only to be left in confusion because so many worked. Then I began to worry about overoptimization..lol.
I'm a mathematician who has been studying the stock & commodities markets for 15+ years, but the reality is that there are an infinite number of philosophies of how to deal with the mass of data that you're analyzing. You can backtest and "overoptimize" the data and produce positive results ONLY for the data you are testing and then fail when applying it. I still won't give up and I'd really like to get excited about your presentation. I was like you when I started.
Yes you are right, mathematics alone cannot gaurantee sucess. It is simply one piece of an overall system of risk analysis and managment, that when combined, lead to sucess. There is definitly a 'soft' piece to the system, which involves news, and other fundamental characteristics. What I have found, however, is that the computers systems make targetting and finding the good stocks much easier, and quicker.
Thanks for explaining that stuff - very educational. But the idea to mathematicly or statisticly predict the stock markets (and similar dynamical systems) is - I think - not very new. Isnt it?
No the idea isn't very new. Typically it's been used by large institutions as they have always had access to the data, and had large computer systems. Technology is just making it easier for the smaller guys to compete on the same level.
Predicting stock market turns based on past data may well be like predicting fair coin toss outcomes based on past data - useless.
clray123 4 years ago
....anyone seen "Pi"......?
superzerox1 4 years ago
Neat test!! I can see how it would work when purchasing several stocks but I don't see it working for 1 stock because there's always factors that are beyond our control, like somebody signing or not signing a contract with a company. Signing contracts with bigger companies boost stock prices.
JayZoop 4 years ago
That's why we buy 30 - 50 stocks a year. Short 3-7 day holds.
tmscase 4 years ago
You're right, tmscase, but there is still the question of which methodology you will apply your computer testing to. It can produce positive results with just about any reasonable methodology you choose. I used to get excited until I saw so many systems testing positive. I've researched many of them and have written software to test them, only to be left in confusion because so many worked. Then I began to worry about overoptimization..lol.
cgprnd 5 years ago
I'm a mathematician who has been studying the stock & commodities markets for 15+ years, but the reality is that there are an infinite number of philosophies of how to deal with the mass of data that you're analyzing. You can backtest and "overoptimize" the data and produce positive results ONLY for the data you are testing and then fail when applying it. I still won't give up and I'd really like to get excited about your presentation. I was like you when I started.
cgprnd 5 years ago
Yes you are right, mathematics alone cannot gaurantee sucess. It is simply one piece of an overall system of risk analysis and managment, that when combined, lead to sucess. There is definitly a 'soft' piece to the system, which involves news, and other fundamental characteristics. What I have found, however, is that the computers systems make targetting and finding the good stocks much easier, and quicker.
tmscase 5 years ago
Thanks for explaining that stuff - very educational. But the idea to mathematicly or statisticly predict the stock markets (and similar dynamical systems) is - I think - not very new. Isnt it?
vipmarco 5 years ago
No the idea isn't very new. Typically it's been used by large institutions as they have always had access to the data, and had large computer systems. Technology is just making it easier for the smaller guys to compete on the same level.
tmscase 5 years ago