[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y ] [Search | Home]
4Archive logo
Prediction Algorithm - Should I publish it or not?
If images are not shown try to refresh the page. If you like this website, please disable any AdBlock software!

You are currently reading a thread in /adv/ - Advice

Thread replies: 34
Thread images: 2
File: newseventsimages.jpg (77 KB, 700x484) Image search: [iqdb] [SauceNao] [Google]
newseventsimages.jpg
77 KB, 700x484
Hi everyone. I'm a computer science phd student and I've developed a financial markets prediction method that has consistently generated profit for more than 5000 hours in different currency exchange markets. The profits were calculated by doing a simulation.

I was writing a paper to publish the method, but then I began thinking if this is the correct thing to do. What is your opinion?
>>
You could sell it to somebody and make mad money. Or just let them pay you for your predictions and don't tell them how you do it. Get rich and fuck bitches
>>
Before posting a reply, it proved to be profitable for 30,000 hours, not 5,000, in 6 different currency exchange markets.

>>16863945

How about publishing it and use it for myself? Use it myself to trade (I'm actually afraid of this, and I now think this is a bad idea)

Now, should I sell it or sell the predictions?
>>
>>16863945
why would you sell an unproven algorithm to someone, only to watch the make insane amounts of money off of it? if it actually works why not keep it secret and make insane amounts money, then publish it on your deathbed and explain why it works
>>
>>16863964
This, anon. You would become a legend. Do it! NOW
>>
lets talk [email protected]
just some email i made up
>>
Trading is all algorithms in the first place. If you have a good one, use it.
>>
>>16863935
Why would you even tell someone about it? Get rich and let someone discover ot after that... i work in the financial industry and every day i doubt if it does any better to the society.. you are an inteligent being, the money is better kept in your poket than the big banks
>>
>>16863964

Well, that's another option. I could use it and publish it after some years.

What do you guys think about selling the predictions?
>>
>>16863990
Why not use it yourself? If you get anybody else involved they might try ro fuck you over. You would prob need to have decent knowledge about the topic to even write an algoritgm in the first place, so go and become a legend! Once in a lifetime opportunity. They will make movies about you when you die.
>>
>>16863990
Man you dont sell a gold machine.. use it until you milk it all the way... i seriously think you can use the money in a good way
>>
>>16864007
Oh fuck. If you use the money for some awesome shit, you will not only be remembered as a legend but also as a god amongst humans. You'll make history
>>
File: forest.jpg (32 KB, 480x360) Image search: [iqdb] [SauceNao] [Google]
forest.jpg
32 KB, 480x360
No, you didn't. You're full of shit.

Also, stop watching "Limitless."
>>
You have done something revolutionary. No one has ever done anything remotely like it before.

Oh wait. You're just one of the millions of whiz kids who thinks they have something that no one, in the wide world of finance full of brilliant minds, has ever come up with before. What did you do, back-adjust your parameters until it fits historical data perfectly? Guess what, your predictions will fail spectacularly tomorrow.

If you want to see how your algorithms would actually perform, create an account on Quantopian and try it out.
>>
If this works, why don't you put $100 (any amont of money is fine) into the market and get 5000% profit?
>>
>>16863990
>pls buy my predictions
>they're better than everyone else's! For realz!
>I'm just a student and don't know jack shit about finance but I made something that works! Trust me!
>>
>>16863998
>>16864007

One reason to involve someone else is to get their money to invest with. I could use my money alone, but I guess that would take more time to get a considerable amount of income.

I now really feel like hiding my method and not mentioning it to anyone. I think I needed other people to agree with that idea to make myself more comfortable with my actions.
>>
>>16864025
If it works then invite us to a party in your private yatch in 6 months
>>
>>16863935
Make bank on it for "research"

Then publish a self funded paper/study

Then retire early and have fun forever
>>
>>16864016
Hum... no, I know how to test a time series prediction algorithm. In most of the experiments I used 3000 hours for "training" (I don't use neural networks or a "trainable" algorithm, but it's a similar idea to training), and I tested it for 2000 hours. Profits were consistent.

I've done some other experiments where I trained with as little as 1000 hours and tested it with 4000 hours.

It has also been successful with 15-min timeframe data, and 30-min data.
>>
>>16864021

Well, getting 5000% profit wouldn't be possible I think. As far as I know, you need to get some "borrowed money" or... leverage I think, and you need money to back your losses. Also, my method isn't a 100% win. I'd say it wins 80% of the time.
>>
Thanks everyone for your time. You've helped me decide, and I truly appreciate it.
>>
>>16864037
There are thousands and thousands of people in the finance world who make a living from this. People far more skilled and experiences than you. How can you seriously think you have come up with something vastly superior that nobody has ever done before?
>>
>>16864055
I hope that means you realized your algorithm is nothing special, and you will not find anyone who would want to invest money in you.
>>
>>16864037
And your algorithm is something different and revolutionary because...?
>>
>>16864114

I don't. It's just what my tests are telling me. I keep doing tests and they keep showing consistency. And I keep doing those tests because I'm just like you, I'm skeptical.
>>
>>16864115

I can't understand why you are hoping for this.

>>16864117

Well, it would be different because I spent hours searching for papers which described a similar method, and I couldn't find one.

On the other hand, it would be revolutionary because, if it works, it means financial markets could somewhat be explained with a mathematical model, the efficient market hypothesis would turn out to be false, and because we could begin understanding the interaction of humans in the financial markets.

I hope this clears your doubts.
>>
>>16864127
As mentioned before, try it out on Quantopian. If your algorithm is actually good, you will get results from it.
>>
>>16864150

Interesting. But won't they own the code when I submit it?
>>
Well, considering that op is lying and that if it were true it would crash world markets if published...
>>
>>16863935

Wow, if this is real, you have the power to stop the Chinese on their quest for world domination. Although desu, it sounds like the kind of thing where the CIA, NSA, MI6, FSB and all the goverment assasins will hunt you down and kill you before you make a dime. If i were you, i would microwave, burn and sink all your research to the bottom of the sea. Saved your life
>>
>>16864158

Thank you very much for your response.

I'm just a computer science PhD student, but I have no intention of proving this to you, and it would be a really careless action.

I did a quick read of what market impact is on Wikipedia, and I have the following question: wouldn't I need to trade a huge amount of money to make a market impact?

I am in no way a professional in finance, and that's why I really appreciate your input. I think my next step will be to trade on real markets. Would I need a real account with real money to have conclusive information about my method? I assume a demo account would not suffice.

Lastly, I had to look for the meaning of "shilling for information." I found this definition on Urban Dictionary:

"A person engaged in covert advertising. The shill attempts to spread buzz by personally endorsing the product in public forums with the pretense of sincerity, when in fact he is being paid for his services."

You can rest assured that I won't try to convince someone from 4chan to invest on my idea. My first options would be close friends and people I trust.

HFT wouldn't be the same as scalping? I wouldn't be scalping, my method performs a trade every 1-3 hours.

Again, thank you very very much for your kind response.
>>
>>16864155
Read the terms. And submit whatever partial things you want.
>>
>>16864171

it would only crash the markets if the markets were predictable, but they aren't, so this shouldn't happen. If the paper got published and it was successful beforehand, the markets would adapt and it would lose it's dollar to prediction properties, if for no other reason than there will be other people out there betting with it as well as against it (because it isn't perfect). For example, hedge funds are looking for near-instantaneous low level algorithms that can work instantaneously before other firms algorithms can hit the market. It's doubtful that a CS PhD student trolling 4chan has developed something that rivals this. Not impossible, but with probability one I wouldn't bet on it.
Thread replies: 34
Thread images: 2
Thread DB ID: 546587



[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the shown content originated from that site. This means that 4Archive shows their content, archived. If you need information for a Poster - contact them.
If a post contains personal/copyrighted/illegal content, then use the post's [Report] link! If a post is not removed within 24h contact me at [email protected] with the post's information.