Should mgrunin start a trading thread ?

[Mikesev]

I'm currently creating a semi-automated trading system based on machine learning algorithms. More specificially, I'm in the middle of a massive experiment testing support vector machines, gradient boosting, and a mixture of the two (using linear svr for feature selection and gradient boosting for classification). Should be pretty interesting.

Out of interest, why did you select those algo's?

 

[SUP3RNOVA]

My uncle has netted $x,xxx,xxx+ trading as a hobby (he is an engineer by day). He ONLY trades index funds. When the market is low and he feels all the weather patterns tell him an uptick is coming, he shoves a lot into the Dow. I don't think he's had a miss yet, but as of last year he hadn't invested in over a year. He just quietly sits and waits for the right moment to strike.

Sounds a lot less stressful than day trading, and he has done very well with it. Probably not something you can make a career out of though.

 

[cardine]

I'm currently creating a semi-automated trading system based on machine learning algorithms. More specificially, I'm in the middle of a massive experiment testing support vector machines, gradient boosting, and a mixture of the two (using linear svr for feature selection and gradient boosting for classification). Should be pretty interesting.

Is it based completely on technical indicators?

I've always had interest in developing a real automated trading system, but everything out there right now is garbage since they rely on technical indicators.

So yes, I will agree with you that if you are on the buying side of options, sooner or later you might get burnt UNLESS you play very tight strategies such as candor and butterflies. The ROIs turn out much lower, but the loses are minimized.

The money is in selling options, it's as simple as that. Let me show you just how much bullshit is fed from the market.

For weeks now Apple has been expected to miss earnings. You have all the bears running around saying they'll buy a new lambo after the miss. Then you had all the bulls running around saying Apple #1; they don't miss. Like usual, the option pricing on weeklies for AAPL is ridiculous. The implied volatility is 120%.

Well...Apple misses on EPS, iPads, Macbooks and releases a 30% lower guidance vs estimates. Now..you'd expect a big drop in the aftermarket, right? At least all the bears thought they would see. Well...Apple right now is trading 0.06% up in the aftermarket.

So who wins once again? The motherfucking sellers. Those that sold options in either direction will see the options that they sold lose 90% value tomorrow at the open.

I think you are oversimplifying things. Options are a zero sum game. Sellers lose just like buyers do.

 

[lschmidt]

Is it based completely on technical indicators?

I've always had interest in developing a real automated trading system, but everything out there right now is garbage since they rely on technical indicators.


I think you are oversimplifying things. Options are a zero sum game. Sellers lose just like buyers do.

Yes just technical indicators. I believe that technical indicators hold enough information to predict prices accurately enough to make money. The thing is, there is no simple formula involving technical indicators will work. That's where machine learning comes into play -- it can detect patterns that would be impossible to detect with traditional analysis. Most people don't have the knowledge, patience, or skills to implement such a strategy though. And nobody is going to sell you something that actually works and makes money.

You'll need to have a pretty firm grasp on calculus, probability, statistics, linear algebra, time series/regression analysis, and then some advanced stuff like probability models and finally machine learning theory. Oh and you'll need to be a pretty damn good programmer too. Not something your average joe can implement.

 

[cassius]

Is it based completely on technical indicators?

I've always had interest in developing a real automated trading system, but everything out there right now is garbage since they rely on technical indicators.


I think you are oversimplifying things. Options are a zero sum game. Sellers lose just like buyers do.

Most of the successful automated systems I have encountered employ various scalping strategies, the legitimate form of scalping, and are nearly completely driven by technical indicators. Of course the truly successful systems are almost always private and not for sale. The only reason I have even seen them in action is because they were developed by friends or friends of friends. You do have to regulate and revise the systems regularly though, especially when jobs/housing/earnings/etc. reports are due. The most effective automated systems, as far as my experience goes, will actually avoid volatility for the most part. You want small spreads with predictable movement in massive trade volume. Personally, I am weary of putting any significant funds into an automated system. I know they can work very well when properly executed, but I don't like the feeling of not controlling every trade.

 

[guerilla]

I believe that technical indicators hold enough information to predict prices accurately enough to make money.

If that was true, the USSR wouldn't have collapsed.

Prices are psychological, not technical, phenomena. Learn economics.

 

[cassius]

If that was true, the USSR wouldn't have collapsed.

Prices are psychological, not technical, phenomena. Learn economics.

This is why scalping is popular in automated trading. However, technical indicators are very valuable in evaluating probabilities and potential magnitude of PPS change.

 

[wickedDUDE]

mGrunin

mGrunin: What are your thoughts on index funds? I am specifically looking to invest with Vanguard because of their low expense ratios:

Total Stock Market Index Fund (VTI)
Total International Stock Index Fund (VXUS)
Total Bond Market Index Fund (BND)

I plan on investing over 6 figures, but don't know if I should get in now or wait until the election is over...

 

[Mikesev]

Yes just technical indicators. I believe that technical indicators hold enough information to predict prices accurately enough to make money. The thing is, there is no simple formula involving technical indicators will work. That's where machine learning comes into play -- it can detect patterns that would be impossible to detect with traditional analysis. Most people don't have the knowledge, patience, or skills to implement such a strategy though. And nobody is going to sell you something that actually works and makes money.

You'll need to have a pretty firm grasp on calculus, probability, statistics, linear algebra, time series/regression analysis, and then some advanced stuff like probability models and finally machine learning theory. Oh and you'll need to be a pretty damn good programmer too. Not something your average joe can implement.

I'm doing a MSc in ML at the moment (inb4 lol college herp derp) hence the curiosity. If you fancy having a chat about this stuff let me know.

 

[lincolndsp]

i spent half a year with a Neuroshell Trader (neural networks software - sort of a lego constructor and trainer for neural networks) but could not make it work for me

 

[ericchristopher]

If that was true, the USSR wouldn't have collapsed.

Prices are psychological, not technical, phenomena. Learn economics.

This is true.

However, technicals show market history. Market history suggests order distribution.

Visual estimations of underlying distribution DO NOT suggest price direction, but they do -- with a study of market participant behavior and the governance of a strong understanding of the fundamentals -- suggest an understanding of the battle at hand.

I'm wrong on direction 59% of the time. (I'm a counter-trend trader.)

But I don't need to be right to make money.

 

[Mike]

What are some good resources to learn basic option trading or trading at all for that matter?

I think mGrunin already pointed this one out, but it bears repeating: Investopedia

Some more:
Education: Options 101 | Webinars | Strategies | Conferences | Coaching
https://education.trademonster.com/
https://us.etrade.com/investor-education?ploc=p-TopNav
Learn About Options, Stock & Futures Trading Strategies at optionsXpress
Investment Education | Investment Products
https://www.tdameritrade.com/education.page

Almost every online trading platform has an education section. Some are better than others. Just look around. Start reading. Watching videos.

OptionsXpress has a cool calculator that lets you enter a trade and then run a calculator on it to see where you'll make money and where you'll lose before committing to the trade. I think eTrade does too, I just can never find the damn thing when I want it.

If you want to practice trading options, equities, whatever. Get out Excel and enter your opening trade, costs, commissions, etc. Then when you reach your exit, enter your closing trade and see how you did.

Great way to learn without any money.

I believe you can also do "fantasy trading" on a lot of sites if you want.

Go learn, have fun.

 

[netbert]

Edit: Doh, didn't see your own reply.

 

[OrganicJuice]

Question from an investing noob:

Anyone watch the SERPs after an update, see who benefits, and invest accordingly to who is currently ranking well? Or does that not effect things as much as I imagine it might?

 

[mx=headroom]

Question from an investing noob:

Anyone watch the SERPs after an update, see who benefits, and invest accordingly to who is currently ranking well? Or does that not effect things as much as I imagine it might?

I don't always invest but when I do I make sure it's with acaiberrryplusweightlossforever.com.

 

[cassius]

Question from an investing noob:

Anyone watch the SERPs after an update, see who benefits, and invest accordingly to who is currently ranking well? Or does that not effect things as much as I imagine it might?

It would likely only be noticeable in micro-caps.

 

[mGrunin]

I think you are oversimplifying things. Options are a zero sum game. Sellers lose just like buyers do.

Options are a zero sum game with the sellers having an upper hand. That's all that I'm saying. Biggest variable to option pricing is time. As a buyer time is against you. As a seller time is helping you. That's a big advantage.

 

[cardine]

Options are a zero sum game with the sellers having an upper hand. That's all that I'm saying. Biggest variable to option pricing is time. As a buyer time is against you. As a seller time is helping you. That's a big advantage.

But options are priced accordingly. If that was such a big advantage then it would be less profitable to sell options because everyone would be doing it, which would cause the price of options to plummet making it no longer a good deal to sell options.

 

[mGrunin]

But options are priced accordingly. If that was such a big advantage then it would be less profitable to sell options because everyone would be doing it, which would cause the price of options to plummet making it no longer a good deal to sell options.

It's not a case of everyone being able to do it.

To sell 10 naked put contracts for AAPL that is a couple strike prices out of the money requires $160,000 capital to be withheld. So in other words, if you are attempting to collect $160 in premiums on 1 contract; then $16,000 (100x) will be held from the brokerage in case you lose out on the position. So..you tell me if anyone can do this? In other words, you are risking $100,000 for less than $1,000 in possible gains and that's if the options you sold expire completely worthless.

And yes, all large institutions are on typically on the SELLING side.

 

[mGrunin]

The stock market is officially closed for tomorrow and possibly Tuesday. Guess who's benefiting from this? Once again, the sellers.