Investment BS Thread - Stocks/Futures/Mutual Funds/Bonds/Commodities/Options/ETFs/401ks/Etc

eh, Bank of America, Wells Fargo, etc.. they all kind of suck. Crappy interest rates, dumb fees left and right, etc.

 

Uh yeah, that one is obvious. But not many people get car loans at 2%? It may feel that way since most of the people in our little echo chamber on this thread are financially savvy and value trucks. Most folks get car loans for 3-4%, dont have 300k invested, and dont have it invested in a way that returns them 1k a month.

Make it harder and more realistic. What do you do if you work full time, making 100k a year, have 100k in school loans, your wife's car has broken down and you need to buy a new one within a week, you have made contributions to your 401 and IRA over the last few years since you have started your job, but don't have too much saved up (maybe 30k at this point), no dividends. However, you do have 10k in a brokerage account from some investments you made ten years ago. You can only get a 4% loan and you are looking at a 25k used crew cab truck so you can start thinking about a family and go camping.... You know, like real people.

 

A combination of C++, R, and Python. Crypto is 24/7, so its all-the-time-trading. Yes on taxes, my account get reported to the IRS. I trade through US based platforms only. My account on coinbasePro is FDIC insured.

My intention is to build this in the context of crypto because its so cheap and I use tiny amounts of money that I dont mind loosing, and then adapt it to stocks. Its basically done. I need to rewrite a few things to let it compound so the account grows faster, and then I'm thinking I might look into a trading platform like interactive brokers or seeing what robinhood will cost with the volumes I move.

 

I take it you are using C++ for the API? I am familiar with R and trying to get more into machine learning. I have barely dabbled in Python but I know I will need to eventually.

As for high volume trading, you might run into issues doing that on a non-crypto exchange. Sounds like you've got a good thing going with what you are doing right now.

 

I understand how real life can be different. I was just wondering, numbers wise, what your take would be.

And in your scenario, if I was the person making $100k, I would be living off of $30k or so, and would quickly have that $300k portfolio. That is exactly what I do (with some of the numbers in the example changed). Maybe that's the problem, people don't live below their means. I think I am living off of 30% of my income right now, investing 70% of it. Why don't more people do that, it's not hard?

 

I use python and r to talk to exchanges. I use c++ to do the computationally taxing stuff.

My recommendation is to pick a strategy that makes sense to you. I look for ideal buy ins and no matter what the market does, I only take 0.5%, 1% or 1.5% profits - no mater what. I do it all the time, so I make money. In truth, when BTC does these huge runs I average about as much profit as I would make if I had just bought in and left it alone. However when BTC has these spectacular crashes, I still make money. So in the end, I beat the market. I know people who model the ideal selling points, I know folks who focus on predicting the "next bubble" (the twitter API is great for this purpose - see how many people are talking about crypto and then use text parsing algorithms to determine positive vs negative thoughts), and I know people who focus on only crash buy ins, etc. The point is there are a lot of ways you can make money in this market, but you have to commit to one to ensure you do the right analysis, test it for a year, then give it money and let it trade, and trouble shoot trading, and maintain it. I make it sound like its not a lot of work, because I am on the tail end of investing two years of my time creating my programs, which are now working smoothly. Its not as trivial as most people think. Anyway, I'll be starting the slow and conservative process of making mine more and more profitable. Time will tell.

 

are you using any particular R packages for your algorithms or are you using python to build/train models?

 

Because many people are forced to live off 30% of their income due to the debt they have incurred. Lot of people come out of college with 250k in debt. So instead of getting a 250k investment portfolio over five-ten years, they pay off their loans. By this time, they want to have a family, expenses go up, and now they are still under shooting their retirement contributions. This is the reality most people face, even with high paying jobs.

Oh yeah, and add a mortgage and truck payment to this too.

 

Awe come on now! Im not going to give you my programs. Pick a strategy you believe in, and I'll advise from there. Also, tell me which market and platform you want to trade in.

 

I have $120k debt in student loans. I am on a pay as you earn plan for the next 25 years. This goes back to using debt correctly. With an income of $100k you would owe a few hundred bucks a month on a student loan the size of mine. I think people just don't have discipline, so they make a lot of excuses instead of learning to handle the situation. Building wealth can be done, but people don't want to do what it takes to do it.

 

I'm genuinely curious - I just want to learn more machine learning. I used to use SAS Enterprise Miner.. which was fairly point/click. But I have struggled with model building in other programs and my company doesn't seem willing to buy a SAS EM package just for me to dick around with. Has nothing to do with trading really

 

Ok, close the thread. The truth is out. Its finally been said. I 100% agree with you in every way. I have done college and graduate school with no debt (scholarships and fellowships), and bought a house, and have been ramping up my retirement contributions in recent years. Im behind you, I think, in terms of my investments, but ahead in terms of debt. It goes to show how regardless of your strategy to building wealth, its the commitment and disciple that most lack.

 

No need to be a dick, it's just a discussion man..

 

The context was as it relates to an auto loan. And it was still not good advice.
I have a 1.5% auto loan. I am not going to pay it off unless I have to.

I also sold one house for $1.0M with $600k equity. I only put down $250k on the next house and took out a $1.0M loan at 3.75% interest on December 14, 2017 with the $1.0M limit on interest write off still in place. The effective rate of the mortgage drops to just over 2%. Debt is a tool. Like any tool it needs to be used wisely. You can smash your hand with hammer or build a house.

 

Start with R and a coinbase pro account. Crypto APIs are free, which is why I started here. PM me and I'll send you some code you can use to recursively download data to play with. Do you have a linux machine? I like to use the crontabs feature to make recursive data requests.

The exact programs you will want to use to model will depend on your strategy. Most folks want to go straight for the neural nets for no particular reason except that then you can tell people you are developing an AI algorithm. Start here if you want neural nets:https://www.datacamp.com/community/tutorials/neural-network-models-r

Once you are ready to get serious about it, you need to brush up on your math in a huge way. I read primary math literature and text books. Read about fractals, chaos theory, and non linear dynamics. Be very solid in differential equations and linear algebra. What you need to know is that most people are not successful in these endeavors because they do not take the time to understand the math that is being implemented under the hood of these libraries. You will realize very quickly that most of this stuff will not help you until you understand the math. MATH MATH MATH! Finally, realize the truth about what you are doing, which is modeling human behavior. Don't forget this.

 

Everyone thinks that they are entitled to a $250k education. I worked full time and put myself through a Cal State University. Started with nothing now make nearly $600k per year. Not bragging but proving a point. I had to do a lot of hard work to get here. For starters if you can't afford a university, education debt is just stupid debt.

 

Not my intention. I'm trying to emphasize how emphatically I agree with you. Its about discipline. Its literally the best piece of advice on this thread, and the root to building wealth. Its the one and only thing people need to see. Pick your strategy and put in the work. Done.

 

My youngest son is in Computer Engineering at UCSB. He needs to read this. Pretty awesome stuff.

 

Sorry dude I read it wrong, kind of realized after that you might be agreeing lol

That is all it comes down to though, whether you make $30k or $500k a year.. either of these people can be broke or rich after a while, depending on what they decide to do with what they earn.

 

LOL neural nets... great for sounding cool, impossible to interpret what the hell is going on.

"well it did something"... "but what did it do? what are the biggest predictors?"... "ummmmm.... ermmm"

I don't have linux though, just regular ol windows 7