Top rated private investing Course ?

[skippybangkok]

Hi

 

Retired a while back, and been real heavy into python programming language ( self taught after retirement ) - can get any data set I want ie 2500 ETF tickers going back 15 years if avail....  used some algorithms ie. Efficient Frontier etc, had some success, overall still made decent money, but really feel I need a more educated approach to it.

 

Anyone know of any really courses ?

 

found this one - not cheap - and no clue good or bad

 

https://www.warriortrading.com/trading-courses/

 

any advice appreciated 

Edited January 23, 2022 by skippybangkok

[Yellowtail]

Why would people that know how to predict the market be teaching others how to do it?

 

 

[skippybangkok]

2 hours ago, Yellowtail said:

Why would people that know how to predict the market be teaching others how to do it?

 

 

very helpful....  you must of been a wonderful motivational speaker at one time in your career 

Edited January 23, 2022 by skippybangkok

[timendres]

I would not recommend Warrior. In fact, I would not recommend any course that charges real money. Most of the information you seek is available for free and in quantity, and is simply regurgitated by these pay-for services. What you need to start with is what style of trading you are interested in - stochastic, trend following, mean reversion, etc. And what timeframes and markets you are interested in. Once you have some idea of what you are most interested in, simple Google searches and some sifting through the chaff will reveal many good sources of information. As intimated by @Yellowtail anyone with a working strategy will never reveal it to you. Even for money. Once a strategy is known by enough traders, they will pile on sufficiently to move the markets and kill the strategy.

 

The firm I worked for years ago was running risk-arb pair trading on mergers and acquisitions. It was raining money, with spreads from 50 cents to a dollar. After about 4 years, many other firms piled on, and in the end they were trading for four TENTHS of a CENT spreads, until finally the spread dropped below their costs and they turned it off.

 

Having coded many automated trading systems over the years, both personal and for trading firms, I can say, without much hesitation, that you will find it is not simple. Many people will present trading systems based on "signals" (macd, rsi, vwap, etc.) with backtesting showing 60% plus win rates. Don't believe it for a minute. I have personally taken dozens of these systems that "demonstrated" success with backtesting and proved that they simply do not work over a large enough dataset. Easy to cherry pick a time series and produce a backtest with positive results. But when you run over years, you will see they fall apart

 

The problem is that all of these signals are "backward looking". They are never predictive. A number of very solid papers have shown that trading on these signals generally fails. In fact the only consistent approach appears to be price action analysis combined with support and resistance. Problem is, these are difficult things to program.

 

I coded a pair trading system that was producing 35% annualized returns with live trading. It was great until it wasn't. Pair trading is complex, and even the risk-arb guys get toasted from time to time. It is difficult to succeed with this approach without significant resources and vey low costs.

 

My most recent success has been with a momentum trading strategy using moving stops. It does quite well, but requires that I trade a narrow set of stocks that pass a specific filter. What I dislike is that the strategy requires holding positions overnight, which halves my potential margin, and exposes me to overnight risk. I prefer strategies that enter and exit within a single trading session. But the best of these tend to be stochastic algorithms that require very significant capital and margin, since you need to trade enough positions that the statistics play out as expected.

 

[talahtnut]

Waste of time predicting for peanuts, you've got to be in the know,

being a politician helps

[OneMoreFarang]

What do you want to learn? Trading or how to program?

I am pretty good in programming but I have little knowledge in trading.

 

If you know a lot about trading and you want to "translate" your knowledge into a program then good luck with that. There are hordes of well paid professional programmers out there. And it seems they didn't all quit their job because they write programs which trades for them and makes them rich.

[skippybangkok]

13 hours ago, timendres said:

I would not recommend Warrior. In fact, I would not recommend any course that charges real money. Most of the information you seek is available for free and in quantity, and is simply regurgitated by these pay-for services. What you need to start with is what style of trading you are interested in - stochastic, trend following, mean reversion, etc. And what timeframes and markets you are interested in. Once you have some idea of what you are most interested in, simple Google searches and some sifting through the chaff will reveal many good sources of information. As intimated by @Yellowtail anyone with a working strategy will never reveal it to you. Even for money. Once a strategy is known by enough traders, they will pile on sufficiently to move the markets and kill the strategy.

 

The firm I worked for years ago was running risk-arb pair trading on mergers and acquisitions. It was raining money, with spreads from 50 cents to a dollar. After about 4 years, many other firms piled on, and in the end they were trading for four TENTHS of a CENT spreads, until finally the spread dropped below their costs and they turned it off.

 

Having coded many automated trading systems over the years, both personal and for trading firms, I can say, without much hesitation, that you will find it is not simple. Many people will present trading systems based on "signals" (macd, rsi, vwap, etc.) with backtesting showing 60% plus win rates. Don't believe it for a minute. I have personally taken dozens of these systems that "demonstrated" success with backtesting and proved that they simply do not work over a large enough dataset. Easy to cherry pick a time series and produce a backtest with positive results. But when you run over years, you will see they fall apart

 

The problem is that all of these signals are "backward looking". They are never predictive. A number of very solid papers have shown that trading on these signals generally fails. In fact the only consistent approach appears to be price action analysis combined with support and resistance. Problem is, these are difficult things to program.

 

I coded a pair trading system that was producing 35% annualized returns with live trading. It was great until it wasn't. Pair trading is complex, and even the risk-arb guys get toasted from time to time. It is difficult to succeed with this approach without significant resources and vey low costs.

 

My most recent success has been with a momentum trading strategy using moving stops. It does quite well, but requires that I trade a narrow set of stocks that pass a specific filter. What I dislike is that the strategy requires holding positions overnight, which halves my potential margin, and exposes me to overnight risk. I prefer strategies that enter and exit within a single trading session. But the best of these tend to be stochastic algorithms that require very significant capital and margin, since you need to trade enough positions that the statistics play out as expected.

 

Thanks for the great insights. Really helpful

 

I have no passion for day trading, but prefer more the value investing style or crunching large data which I can access 

 

You are spot on with the back testing. I have become quite good with python and have a set up where I can cut weekly data ie   01 Jan 2013 to 15 Dec 2019 and then using the algorithm, ask it to predict what to buy..... then add an extra week of data and see what it plans to sell, and then buy ( so ie 01 Jan 2013 to 22 Dec 2019. As such, there is no ability to see the future with this approach and the model will deliver or not

 

Tested an enormous amount of code on GitHub all saying can make x% - as you said, most are <deleted>. found one which works, but not happy as I see too much inconsistency with different data sets - trying to find root cause.

 

Currently using 2 approaches

- value investing - looking for undervalued stocks

- efficient frontier ( returns / risk maximisation )

 

So far, have faired well, survived the covid crisis - but really want to improve 

 

found one free resource which is extensive and well written. Not exactly what I was looking for, but never the less, I can learn from it.  Zerodha

 

Thanks again 

 

 

 

 

 

 

 

Edited January 24, 2022 by skippybangkok

[skippybangkok]

13 hours ago, OneMoreFarang said:

What do you want to learn? Trading or how to program?

I am pretty good in programming but I have little knowledge in trading.

 

If you know a lot about trading and you want to "translate" your knowledge into a program then good luck with that. There are hordes of well paid professional programmers out there. And it seems they didn't all quit their job because they write programs which trades for them and makes them rich.

Thanks

 

Not the worlds best programmer, but have done some cool things with python ( ie 400MB data downloads and processing ). Also not a trading whiz.... so a nothing ????

 

 

[CrunchWrapSupreme]

14 hours ago, timendres said:

anyone with a working strategy will never reveal it to you. Even for money

Heh. Same here in the teaching arena. I've found my fellow teachers some of the most unhelpful when it comes to our profession. They each feel the other could potentially take their job.

 

This has now been amplified with online teaching and everyone vying for online students, especially with the Chinese govt cracking down on servicing what was once the most lucrative group of students.

 

If someone's stumbled upon the neat trick that nets them more students and longer regulars, they're certainly not sharing it. Other than the obvious: be an attractive, perky young lady. ????

Edited January 24, 2022 by CrunchWrapSupreme

[KannikaP]

44 minutes ago, skippybangkok said:

cut weekly data ie   01 Jan 2013 to 15 Dec 2019 and then using the algorithm, ask it to predict what to buy.....

Do ALL investment companies/advisors have to say that 'past performance is no indication of the future' and 'You may end up with less than you invested' ?

[kingstonkid]

1 hour ago, KannikaP said:

Do ALL investment companies/advisors have to say that 'past performance is no indication of the future' and 'You may end up with less than you invested' ?

Yes unless they want to be sued and the government shut them down.  No ONE can guarantee that you are going to make money.  No one can guarantee that the stock you choose will not go down or up in value over a long period of time.

 

 

[skippybangkok]

8 hours ago, KannikaP said:

Do ALL investment companies/advisors have to say that 'past performance is no indication of the future' and 'You may end up with less than you invested' ?

Thaksin came back few times, and who knows, maybe Trump in 2024