Speaking of day trading... what platform? IRA content

Been wanting to play around with some algorithms and try my own hand at managing my retirement funds. I have enough to sacrifice about a years worth of contributions, which is what I intend to play with. 

But what trading platform should I be using? Figure that I would make about 10 transactions a month max. I'm currently all vanguard, but their structure makes frequent  trading difficult. Any thoughts?

oh, and feel free to be a naysayer, I think it is good to have in any thread in case someone else comes along for advice in the future. But I'll say that I've read more books from Jack Bogle and Warren Buffet than I can count, have a degree in Mathematics, Economics, and Business, and have been in finance my entire career. I'm well aware that I could lose it all by playing with it. 

Oh, and to be clear: I want to do it within my IRA. Save on capital gains and all. 

I think the big question is how automated the process needs to be. I haven't dug too deeply but I don't think the usual low cost brokers (Vanguard, Schwab etc) have APIs. Might be wrong about them, but for proper no-human-involved algo trading, you want a broker than offers an API. Here's an older question on stackoverflow that lists brokers with an API: https://stackoverflow.com/questions/59327/what-online-brokers-offer-apis

For algorithmically triggered trades that you stick in manually, I'd also look at Charles Schwab - they seem to have reasonable trading fees ($4.95 vs $7.00 at Vanguard if I deciphered that correctly - Vanguard fee assumes you have > $50k with them). Traditional brokers tend to be pretty expensive for options etc, so for those you'll have to find a specialist broker.

One other option that I've heard about is to use something like the Robin Hood app for trading as they don't charge fees. I don't have any idea what their spread is though and what kind of restrictions they have on number of trades. Plus, I suspect you'd have to feed the trades in by hand again. But it might still work as a cheaper way to trade.

As an aside, I quite like the book "Quantative Trading" by Ernest P Chan. It's got useful information on strategy development and backtesting in it. Just in case you haven't come across it already.

Speaking of which, backtesting is important and requires historical market data. Sources of those tend to be on the expensive side for the real thing so you'll have to try and find a source in your budget.

Oh, and the other question is where you get live market data for algo trading from. That might be supplied by your broker, definitely something to check. Depending on your strategies you don't really want to trade on the standard public market data feed that's delayed by 20 minutes or so.

I'll likely be entering the trades manually. I like the option to override it. 

I think most of the high volume traders use Interactive Brokers. Low fees and a lot of rope to hang yourself with! (Meaning ready access to margin trading and shorts.) This is where I usually discourage this but it sounds like you have your eyes open and don’t have unrealistic expectations. Let us know how it works out!

dculberson said:

I think most of the high volume traders use Interactive Brokers. Low fees and a lot of rope to hang yourself with! (Meaning ready access to margin trading and shorts.) This is where I usually discourage this but it sounds like you have your eyes open and don’t have unrealistic expectations. Let us know how it works out!

Will do. I've been following a specific algorithm for a while, actually I've been following it and doing my own picks and comparing to index... It wins over a 15 year period, which is all the data I have. Actually, it hasn't really lost at all for any period more than 2 months. Still hesitant to go full in though, so we're just dipping my toe in for now. 

Trading inside the IRA makes so much sense. One of the downsides of frequent stock trades is the tax risk - you can make a gain on every trade and end up with next to nothing after fees and taxes. Lower your fees as much as you can and do it inside the IRA and you eliminate that. The main downside is still left: finding out you're not as good as you think. ;-) But maybe you are!

Just remember it's easy to be a hero when everything's going up. We've had a bull market for close to 10 years now and it'll be interesting to see what happens when it's more historically average. I hope my mindset of "buy and hold, basically forever" can hang on once I start seeing some negatives in the returns columns. I think I'm temperamentally suited to it but only real experience can show for sure.

In reply to mtn :

Do your trading on paper or for practice to prove to yourself that you can at least beat the DOW. 95% of professional traders don’t beat it over a 5 year period. 

If you feel smarter than the pros,  you very well might be.  But prove it to yourself first.  

frenchyd said:

In reply to mtn :

Do your trading on paper or for practice to prove to yourself that you can at least beat the DOW. 95% of professional traders don’t beat it over a 5 year period. 

If you feel smarter than the pros,  you very well might be.  But prove it to yourself first.  

I say day trading, but it really isn't. Active trading, but not day trading--figure about 4-5 trades a month. 

On my simulations, which I can only have go back to 2002, I'm beating just about everything index wise, blows the DOW out of the water. My big concern is that 16 years isn't much of a history. 

In reply to mtn :v

then that’s your answer

In reply to mtn :

Depends on how you use the data. I'd specifically test around the "bad" periods during that time to see if your strategy ends up blowing the DOW out of water in both directions. That might be bad...

The big issue with markets like we've been having is that a lot of people tend to back test for a year or two, and we've been in a period a trader I used to work with once described as "a monkey with an oujia board can make money in this market". You really want to see what your strategy would have done in face of the Lehman failure and the shenanigans around that time.

BoxheadCougarTim said:

In reply to mtn :

Depends on how you use the data. I'd specifically test around the "bad" periods during that time to see if your strategy ends up blowing the DOW out of water in both directions. That might be bad...

The big issue with markets like we've been having is that a lot of people tend to back test for a year or two, and we've been in a period a trader I used to work with once described as "a monkey with an oujia board can make money in this market". You really want to see what your strategy would have done in face of the Lehman failure and the shenanigans around that time.

With data that only goes back to 2002, what periods would you use? I've obviously got 2007-2009, as well as the max period... But I've missed out on the tech bubble, as well as 1973, simply because I can't run the algorithm that far back. 

2007 to 2009, the algorithm holds up. It had growth until June 1st, 2008, when it had you hold only cash for an entire year.  

Would you be able to compare your algorithm's performance to the market during small corrections (10% or more)? The S&P saw 2 or 3 corrections between 2010 and 2013 depending on how picky you want to be with rounding. http://www.businessinsider.com/history-of-stock-market-corrections-2013-7

No idea why I didn' think of that. Yes, I can basically do it across any two date ranges. It is pretty useless on anything shorter than 2 months, only because it really only has you trading once a month. 

FWIW, I use VTI, VTS, and BRK_A as my benchmarks. Basically just use $1,000 on start date, and see where it goes between the two dates--the algorithm taking into account fees, the benchmark just holding (I have to use 100k or 200k for BRK_A, but the same idea holds true)

Just comparing to VTI....

For 11/27/2002 to 3/11/2003 correction (using 11/1/2002 to 4/1/2003), the algorithm returned 0%; VTI -2.28%

For 4/23/2010 to 7/2/2010 correction (using 4/1/2010 to 8/1/2010), the algorithm returned -15.46%; VTI -5.14%

For 4/29/2011 to 10/3/2011 correction (using 4/1/2011 to 11/1/2011), the algorithm returned 10.69%; VTI -8.11%

For 5/21/2015 to 8/25/2015 correction (using 5/1/2015 to 9/1/2015), the algorithm returned 1.78%; VTI -7.67%

For 11/3/2015 to 2/11/2016 correction  (using 11/1/2015 to 2/11/2016), the algorithm returned -4.23%; VTI -8.31%

...Interesting. It generally wins compared to VTI. 

HOWEVER: For more comparisons,

From 12/1/2009 to 2/1/2012, which incorporates two corrections, the algorithm returned 5.1%; VTI 29.55%. Big loser.

From 1/1/2015 to 4/1/2016, also incorporating two corrections, the algorithm returned 9.67%; VTI 1.91%. That wins, but not near as big as it just lost.

12/1/2009 to 4/1/2016, which has four corrections: algorithm: algorithm returns 73.52%; VTI 117.36%. Another big loss.

11/1/2002 to 4/1/2016, which has five corrections, and a recession, the algorithm returns 758.83%, VTI 212.06%. Huge win.

Now I'm thinking hard before jumping in. It seems that it is phenomenal, as long as you don't jump in at a peak. We could definitely be at a peak right now.

In reply to mtn :

2007-2010 would be the big one in the timeframe you have.

BoxheadCougarTim said:

In reply to mtn :

2007-2010 would be the big one in the timeframe you have.

01/01/2007 to 01/01/2011, the algorithm returns 130%; VTI -1.21%

06/01/2007 to 03/01/2011, algo is 170%; VTI -4.06%

06/01/2007 to today algo is 622.96%; VTI 125%

Everyway that I can play with this (start a few days/weeks/months before and after October 9 2007, finish a few days/weeks/months before and after March 5, 2009), the algorithm just blows VTI out of the water. The worst I can get is 59% for the algorithm, although I'm sure if I played with it longer I could do even worse than that (i.e. $1000 initial investmnet does better than $100,000 initial investment).

BTW, if anyone is interested in what I'm doing, I'm happy to share offline. Send an email to [username]ovak1 at yahoo dot com

Current markets have me wondering how this is looking. Any updates that take recent situations into account?

I went with Fidelity, since I already had an account and don't need immediate transactions. 

Since I started, I'm up 1%. VTI and SPY are both down in that timeframe so I feel like I'm doing ok so far. 

In May of 1986  I started a highly leveraged program of short selling.  Working through a broker I had fantastic returns.  Then Oct of 1987 happened.  By the time I heard the news and contacted my broker every last penny was gone.  

So I started very conservative retirement program based on the Dow.  When the Dow goes up I profited by exactly the same amount. Since95% of professsional brokers fail to beat the DOW over a 5 year period  I was content. 

Then 2008 happened and not only did I lose my career( and income) but my retirement started taking a massive beating.  When I was at 50% of my lifetime savings I chickened and pulled what was left.  Doing so cost me the penalties since I wasn’t old enough yet to draw without them.  While I was able to pay my bills and save my home,  I was left with no retirement income.  

I understand Warren Buffet and a few others have done very well with the market and it’s the only way to real riches for me burned twice is enough. 

I do hope not all your eggs are in one basket. So when unpredictable events happen you can ride things out

In reply to frenchyd :

Had you stayed in, you would have more than recovered your investment losses from 2008.  

frenchyd said:

In May of 1986  I started a highly leveraged program of short selling.  Working through a broker I had fantastic returns.  Then Oct of 1987 happened.  By the time I heard the news and contacted my broker every last penny was gone.  

So I started very conservative retirement program based on the Dow.  When the Dow goes up I profited by exactly the same amount. Since95% of professsional brokers fail to beat the DOW over a 5 year period  I was content. 

Then 2008 happened and not only did I lose my career( and income) but my retirement started taking a massive beating.  When I was at 50% of my lifetime savings I chickened and pulled what was left.  Doing so cost me the penalties since I wasn’t old enough yet to draw without them.  While I was able to pay my bills and save my home,  I was left with no retirement income.  

I understand Warren Buffet and a few others have done very well with the market and it’s the only way to real riches for me burned twice is enough. 

I do hope not all your eggs are in one basket. So when unpredictable events happen you can ride things out

You missed out on the best returns the stock market has ever seen by pulling out. 

FWIW, not all my eggs are in one basket. About 20% are in the "day trade" basket, which could be in a single stock for a given period of time. 

In reply to mtn :

Thats my point. Events beyond our control affect our lives. 

If I total money lost on the stock market it amounts to over 3/4 of a million. Almost a Million.  

Now to be fair only some of that money was from initial investment.  A good deal of it was profit from investing. Plus those sums are from 1987 and 2008/9.  

My real-estate investment was as erratic  but at least I  had a place to live which provided me with shelter.  

Food, clothing and shelter are the three requirements of life.  A healthy stock portfolio isn’t. 

So far so good. Approximately a 30% return since mid January. 

In reply to mtn :

That's impressive! Has to feel pretty good, but don't get cocky.