Anyone Tried Day Trading After Retirement?

[ESRwannabe]

That sounds more like value investing than day trading to me.



When I played with individual stocks, I would discount my expected future value (typically conservatively estimated future EPS along with a conservative future PE ratio) by a desired rate of return (15% in a low inflation environment) and would buy companies I felt were relatively safe (good established business and not highly leveraged -watch those balance sheets). I did well and beat the market over the next 2 years on almost all my purchases but was not disciplined at selling my winners and rode a few down losing some of my gains. I then made myself a rule to sell half when it doubled if I still liked the stock. -Then I felt I missed out as they kept rising but at least I didn't ride them down! I only own one stock at this time and am liquidating it for living expenses first. I've "proven" my analysis to work well but don't have any false security that would be able to beat the market consistently over a longer period and would have long periods in cash when I couldn't find anything I liked that was undervalued.

Maybe 2020 to 2022 was an exceptional time for finding mispriced stocks, with the two large crashes and the giant rally in between them.

 

[NW-Bound]

...
I basically doubled my taxable investments ($500k to $1M) from 2020 to 2022 primarily due to trading. Which is why I am stopping and have moved it all into etfs. I have won enough that I'd rather re-allocate for less risk and more tax efficiency.
...
I started trading in 2020 and I basically matched the S&P 500 in 2020 and 2021 (when growth stocks had a "perfect" environment), and then in 2022 I have done much better than the S&P 500, YTD returns of 29.48% vs -13.26% as of 12-3-2022.
...
I'm now getting roughly $30k in dividends from the etf SCHD.

https://www.schwabassetmanagement.com/products/schd

That's impressive.

You make me look at my portfolio.

From Jan 1, 2020 till now, I am up 41%.

Quite meager compared to your 100% gain, but still OK compared to the S&P number of 36%, particularly as I was never 100% in stock.

 

[RetiredAt49]

Any update for the past 3 weeks?

-ERD50

I did very little trading because we spent 10 days at the cabin…. But my original $500 account is now $1,291.32.

I just hooked up my software to TD Ameritrade’s API and have streaming stock quotes, level 2 data, etc. flowing into my program but I have many weeks/months before it’s ready to place trades automatically because I’m still working on writing the code for the core algorithm. Once that’s complete, I’ll back testing it across many equities and FOREX pairs.

Meanwhile I’ll still place some manual trades using ThinkOrSwim’s platform.

 

[RetiredAt49]

Brief update for those who’ve asked…

I have flushed out and finalized my version 1 algorithm which leverages ideas from patents I created for a different vertical. I then back tested my algo using both commercial and open source back testing stock market frameworks. Thus far I have tested data across every S&P 500 stock ticker including some ETF’s like VOO, VTI, UPRO, QQQ, TQQQ, SQQQ, etc.

Excluding a few individual stocks that have huge daily ask/bid spreads, over any rolling 6 week period (including the bear market we are in) my algo generated positive returns for 98% of the 500+ equities I tested. The average 6 week return across all datasets was 57.3% with an average max drawdown of around 3.7%. The max return in last 6 week period was 281.93% on dish (haven’t investigated why that one was the highest). What I also noticed is that out of the individual equities that have loss money using a buy and hold strategy, my algo generated positive returns for 99% of them.

I’m now working on plugging this algo into my trading bot I’m working on and plan to have a version 1 running within the next couple of weeks and running against 1-3 ETF’s but on using a fraction of my day trading brokerage account.

I anticipate I’ll get some slippage (the time it takes to execute trades once a signal fires) as my algo is designed to jump in/out (either long or short) in seconds/minutes and occasionally hours if the stock has a really strong trend.

 

[ERD50]

Brief update for those who’ve asked…
....

I anticipate I’ll get some slippage (the time it takes to execute trades once a signal fires) as my algo is designed to jump in/out (either long or short) in seconds/minutes and occasionally hours if the stock has a really strong trend.

Interesting, hope it keeps working for you.

With those fast moves, I hope this is in an IRA/401K or something that won't require tracking the trades for tax purposes.

-ERD50

 

[RetiredAt49]

With those fast moves, I hope this is in an IRA/401K or something that won't require tracking the trades for tax purposes.

-ERD50

Tracking the trades… for what purpose other than performance/return?

My bad… missed “for tax purposes”… yeah, I’m not trading same equities as my long term portfolio (wash sale rules, etc.) and this account is listed as a PDT (pattern day trading) so year end statement will have capital gains/loss

 

[finnski1]

Tracking the trades… for what purpose other than performance/return?

For tax purposes as he said. Could be tons of trades in a taxable account. Could end up being lots of work.

 

[MarieIG]

Glad to hear its going so well!

 

[RetiredAt49]

Quick update from me (as requested)…

My version 1 algorithm and trading bot are complete… i have back tested it against a ton of historical stock market data and across all of the individual S&P 500 tickers, various ETF’s, and some small/mid cap stocks.

The TDAmeritrade API platform has been a PITA to work with but I had my v1 bot trading live yesterday for the first time (just a few shares of a single stock). My multithreaded , multi-stock algo v2 bot is nearly complete and I’m really excited for this one because I’ll be trading between 10-30 stocks/ETF’s in real-time because I like to diversify my strategy just like I do in my long term investment accounts.

We are headed to Alabama and Louisiana next week for about 10 days for some volunteer work but I will post here again once I return and have time to run/verify that it’s working well and I’ll unleash it using a larger brokerage account (e.g. $30k or so) and report in.

 

[Sunset]

^^^^ thanks for the update, sounds exciting.

 

[OverThinkMuch]

An entirely separate account sounds like a really idea to limit trading activity.

My version 1 algorithm and trading bot are complete… i have back tested it against a ton of historical stock market data and across all of the individual S&P 500 tickers, various ETF’s, and some small/mid cap stocks.

I'm curious about the steps you took to get historical data and simulate trades. What website's historical data did you use?

Most stock price data I've seen only show open, close, high and low. I imagine you have data with bid-ask spreads throughout the trading day?

 

[RetiredAt49]

I'm curious about the steps you took to get historical data and simulate trades. What website's historical data did you use?



Most stock price data I've seen only show open, close, high and low. I imagine you have data with bid-ask spreads throughout the trading day?

TD Ameritrade (my day trading broker) has a ton of historical data going back to at least 2009 and available via API’s…. I also used Kaggle and other data science platforms (e.g. QuantConnect).

For simulating trades I use Backtesting.py, VectorBT, QuantConnect, and BackTrader.

My backtesting framework pulls any/all historical stock data and runs it through my algorithms, which will eventually feed recommendations to my live trading bots but that’s a little further down the road as I’m just manually “instructing” my bots which stocks to trade

 

[RetiredAt49]

Forgot to answer the other question… the TD Ameritrade API’s let me stream level 1 and 2 data so I have access to things like bid/ask spreads… order book details, etc. and my algo uses all of it to automatically place long and short trades all day long…

I’ve tested my algo’s against forex pairs (and it’s done extremely well) but for now just focusing on stock/equities.

 

[SunGuy]

Not retired, but I have been trading forex for quite a while, 8+ years now.

 

[RetiredAt49]

Well a couple of weeks ago I officially turned on (full blast) my automated day trading algorithm and it was running great until TD Ameritrade contacted me and informed me that I was placing too many trades per day on their platform. Long story but neither their online material nor developer API documentation suggested they couldn’t handle lots of orders/transactions per day. Within 1.5 weeks my algo raked in just under $6,000 in net gains (and only using a $25k brokerage account and no margin). I must admit that it has been a blast building it, watching it run, automatically placing trades, and making some amazing returns regardless if the market each day is going up or down.

I’m actively swapping out TD Ameritrade with another brokerage platform that has no limits and is commission free. They also have real time streaming tick data which I’m looking forward to leveraging too

 

[cyber888]

Well a couple of weeks ago I officially turned on (full blast) my automated day trading algorithm and it was running great until TD Ameritrade contacted me and informed me that I was placing too many trades per day on their platform. Long story but neither their online material nor developer API documentation suggested they couldn’t handle lots of orders/transactions per day. Within 1.5 weeks my algo raked in just under $6,000 in net gains (and only using a $25k brokerage account and no margin). I must admit that it has been a blast building it, watching it run, automatically placing trades, and making some amazing returns regardless if the market each day is going up or down.

I’m actively swapping out TD Ameritrade with another brokerage platform that has no limits and is commission free. They also have real time streaming tick data which I’m looking forward to leveraging too

So how does your algo interface with TD Ameritrade?

 

[RetiredAt49]

So how does your algo interface with TD Ameritrade?

Via their API’s

 

[sengsational]

It's admirable, you putting your design and build skills to work like that. Any smothering skills I have get put into apps I'm lucky to give away, hehe! Given you're using their API, you can code. Was that your craft before retiring? We're you in financial or other?

 

[RetiredAt49]

It's admirable, you putting your design and build skills to work like that. Any smothering skills I have get put into apps I'm lucky to give away, hehe! Given you're using their API, you can code. Was that your craft before retiring? We're you in financial or other?

I was a software engineer specializing in software algorithms/patents/designs for about 25 year so I enjoy this kind of stuff as it keeps my mind/skills sharp.

 

[Koolau]

Forgot to answer the other question… the TD Ameritrade API’s let me stream level 1 and 2 data so I have access to things like bid/ask spreads… order book details, etc. and my algo uses all of it to automatically place long and short trades all day long…

I’ve tested my algo’s against forex pairs (and it’s done extremely well) but for now just focusing on stock/equities.

Heh, heh, like they say. It w*rks until it doesn't. Best luck!

 

[MarieIG]

Well a couple of weeks ago I officially turned on (full blast) my automated day trading algorithm and it was running great until TD Ameritrade contacted me and informed me that I was placing too many trades per day on their platform. Long story but neither their online material nor developer API documentation suggested they couldn’t handle lots of orders/transactions per day. Within 1.5 weeks my algo raked in just under $6,000 in net gains (and only using a $25k brokerage account and no margin). I must admit that it has been a blast building it, watching it run, automatically placing trades, and making some amazing returns regardless if the market each day is going up or down.

I’m actively swapping out TD Ameritrade with another brokerage platform that has no limits and is commission free. They also have real time streaming tick data which I’m looking forward to leveraging too

Thanks for the update!

 

[Car-Guy]

Call it want you like, "New Years Resolution" or "Giving it up for Lent", but I've stopped all day/swinging trading, at least for now. Too much like work and too little in "potential gains" and too much in "potential losses". With fixed interest rates continuing to rise, why not go for a sure thing.

 

[target2019]

I was a software engineer specializing in software algorithms/patents/designs for about 25 year so I enjoy this kind of stuff as it keeps my mind/skills sharp.

Here's an upvote for your ambition. Sounds like fun.

 

[RetiredAt49]

Call it want you like, "New Years Resolution" or "Giving it up for Lent", but I've stopped all day/swinging trading, at least for now. Too much like work and too little in "potential gains" and too much in "potential losses". With fixed interest rates continuing to rise, why not go for a sure thing.

There’s no way I would sit behind a desk and day trade manually (which would be a full time job) - which is why I’ve built a full automated day trading algorithm that runs in AWS. Like I mentioned previously, I had to alter my code to a different brokerage platform/API because TD Ameritrade’s API platform couldn’t handle the volume of trades my algo performed daily. I’m hoping to have the swap done within a couple of weeks but we’ve been vacationing a lot lately.

 

[RetiredAt49]

Well it’s been a PITA transitioning my algo trading bots to use multiple brokerage platforms (had too because the first brokerage platform - TD Ameritrade - couldn’t “handle” the volume of trades) but v1 is finally complete.

These algo bots (which mostly just scalp individual tickers within the S&P 500) have performed thousands of full automated trades and the account has grown from an initial account size of $25,000 (to be eligible for the pattern day trader SEC rule) and it’s now just over $150,00 (with no margin/leverage) and it has only run fully autonomously for a couple of months (with some tweaking here and there and some significant down time). All those patents I did (working for multiple Fortune 500 companies in non financial sectors) are now personally starting to pay off for me personally besides just have plaques on the wall [emoji16]

One thing to note is that most of the results and videos you’ll get while searching Google/YouTube on “how to day trade” or “best technical indicators for scalping” is all a bunch of [emoji90]. In addition to my algorithms, I fully and programmatically backtested nearly every technical indicator/strategy found on Google/YouTube and most of them do not work unless they are significantly adapted in dealing with various intraday market conditions (up, down, and/or sideways). BTW - tight stop losses are a game changer (when done programmatically and not via a preset value). My next version will
Incorporate TQQQ and SQQQ - they are both triple leveraged ETF’s and are solid strategies if unsure which way the market is going to turn (e.g. in a sideways market). Backtesting a decade worth of those still but results look very promising.

YMMW