Starting out NOW

J

JoeAnonymous

Background:
I am an IT engineer, from Holland. I live in Asia and I have a full time job.

I "played" some stock before won big with the banking crises (buying banks when they were at their low) and lost even bigger buying uranium mining companies after the Japan nuclear mishap. Down 30K USD to 7.5K (still have it). That is a while ago.

I don't know quite how this got started, but I found this online website where you could analyse charts (with MACD, SMA, EMA) and kind of "play" some stocks historically. If you were "reading" the charts right you could consistently win, maybe, and bring up your capital. I don't remember the website, but it's not relevant to the story. Basically I spend (used to spent) a lot of time at work being bored.

That got me started on creating my own stock charts. I researched what software I could use for that, such as Excel or other tools. I tried various Java / JavaScript / other types of tools. I ended up with R. I think a module called quantmod. Well I worked on creating graphs and trying to see how to enter winning trades for a while. Then at some point I found something called quantstrat which is a module for R where they let you model quantitative strategies.

Basically R with quantstrat lets you download historical data for stocks or what not from Yahoo or Google and use whatever strategy you have programmed to invest this over the years.

Well that is kind of where I am at now, except that I think I found a strategy that is not too bad. I tried strategies where I would buy / sell stock very short term initially. That did not work out too well. I tried various indicators and variations of indicators and combinations thereof such as MACD, EMA, SMA, RSI, ATR, etc.. Then I switched to long term. I did find one long term strategy that was not too bad, but nothing impressive.

Now my latest strategy I'm pretty happy with. I would call it swing trading, but I'm not expert. I stay in stock anywhere from a few days to ... letting it ride for a long time with a trailing stop. It's swing trading because I buy when I think it's swinging up. I don't buy on the swing down. I don't short anything. I just stay long in stock that I think is doing good (or anyway the program thinks so). By the way I buy on market open because that's how quantstrat works the way I model things and I have a day job anyway.

Having said that it's all paper. I tried paper trading with Interactive Brokers. Happy with the experience over all, except that it's not accurate on paper. They don't have OPG (buy on open) for paper trading accounts and somehow my SELL LMT was being hit while actual stock never got that high that day, so I was making a lot of profit very quickly, which was definitely not realistic. I think that's how the paper accounts work.

I'm in a real account now, funded it with 42K USD (well my wife). I'm testing with 4K out of that 42K right now. One trade / stock is 200USD, so I can trade 20 at a time until I start losing or gaining.

That's where things are right now.

Oh and I don't look at graphs personally. I select stocks from all stocks on NYSE using a script. Then I do some extra selections (based on volume, etc..) and I end up with like 1600 stocks. I let my script take a random selection of those and run analysis and show me what trades it wants to enter. So it's all a script (R with quantstrat) and no emotion (I hope) until I start losing big I guess ;-)

First day of trading I entered like 7 stocks. I exited 2 automatically the first day (because of the stop). I bought another 6-7 today. I think it takes like a week or a little more for me to be "all-in", which means all the cash will be in the market around that time. Anyway it's just 4k right now.

I have huge hopes for my strategy, but well my strategy looks too good to be true. Basically no drawdowns (yearly) for last 16 years and very high annual return. But that's fine too right, if my real trading shows significant differences with what my program is showing me I can try to fix it and go back to the old drawing board. With my trailing stops my losses can't get very huge fast anyway unless I keep re-investing, unless market loses a lot overnight in all stocks.
 
Sounds like you did some back-testing.. got a good enough result to test forward live and now you have to see how it works out.

Sounds like a similar swingish strategy I started out on years ago (2009). I used to even hand draw swing charts for days on end (because of poor advice from a poor trading education firm) - surprisingly the swing trading system, if it got predictable entries and exits - worked very well. When put to work with Australian stocks though - the obvious swing stops and stop entries (beyond the swing extremes) got triggered all the time and burned me, which was when I realized for stocks at least - the market makers new where the retail orders' were and could screw 'em. I personally went down other avenues after I was watching a very basic DOM for the first time and saw a huge order that was spoofing the market.. now I just watch the DOM.

Good luck with the strategy though and obviously don't meddle with it until you have enough feedback. As you have pointed out - the entries and exits are key. Pro's know where retailer traders have their stops.
 
Good luck to you.

My advice would be start even smaller. If you have a system that is truly sustainable, what's the hurry? Start carefully and give it time to prove itself.

There is a lot to be said about automated trading. Fills, how you are going to deal with drawdowns, planning the exact criteria by which you will or will not intervene in your system, etc, etc.

There was interview on the Chat With Traders podcast that says a lot more a lot better than I can so you could check that out.

(Hope we are allowed to link to stuff we like, I am definitely not affiliated, I just like it)

http://chatwithtraders.com/ep-070-jeff-davis/
 
Trading individual stocks wrestles with "idiosyncratic stock risk" and sector specific risk. Each stock has it's own price ( trending ) behavior within the market's perception of it's expected earnings and reaction to those earnings. And the sector in which it resides, can be affected by sector specific news, which can consequently affect an individual stock's price behavior; many times negatively ( in a negative context ) even if a company is sound financially and has healthy expected earnings. This is where GTC stops can be hit and executed "unfairly". Over the short term, stock price behavior is dictated by buy and sell flow ( randomness ), Over the long term, it is expected earnings and, in many cases, the trend of the overall market and economy. It can be helpful to identify the overall market and economy's trend "risk" as many times. money is made by sitting on one's hands / sitting in cash ( market risk is "high" ).
In my experience, it has required less work and analysis in quantifying the risk and trend of the "market" and consequently investing in "indices" ( hundreds of stocks ) vs. wrestling with idiosyncratic stock risk, sector risk, and stop loss management involved with handfuls of individual positions.

James
Director of Quantitative Research
XOXOX
Boulder, CO

"- Open a Roth IRA
- Don't use leverage
- Don't quit your day job
- Don't be hostage to the markets
- many times money is made by sitting on one's hands"
 
JoeAnonymous said:
Background:
I am an IT engineer, from Holland. I live in Asia and I have a full time job.

I "played" some stock before won big with the banking crises (buying banks when they were at their low) and lost even bigger buying uranium mining companies after the Japan nuclear mishap. Down 30K USD to 7.5K (still have it). That is a while ago.

I don't know quite how this got started, but I found this online website where you could analyse charts (with MACD, SMA, EMA) and kind of "play" some stocks historically. If you were "reading" the charts right you could consistently win, maybe, and bring up your capital. I don't remember the website, but it's not relevant to the story. Basically I spend (used to spent) a lot of time at work being bored.

That got me started on creating my own stock charts. I researched what software I could use for that, such as Excel or other tools. I tried various Java / JavaScript / other types of tools. I ended up with R. I think a module called quantmod. Well I worked on creating graphs and trying to see how to enter winning trades for a while. Then at some point I found something called quantstrat which is a module for R where they let you model quantitative strategies.

Basically R with quantstrat lets you download historical data for stocks or what not from Yahoo or Google and use whatever strategy you have programmed to invest this over the years.

Well that is kind of where I am at now, except that I think I found a strategy that is not too bad. I tried strategies where I would buy / sell stock very short term initially. That did not work out too well. I tried various indicators and variations of indicators and combinations thereof such as MACD, EMA, SMA, RSI, ATR, etc.. Then I switched to long term. I did find one long term strategy that was not too bad, but nothing impressive.

Now my latest strategy I'm pretty happy with. I would call it swing trading, but I'm not expert. I stay in stock anywhere from a few days to ... letting it ride for a long time with a trailing stop. It's swing trading because I buy when I think it's swinging up. I don't buy on the swing down. I don't short anything. I just stay long in stock that I think is doing good (or anyway the program thinks so). By the way I buy on market open because that's how quantstrat works the way I model things and I have a day job anyway.

Having said that it's all paper. I tried paper trading with Interactive Brokers. Happy with the experience over all, except that it's not accurate on paper. They don't have OPG (buy on open) for paper trading accounts and somehow my SELL LMT was being hit while actual stock never got that high that day, so I was making a lot of profit very quickly, which was definitely not realistic. I think that's how the paper accounts work.

I'm in a real account now, funded it with 42K USD (well my wife). I'm testing with 4K out of that 42K right now. One trade / stock is 200USD, so I can trade 20 at a time until I start losing or gaining.

That's where things are right now.

Oh and I don't look at graphs personally. I select stocks from all stocks on NYSE using a script. Then I do some extra selections (based on volume, etc..) and I end up with like 1600 stocks. I let my script take a random selection of those and run analysis and show me what trades it wants to enter. So it's all a script (R with quantstrat) and no emotion (I hope) until I start losing big I guess ;-)

First day of trading I entered like 7 stocks. I exited 2 automatically the first day (because of the stop). I bought another 6-7 today. I think it takes like a week or a little more for me to be "all-in", which means all the cash will be in the market around that time. Anyway it's just 4k right now.

I have huge hopes for my strategy, but well my strategy looks too good to be true. Basically no drawdowns (yearly) for last 16 years and very high annual return. But that's fine too right, if my real trading shows significant differences with what my program is showing me I can try to fix it and go back to the old drawing board. With my trailing stops my losses can't get very huge fast anyway unless I keep re-investing, unless market loses a lot overnight in all stocks.
More...

Now, do not feel put back, and nothing personal, but I predict now that you are wasting your time, especially if you are not going to "watch" the markets every single day.

There are no magical indicators or systems that can accurately predict market movements - if you think of how people worshipped false gods (now that in itself is ironic) centuries ago, then you might get a better understanding of why "systems" will never work "all of the time".

However, it is possible to "win" each and every day, week, month, year, that you trade, but it requires hard work and effort, same as any other high paying job!

I hope you prove me wrong, but I am about 99.999% right, so, your chance is not that great.

J_S
 
JoeAnonymous said:
I'm in a real account now, funded it with 42K USD (well my wife).
More...

Huge mistake. Return the funds to your wife until you can fund your own ride, and are a proven winner.

Why put her hard-earned money at risk so that you can prove something to yourself?

Quite frankly, there's nothing in your post that gives me any confidence that you know what you're doing, and you shouldn't ask her to jump in your boat.
 
Last edited:
lindq said:
Huge mistake. Return the funds to your wife until you can fund your own ride, and are a proven winner.

Why put her hard-earned money at risk so that you can prove something to yourself?

Quite frankly, there's nothing in your post that gives me any confidence that you know what you're doing, and you shouldn't ask her to jump in your boat.
More...
that's true, traders should make money by compounding from a small amount, and never never borrow money to fund their accounts.
 
I have been doing a lot of back-testing, but I have already found an issue with that. With the live trading I am "exiting" trades before my back testing software exits trades. My trailing stop is being hit (rightly so as far as I can see), but somehow my backtesting software (quantstrat) is not hitting that stop. I think it may be an issue with it not hitting on intra-day Low for some reason especially on the first trade day. I googled a bit and I think it's a problem with the software and/or my usage of it, so I need to investigate.

It might be a variety of things and I need to dig into that. Which will take time. But in the meantime I will continue with the strategy. When I get my software to behave exactly like the real world I will have to run more backtests to see that all the variables are optimal (such as point to exit, etc.).

As regards to some of the people that suggested do not use other people's money, well my wife's money is my money too. And my money is hers. I did say she funded it, because here in Asia she keeps most of our savings as she's better equipped to deal with the local banks and what not. But I consider the money mine as well as hers. So I kind of put everyone on the wrong foot about that, sorry! We have other savings as well and even if I lose the whole 40K it's a substantial loss, but life will go on. We both have decent jobs. I am using 40K (or 4K right now) as it is the same as the initial equity I have been back testing with. I asked my wife to transfer 42K to allow for slippage and testing and what not (which I am doing now).

Ideally if the system works, I would not interfere ever. Max drawdowns should be less than max market drawdowns with the trailing stops. But I am back IN and back IN very quickly so theoretically I could draw down a lot more. Back testing for 2008 shows that. But year over year results are much better than market and always positive (for last 16 years).

Some people suggested looking and researching markets and sentiment and performance. I think my strategy performs better when markets are volatile. Back testing shows huge positive results for some of the most volatile times over the past 16 years. Again, I have to go back to the drawing board with trailing stops not triggering the same live as in the software, but I will continue the strategy at 4K capital for now as is. As to research and having a feel for the market, that is not my strength I do not think and it would probably take me 2-20 years to make it my strength. I think those times are over too. With more computerized and automated trading I think the market is becoming kind of like a roulette table, more random and random. I think my only way to figure out a way to beat the market is to let a computer calculate various scenarios (based on some guesses I make, which admittedly do use indicators) and just see how that would perform for the past X years, and then use that system for today's market. One concern I have with that is markets are changing and becoming ever more random (I feel).

I guess there are two major concerns right now:
1. my back testing not reflecting reality
2. past results being no guarantee for future results

Number (1) I need to address to my satisfaction before I start trading more. Number 2, well I do not think any of us can ever completely eliminate that factor.

Oh one other thing, I did not quit my job and I do not plan to. Not sure why my thread was moved to career trader!
 
Sounds like you have a good plan going on. Don't mind the other traders here. I think you're starting off 100x better than most new traders, and if you want to be an algo trader then so be it. Just remember, like you said, the market is becoming more computerized, so it's your robot vs theirs.
 
You must log in or register to reply here.
Back
Top