Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
A Small Algo Lesson
Last night I decided to add another instrument to my current live algo 'stack': micro gold (MGC), in the spirit of diversification. It like it, performs very well with the current algo I selected and I have plenty of margin available. It is a low risk, high reward strategy. When I added the instrument and strategy last night, I ended up with a bunch of trades that I probably should not have had. The same happened on the forex pairs I am testing. Let me explain why I believe this happens and how to prevent it.
First, consider that I am a medium frequency trader, I am running strategies 23 hours per day (24 for forex), and for this week I am averaging about 1 hour and 35 per minutes in each position. I have written earlier about 'normalizing' chart data, primarily indicators/studies, but also any specialized chart (Range, Renko, etc.). I'm not sure why, but if I start a strategy mid-week and mid-session, I start with some bad trades. After that, everything seems to normalize. I typically do not see this Sunday night, though I put in a delay (10-15 minutes) for my trading to commence, to allow for any gaps from the Friday close. Sometimes I will use number of bars, if using an indicator that looks and the last n-bars (e.g. 14 bar ADX).
By starting the micro gold early (2100 ET last night), I probably gave away some money on bad trades. Everything normalized by 300 ET this morning, and my Strategy Performance Report entries now match (give or take a little slippage) my actual entries and exits.
The goal is to prevent leakage of the strategy, and I believe starting at the beginning of the week addresses this issue. Lesson learned.
Today's chop just made any chance of profitability a tough proposition. Drawdown was not bad, but still couldn't break above zero. My execution was fine (I didn't screw anything up). I had some extra trades that I am researching, so execution was not perfect. The extra trades appear to be during consolidation.
I am happy thus far with our live systems, though they are picking up some stray orders. We will be watching this carefully. We are not seeing this issue with our forex orders, but probably because forex moves like molasses in Alaska in winter.
Forex Goal
I have set a forex goal for my strategy, running these three pairs: EUR/USD, GBP/USD, EUR/GBP. The goal is to return my practice account to $100,000, or about $2500, before (or if) I start trading live. I'm trading 100k lots. This is full auto, so I need to execute (meaning, not screw it up by fiddling with anything) and stay connected.
Forex is just a side project for now, and if it generates a little revenue, that's okay by me.
This is one of our strategies today on micro gold (MGCQ20), on full-auto:
This is a typical day, if there is such a thing. Lots of give and take, hopefully coming out on the right side of in the end. Yesterday was worse, today is better, and tomorrow?
Profitable for the day, trading MNQ and MGC; both instruments profitable
43 round trips
42% profitable
1.45 profit factor on closed positions
2 open position at close: one profitable/one not
Profitable for the week (3 of 4 days)
Today was a wild ride. Micro gold did very well, all things considered, at a 50% win rate. That is good, since I got off to a rocky start with it. Micro NQ did okay, too, though we gave a lot back this afternoon. Overall drawdown was not bad (~3.5% of initial capital).
Execution
Things were a little wonky in this arena. I thought that my strategy had missed a very important reversal, so I intervened. It just happened that my indicator setting did not match my strategy, but I thought it was a strategy glitch. In the end, the strategy was okay and the indicator was wrong. Anyhow, no harm the first time and I was able to pickup a few extra $ on the second time (by the time I caught my own error). So I give myself a 'D+' for execution today, and only a + because I was able to squeeze out a few extra $$.
System execution was fine, but TradeStation got weird. I am showing an extra position here that does not exist:
Extra Trades
We figured out the extra trade problem and are trying to code around it. Since I am using Renko on this particular strategy, we need to account for the the price action above or below a Renko bar. Unfortunately we cannot see it on the chart. If you are not familiar, here is Renko in Sierra Chart, showing wicks:
Renko with Wicks
I think we have figured it out (and by 'we' I mean my brilliant wife/trading partner).
Forex
Things were just weird with this today. Testing continues, but we are moving sideways. I have not had time to check transactions against the strategy performance report, but maybe this weekend.
Profitable for the week, profitable for the month (after data and all other fees)
I don't trust the reports, so I am not posting win % and # of trades until I have full details
End of week automatic exits worked fine
Overall, MNQ did very well; MGC did not do so well, ending negative for the week. Though I started MGC on Tuesday evening, I don't think it would have made any difference starting Sunday night. I have a better way, I believe, to trade this, so I will have run it next week. MGC was not terrible, but it was not profitable. There were plenty of ups and downs to keep things interesting.
Execution for the week was mostly clean. Not perfect, but pretty close. I made a couple mistakes Thursday, but was otherwise able to let everything run. The systems ran fine, though we do generate some extra (though expected) orders that we need to deal with.
The virtual machine, which runs our live strategies, was perfect.
Forex test
This testing is a nightmare right now. I don't have the time to dedicate to it and kept goofing up my positions. I will run a set starting Sunday night again, using a basket of pairs (not pears) and just let it run. These are my pairs for next week:
AUD/USD
GBP/USD
EUR/USD
EUR/GBP
I am lazy to find any pics to share, so here is my Forex screen (MultiCharts) today:
You will see I have two positions open over the weekend. Oops.
Other Accounts
I worked to get my Roth IRA setup with TradeStation. Figuring out our architecture for trading (professional and personal) has been a little challenging, but @SMCJB and @kevinkdog helped me a lot in figuring out my data fee dilemma (thanks guys!). The main thing is that we need to keep trading live, but also developing strategies, without paying through the nose for data. I think we're good now.
So going through our results this week, I realized that I had a bug in my fractal code function, which affects all strategies and indicators that use it. Who knew that addition and multiplication were different!
Thankfully, finding this error actually helped me to (potentially) find a solution to a much bigger problem.
Ultimately it would have not impacted our trading this week, but it was a stupid oversight. Lesson learned.
Post-mortem from last week on the strategy I'm running on MNQ caused me some consternation. We were nicely profitable for the week. However, I'm looking at the trades I had versus the trades I should have had, and I think I may have been lucky. As Bill Williams famously wrote/said: even a blind chicken finds some corn eventually. I had a whopping 40% more trades than I should have.
I believe luck is where opportunity meets preparedness. (Thanks for the words of wisdom, Dad!) So I put my head down and worked on some of this over the weekend, creating my own luck. This spawned a few solutions, so we are testing those.
In testing these (in simulation) today....
MNQ did well
Silver did well (made a mistake on chart settings)
NQ did horribly (should have tracked with MNQ, but my chart settings were incorrect)
I am starting a new cycle at market open, hoping we have eliminated the extra orders. I will report tomorrow on the results.
Here is my journal entry of the week. I'm not journaling this week.
Seriously though, I am taking a short break from journaling, mainly because it is more of the same that I'm doing, but also I'm taking a couple weeks off and just letting the robots run the shop. I will sit near a large body of water and hang out with the family, and that is about it.
I'm back! I returned yesterday from a nice 15 days away from trading and other work. It was a much needed reset. We were staying at Lake Norman in North Carolina (USA), which is a beautiful man-made lake in roughly the center of the state. I am fortunate to have a family member who owns a house there, who was generous enough to block out his AirBnb rentals so we could stay. We had an absolute blast. If you need a getaway, Lake Norman has a lot to offer.
Trading
I started trading the Sunday before we left, and weathered the tropical storm early Tuesday that brushed by us and stirred up a nasty tornado in our area. This is where having the VM running my automated strategies was very useful..... I wasn't worried about a power or internet outage (we had a power line down in our neighborhood when we left for NC). I was able to safely exit my positions late Tuesday morning. It was a profitable day and a half and was focused solely on metals (precious and base). This is mainly because metals work well with one of my algos.
It was tempting to continue trading through vacation, but I vowed to not touch anything until today.
I don't like starting to trade in the middle of the week. I may be leaving some money on the table, but I prefer to look at my performance by week, rather than by day. If I have a two day skid, I'll just be annoyed all weekend.
Forex re-re-re-revisited
I'm an idiot. My wife/trading partner has been telling me how profitable her forex trading has been, but as I recorded earlier in this journal, I have not been able to get forex to work and my results have been less than stellar (read: unprofitable). Well, it seems that were were using slightly different pieces of code. The difference is around Boxsize, which is the size of a Renko box. EasyLanguage (TradeStation) and PowerLanguage (MultiCharts) implements this a little bit differently.
Long story short, they don't behave the same way. Unfortunately, my code was slightly different than her code. My code was wrong, now that I look at it. I don't know how we got out of sync, but we don't really have an effective method for version control (think SVN or Git). I actually run an SVN server, but haven't implemented the clients. My bad.
Anyhow, the problem is fixed and now I can start testing on some pairs. If all goes well, I may start trading live in a week or two.
The question may arise: why trade Forex if you can just trade currency futures? Answer: for whatever reason, I have an algo that does not work as well on currency futures. I cannot explain why. It could be differences in platforms (TradeStation versus MultiCharts), but personally I don't care. If I have a profit center that happens to be some forex pairs, then I'll trade them, even if the lots are small. I think of it in terms of the old bitcoin faucets.... earning by drips.
Trading forex pairs continues the theme (and goal) of diversification that I have been writing about for the past couple months. Hopefully it works.