So, as the video says, I turned on an autotrading strategy on ES today. Within 2 hours, it had made me 9.75 ES points. I know that I technically should let it keep going, but I just couldn’t do it. I’m a puny weak human, after all. So, I turned it off and will just watch it for the rest of the day.
Needless to say, I am very happy with my first day of autotrading results.
And, if you have been keeping up, you know that I have been fretting about the stops required to take the signals. Via backtesting, I was able to find a balance that had good profits but also limited my max stop loss to 3pts ($150/contract). So, I am pretty comfortable with it, for now. We’ll see how trading goes next week. I plan to do a lot more testing over the weekend, and see what tweaks I can make, if any.
[EDIT: I can see now in the video that it actually went down a little more than 1 point... when trading I use a fixed scale on my chart so that 1 inch = 4pts at all times. It keeps me from being fooled by scaling differences in wide range vs. narrow range time frames. But I usually turn that off for the videos so that they are clearer. So, I thought I was looking at about a point but the scale was different.]
damn! that’s impressive
Sweeeeet!!!!!!!!!
kaaaa-ching
LMAO Rich and then disappear LOL Awesome!
Richard,
Good job Today!! That was cool… Quick question I got The box indicator form Glen and I tried to add the show me indicator on a stock, but it did not work? Do I have to change something on the parameters? can you help me out on how to setup?
Later,
Alex
Hey Alex. For stocks, I use settings: Lookback: 20 or 30 … basically whatever you want. Tolerance: 0.03 so that the boxes can be off by +/- 3 cents. MinSize (or whatever that other parameter is called): 0.25 so it finds boxes 25 cents wide or wider. The default settings are more appropriate for ES, and I suspect the reason you see no boxes is because on ES you have to have 1.5 point boxes or more. Hope this helps.
daayaannng Richard, I was watching your second trade, and I realized it was the same trade that I took! I took that exact same short (but got in earlier @1466.25), but got scared out for 1pt of profit. I am already jealous of your ‘Bot, maybe I should consider letting it autotrade for me……
Oops, I went back to review my chart, and it turns out I actually didn’t take that short trade (I took a long just before that), but I did spot the short that your ‘Bot took, I just didn’t believe the setup.
wow, seems great. Richard did you code this yourself.. I have a simple system that I have been wanting to backtest btu I am not a coder/exceler :P so maybe we could work on something and did you test your bot out on intraday data at all before going live with it? I use sierracharts and you can backtest in that let me know if you need some statistics with your bot.
Hi Richard, the Trading Goddess referred me your way. I am very interested in developing an automated system for my trading method. I was wondering if you had ever heard of http://www.smartquant.com or know of other automated/algorithimic trading platforms that could work with Interactive Brokers API?
Daytrading: Yeah, coded it myself. I’ve been programming computers a LOT longer than I’ve been trading. Yep, backtested on 6months of tick data, which is all TS can give me. If I wanted to backtest on minute charts I could go back years, Maybe I’ll do that just for fun, but I’m not sure it would be very meaningful.
Tom, You might want to look at Multicharts at tssupport.com. It’s essentially a fork of tradestation with slightly different features. I think it interfaces with IB. If you have any programming experience, the IB API is actually pretty easy to work with.. so you could just write a standalone program. I have written at least one IB add-on tool in the past without much trouble. I have no experience with openquant, but it appears that it is built around c#, so you will need a certain amount of programming sophistication to use it.
Congrats on the system. :) With your programming background I think it’s right up your alley.
I’ve looked at a lot of systems lately and my opinion is that any system with a max DD of 25% or less is okay. You seem very risk averse so I am not sure if you’ll ever be comfortable with the long-term performance of automated systems (even the ones that you write).
Two more websites that I find interesting. They are both mathematicians that have written trading systems. I’m not endorsing them but the first one, in particular, may be appealing to someone good at math (that’s not me. lol.).
http://www.meyersanalystics.com
http://www.breakoutfutures.com
Oops. That should have been http://www.meyersanalytics.com
I was surprised by your comment (in the video) about having to test your automated system with real money. I’ve always been lead to believe that automated systems could be paper traded as well as back-tested, but I’m not a Tradestation customer yet.
Michael: Tradestation can backtest, and you can have the strategies on for forward-testing without sending any actual orders. But, if you send an order it has to be real. So, the only way to play with their order matrix, or test out things like manually adjusting strategy orders, is to put real money on the line. A sim mode has been one of the top enhancement requests for tradestation for years now. They just don’t seem like they want to do it, for some reason.
Thanks for the links, I’ll check them out.
I personally think max drawdown is not such an interesting statistic… that’s just how much was lost over some past time period. I would rather predict a theoretical likely future drawdown via other information, and make sure I’m comfortable with it. For systems that trade intraday for a few minutes at a time, a 25% drawdown is ridiculously bad in my opinion. This is just my sensibilities, but to me the whole point of daytrading (and especially scalping) is to give up some of the potential upside in exchange for a relatively smooth equity curve. If I were happy with losing 25% of my account sometimes, I’d save myself the effort and just trend-follow on daily charts.
[...] Richard is “The Man,” so he was able to create the machine! [...]
> a 25% drawdown is ridiculously bad in my opinion
heh. Most of the stock systems that I’ve created on Portfolio123.com (and the systems at minanalyst) will have intramonth DD’s of up to 25%. The guy at minianalyst actually recommends buying into his systems when they have a down month. If you look at his historical stats you will see that the month after a down month often averages 2x the monthly average. It’s like buying low and selling high in an individual stock, I guess…
I prefer to think in terms of months rather than days when it comes to DD. I’ve played with a lot of stop-loss settings on Portfolio123 and optimal was 20% from the 12-month high for the small/midcap stocks that I trade. I keep three months of living expenses in cash so daily draws really don’t matter to me. Your money management strategy may be different, though.
About buying in after a down month… I gotta say… Isn’t that like saying “I haven’t rolled a six in a while, so I’m more likely to roll a 6 next time?” Unless there is some trade-to-trade dependency in those systems, I don’t think there’s a valid statistical reason to believe that it will outperform after a down month. Right? Any system can draw straight down to 100% loss without ever winning again, it’s just a matter of how likely or unlikely it is…
I just know that, trading manually, I have never had anything like a 25% drawdown on any timeframe. I don’t plan to start having them now! If it’s an unavoidable characteristic of mechanical systems, then I’ll play the diversification game and split my capital until the risk is acceptable again. But, I hope it’s not an unavoidable characteristic. We’ll see!
Richard. In the video I hear you say that TS doesn’t have a ’simulation’ mode. I was wondering if you tried the ATI (advanced trading interface) that is available through NinjaTrader(NT). Using NT as a pass-through you could basically forward test your strategies without risking real capital. Do you have any experience with NT? I tried to configure the ATI between TS and NT last night but spinned my wheels…
Also you mention fib retracement levels, what indicator do you use for this? Does it plot fib confluence areas/clusters?
Hi David. I do have ninja, but have never tried to make an ATI link-up work. I know it’s easy enough to make sim trades via ninja’s interface by sucking TS data into ninja. At any rate, I’m over the learning curve now for TS, so I don’t really need that anymore. It’s just scary when you aren’t sure how something works, and you have to use real money to find out.
I’m not sure what I said in this video, but I probably was talking about the eotpro.com fib extensions indicator. It locates fib extensions that you specify from recent swing highs and swing lows and alerts you to them. It doesn’t pay special attention to confluence, but it does have its own way of discriminating between possible short-term highs/lows and possible longer-term highs/lows.