Chabah on Automated Trading

[JimmyJam]

Quote from Chabah:

I think the trading action before an announcement will indicate the direction of the gap around 50% of the time or more, so I am not looking to avoid announcements or events.

Yeah, keep everything totally automated.

This is pretty amazing, considering the fact that you don't have much (if any) actually discretionary trading experience, but rather with the skillsets you've acquired as a programmer and poker player you were able to put together the puzzle pieces fairly quickly.

What does that tell you? ... the rules across the board are the same. Trading is, in its essence, a huge math/logic game - which you were able to figure out by putting together the discrete pieces and quantifying them to the level of "0's and "1's", and breaking everything down to its essential "unit".

But don't rest on your Laurels here ... there's a whole world of futures out there waiting.

Good (no great) trading,

Jimmy Jam

P.S. I'm inadvertently modeling your trading with my own, it's been weeks and months since your broke your system/methodology down, and looking at what I'm doing with the futs now I can see how I've actually mirrored your process - which I thought ingenious when I first read it.

 

[Chabah]

Thanks JJ, I appreciate the encouragement.

I am posting mid-month here because the system happens to be very clean right now - no open P/L, just one position open. So, I will post a couple attachments (1 per post) and describe how I see each picture.

First I should note that my correction to end my last bar at 3:58 so that my end of bar orders execute before the close is working. However, there was a new error where a position bought at the end of the day (as intended) and then bought *again* on the open. That is some sort of execution error, and I need to figure out how to prevent it. In this case it cost me about $230 or 3R, so that kind of sucks. Of course if luck were with me it could have boosted profits, so I can't complain too much.

Here is that chart, on CROX.

No for some reason this isn't rendering quite right, so I will narrate. The bottom long blue line - just ignore that. It's not right at all, because the 125 share position was closed and reversed on 2/8 at 3:30. Also missing is the buy that occurred on 2/12 on the last bar. This is part of the problem - the order executed but even TradeStation doesn't recognize it - very strange.

On 2/13 you can see the buy on the open ALL the way at the top of the bar - that order should never have been placed. The only order *should* have been closing the position using a trailing stop somewhere over 55. The closing arrow is much lower because it is an average of the two orders, and I got KILLED on the erroneous order that I had my wife close manually (haha, I know).

So, these are the type of technical difficulties I have to deal with. Is it worth it? See the next post.

 

[Chabah]

Attached is a screen shot of my results for the year to date. I show everything instead of just February so that the statistics have more meaning. So far February is an 8R or 9R gain, or about $650. It changes a lot every day so I try not to worry about it too much. Note that the system flaw on PMTI caused a loss of over $1000 (including open P/L) instead of a gain, so that is a dominant trade for the month.

The attachment shows a Profit Factor of 2.39, which is great. The short positions have been on fire, which is great considering that the DOW has been making new highs. In other words, we are not correlated with the market - very nice.

We have 95 trades, so this is starting to get statistically significant. We have averaged about 1R per trade, which is of course unsustainably good. 43% winners reflects a string of good luck - it will probably settle down to around 40% in time. The $1886 is exactly the kind of trade we look for with a trend following system, and shouldn't be discounted too much as an outlier.

Honestly I think anyone would be happy with these results in the first 6-7 weeks of the year. In terms of the entire account we are up 16% with an average loss per losing trade of only $70 or .2%. In other words, we still have room to crank up the risk a bit. In time we will.

 

[Chabah]

Finally we have the equity curve. Now I really wish I could show an Open P/L version of this, and compare the two, but I don't have that capability. If I did, then the drop on trade 74 (approximately) would be all the way from $5k, not $3800 or so as shown on the graph. In other words, a lot more of a gut punch than is shown.

Nonetheless, given the closed P/L this is a satisfying curve. The losses don't run too many in a row, and the longest flat period is about 25 trades (maybe 30).

One lucky point is the big $1886 trade, which came right in the middle of the worst string of losses. What can you say - we got lucky!

And that's how it is - there is a lot of luck involved in trading, especially swing trading. Let's hope there is an edge in there as well.

Chabah

 

[Chabah]

Here were the goals from the first post of this thread:

Goals:
1: Get my trading account up and running by November 1st.
2: Start trading December 1st
3: Run fully automated by January 1st, 2007
4: Post a running update here on ET

The first 3 were all accomplished, and #4 is ongoing.

I would like to set some new goals, but aside from running the system and posting updates I am not sure what makes the most sense, especially from a forum standpoint.

Please post any suggestions you might have, and we'll come up with some new goals to accomplish over the next couple months.

Have a good weekend,

Chabah

 

[Chabah]

Yesterday was a huge trend day, and I had high hopes for the system. Unfortunately, it only made a modest gain of 1R ($75) in booked gains and another 2R or so in Open P/L. Much of the Open P/L was gone by this morning as the two short positions, CROX and AAPL, gained in the pre-market.

The system did have the right idea on a couple issues, but the trades were rejected by TradeStation. Basically most of my money was parked on the sidelines for the entire day.

Of course that is better than my long account, which dropped 5%. I don't actively trade that account though so I won't get too excited about it.

FYI, please post any requests you have for future goals for the forum.

One thing I plan to do is create an intraday version of the system and see if we can make that work using shorter bars. Still trend-based (not like my aborted rectangle breakout system), but a little different than the swing trading. That won't be until later this year though, after my other business is settling down in late summer.

Additionally I may return to the Mark Conway systems and see if my more experienced eyes can squeeze some value out of them. Much of my capital remains undeployed on a daily basis so adding another system might be handy.

Cheers,

Chabah

 

[Chabah]

So we hit the end of February. We started with $36750 and ended with $37k including $500 in Open P/L. So, by one count we lost $250 (booked), by the other we made $250 (account value).

Using the booked value, that comes to about -3.5R. As expected, the system error from early in the month, where a stock didn't close/sell short on the close and then gapped against me, was the dominant trade. If that one had worked, the monthly profit would have been more like 6R.

Other problems this month included double buys caused by a new system error that arose in response to my fixing of the earlier error. That error is where a position is opened at 3:58 pm and then doubled the next morning on the open. This doesn't necessarily cost us money - we could get lucky and double our profit instead - but in February it went against us every time. Hard. At least $450 or 6R there.

The other problem was the inability to get short several times. In January this generally benefitted us, as many of the shorts would have lost money. However, in February it bit us on the ass, as again we lost the coin flips - probably close to 10R was not earned because positions didn't open.

So from a purely theoretical system standpoint, it was a good month. If I were sitting at home managing the system - preventing all the errors - we would likely be at around 20R. Now $1500 isn't enough to justify sitting at home all day, but 1R per trading day is a very good rate.

Going forward I need to focus on making sure the system executes as intended, and then we can truly see if it is profitable. March is a long month, and has not started off well, so it will be interesting to see how it develops. With any luck we'll get to nearly 200 trades and a new level of confidence in the system.

Cheers and good luck for March,

Chabah

 

[mischief]

Hi Chabah,

First of all, thanks for this thread and congrats on your success in developing and implementing the system.

I'm currently running a long only automated script and want to implement the short side. I'm interested in understanding why your shorts aren't getting filled, so I can avoid the issue if possible.

Do you think it is the uptick rule, coding problem, dodgy broker?

Thanks,

Mark

 

[Chabah]

Quote from mischief:

I'm interested in understanding why your shorts aren't getting filled, so I can avoid the issue if possible.

TradeStation has a bit of a reputation for not filling short orders on a lot of stocks that they *should* be able to. Their message says to call, and they'll find the shares for you when you call. However, I am not there so I can't call.

Sometimes they just can't get the shares, or can't get what you need. One time I needed 180 and they only had 100.

Basically, it is the broker. I wouldn't say they are dodgy, though. Just limited. Maybe not big enough yet...or maybe there just isn't that much demand or something.

From my perspective, it is just part of the system - some orders don't get through. I just need to deal with it for the time being.

Chabah

 

[Chabah]

So March and the 1st quarter ended on Friday. For March, we were +4R, or $300, to about $36,800. With only 8 stocks going, most of the account is on the sidelines most of the time, which is good and bad. It's good I am not getting killed, but it's bad because my rate of return suffers.

I still don't have much time to hunt down new stocks, so holding about even is OK for right now. Soon I should be able to start up again, once my other side business (and two daughters!) requires less attention.

The first quarter numbers look fine, as we had the great January. Here are some quarterly stats:

Profit Factor: 1.66
Long: 1.28
Short: 3.06

Clearly missing those shorts hurts more than it helps.

Net profit about $4600.
RINA index 96.53
K-Ratio .18

176 trades, 38% winners, avg. winner of $173, avg loser of $64, avg profit $26. Remember most of the profits were made when the R or risk amount was $50, so the average win per trade is .5R. Very nice but half of our trades were in January. Literally.

Please post any questions or comments. Thanks,

Chabah