I am extremely torn, as I am very excited about this new system, and would love nothing more than to write all about it. I used to not worry so much about divulging strategies. In order for any edge to be eroded, many other traders would have to take the exact same signal and use the exact same exit strategies. I know from my limited experience with mechanical system trading that the system has to be closely matched to one’s personality, psychology, and risk profile, in order for him to trade the system successfully over a long time frame. So what are the chances that there are enough traders out there that will follow the system closely enough to erode the edge? I honestly do not know the answer.
And there is also the issue of scalability…This strategy may not be able to be traded by the hedge fund manager with 100 million in capital. So maybe I’m just paranoid, or over-cautious, in my secrecy.
What I do know is that I’m in this to become filthy rich, and not so much to brag, so why take any chances? And then there is the fact that 2 other traders have spent a significant amount of time helping to develop this system. Anyway, for these reasons, I will not divulge the strategy.
That being said, I do want to promote the idea of trading mechanical systems, to anyone who is interested. I have been making more money, and I am less stressed since I started trading almost 100% mechanically. When I wake up in the morning and see the futures up big or down large, I do not really care any more. If the trades stop out, who cares? I did not pick them, the system did!
There are other benefits, too. No more do I have to get frustrated after spending hours researching a trade, only to have the trade turn against me almost immediately. No more do I waste time considering when to sell a profitable position, and then kick myself for selling too early or too late.
What I find satisfaction in now is seeing a profitable system develop from infancy to implementation. It is especially nice when real profits start to roll in after looking at years of historic/hypothetical profits.
This system is being traded out of my Scottrade account, so I do not have the fancy graphics and reports that Tradestation generates. Here are the recent trades, without their symbols:
| Symbol | Date | Entry | Shares | Cost | Exit | Date | Proceeds | Â $ P/L | % P/L | Total P/L |
| $939.95 | ||||||||||
| XXX | 13-Aug | 52.08 | 75 | 3913 | 54.77 | 15-Aug | 4100.75 | 187.75 | 4.80% | |
| XXX | 13-Aug | 56.84 | 75 | 4270 | 56.61 | 15-Aug | 4238.75 | -31.25 | -0.73% | Â |
| XXX | 14-Aug | 17.63 | 200 | 3533 | 18.115 | 15-Aug | 3616 | 83 | 2.35% | |
| XXX | 14-Aug | 28.13 | 150 | 4226.5 | 29.41 | 20-Aug | 4404.5 | 178 | 4.21% | |
| XXX | 15-Aug | 37.98 | 125 | 4754.5 | 37.74 | 26-Aug | 4710.5 | -44 | -0.93% | |
| XXX | 19-Aug | 40.23 | 125 | 5035.75 | 38.83 | 21-Aug | 4846.75 | -189 | -3.75% | |
| XXX | 20-Aug | 35 | 145 | 5082 | 35.65 | 26-Aug | 5162.25 | 80.25 | 1.58% | |
| XXX | 20-Aug | 44.94 | 110 | 4950.4 | 44.52 | 29-Aug | 4890.2 | -60.2 | -1.22% | |
| XXX | 20-Aug | 43.92 | 115 | 5057.8 | 44.56 | 26-Aug | 5117.4 | 59.6 | 1.18% | |
| XXX | 20-Aug | 31.89 | 150 | 4790.5 | 32.47 | 29-Aug | 4863.5 | 73 | 1.52% | |
| XXX | 22-Aug | 31.33 | 160 | 5019.8 | 32.59 | 26-Aug | 5207.4 | 187.6 | 3.74% | |
| XXX | 22-Aug | 32.96 | 150 | 4951 | 32.95 | 29-Aug | 4935.5 | -15.5 | -0.31% | |
| XXX | 26-Aug | 15.28 | 395 | 6042.6 | 15.94 | 28-Aug | 6289.3 | 246.7 | 4.08% | |
| XXX | 27-Aug | 33.24 | 180 | 5990.2 | 34.34 | 29-Aug | 6174.2 | 184 | 3.07% | Â |
Win percentage: 64.29%
Average Trade: $67.14
Average Trade %: 1.4%
Average Winner: $142.21
Average Loser: -$67.99
Commissions of $7 on each leg are included in the results.
Some thoughts about the results up to this point. First of all, two trades were botched. The loss I took on August 21st was from an erroneous entry of a sell-at-market order rather than a stop-loss order. The second error was made when I accidentally mis-typed the stop order amount, by one dollar. Of course, the stop was hit, and then I realized my typo. I did not really track the trades to see if they would have been gains or losses, had all the correct parameters been applied.
I started trading the system with smaller positions, with a stop loss on each trade of 6% from entry. This position sizing limited risk on each trade to about .35% of total equity. On the last two trades, I began ramping up risk to about .5% of total equity per trade. I will eventually settle in somewhere between 1-2% risk on each trade. Obviously, trading with 1% risk would have tripled the profits and losses above. That should provide a good idea about what kind of returns I’m expecting. CAGR in the backtested results is 47.35%, using 1% risk per trade. My current profits represent about 2.3% gain on the total account size, even though I was only fully invested for about 2 days out of the total period.
At $14 dollars per round trip, my commissions are about 21% of my profits. As I increase my position sizing and percent risk, percentage of profits relative to commissions will improve. Still, the ratio is too large, and the system will benefit tremendously by switching to a cheaper broker.
Woodshedder — this is awesome. I always wondered if program trading really worked or not. It sounds like it can, and really reduce a person’s stress. It sounds like it is not truly auto-traded, but your system spits our trade recommendations that you enter manually? Also did you program your system using a computer language, or is it something you assembled by customizing a trading rule or interface available from your broker? BTW I can get you a great broker for $3 a trade – email sloof dot lipra (at sign) gmail dot com. Awesome!
Congrats Shed, nothing like hard work paying off right.
However, you think that you’re gonna win over any fans with this major cock-tease of an article? C’mon.
It’s clear you just want to gloat, which incidentally I am down for. I have a trading style I like, so I think it’s awesome that you have one too. The fact that it’s so great but no one’s invited is south parkian in its humor.
Still though, after reading that I feel like I just left Mons Venus all blueballed or something.
Please forgive all the references to my dick.
Yeah, what he said.
Do you need us to butter you up with some compliments first? Well, those big computer monitors sure bring out your deep, piercing, soul-searching eyes.
Great work, Wood. As usual. On Labor Day weekend, no less.
HopetogetRich: the system does spit out stocks that I enter manually. Usually there are few enough selections on a daily basis that the system could be automated, but every once in a while, there will be more selections than I have cash to purchase. Human intervention is needed in those instances. When that happens, I generally select those with the best liquidity.
The stocks are selected based on a filter which is coded fairly easily. Most screening software should be able to handle the code. I run the filter every evening, and then place my orders.
I’m gonna send you an email about the brokerage you speak of.
Danny, you crack me up. I’m kind of stymied as I want to write about what I’m trading, but literally all I’ve been trading is this system and another one, so it makes it difficult.
Still, I like to discuss systems and system development, so maybe there is some value in the post.
The lowly retail investor with 100K can design systems which will make him or her a millionaire, within 3-5 years. If the system is scalable, those gains can be compounded into 10s of millions.
There seem to be many edges that can be exploited for those without enough money to move markets. Once you get rich, it gets harder to employ the strategies as you will move the market when you enter.
I just find it fascinating, and so, even though I won’t reveal my strategy, I still like to think about it.
DPeezy- you are killing me…if only you were Boca!
But then I’d have to fight Anton!
Well, if it counts I gave Dpeezy positive karma for mentioning your deep soul-searching eyes. Don’t worry about Anton, he’s busy somewhere flipping coins.
Now will you stop teasing us and tell us how you picked the trades that gave you such a tidy profit last month?
don’t make me grovel, dammit! 🙂
I have an idea, you could just tell the people on this thread (who’ve expressed interest) via email or Google Talk and then you won’t have it out there on the Internets forever spoiling your edge.
Adding, check your email for a groveling message.
C’mon Woodie, how about a few morsels for your loyal supporters? I always said you should have top billing at iBC.
😎
Folks, I would like nothing more for us all to get rich together. Also, I would like to provide you the system so that you could evaluate to see if it is right for your personality, psychology, risk profile, time horizon, starting equity, etc. However, I can’t.
One of the other developers just emailed me new figures which include commissions and slippage.
Slippage of .1% on each leg of the trade, coupled with standard commissions, has lowered the CAGR by about 10%.
Lets imagine that there are 10 of us trading this system, and on any given day, there is only 1 stock pick. We all enter market orders, some for 100 shares, some for 500, some more, some less. What is the open going to look like? It would be easy for just a small group of us to cause a stock to gap up. Now I can ONLY trade the open, as I’m at work by 9:30. I do not have a choice. Sure, I could use limit orders, but I hate them, for some reason (a stupid bias, I know). I do not want to build more slippage into this system, which is what would happen should a group of us be buying the same stocks.
Anyway, I hope you all can understand. To make up for this, I am going to post some future systems (which I, nor my friends are trading) for you all to consider and possibly try out. I also think that one of the co-developers is going to have some time this fall to begin some comprehensive blog posts about system development.
You’re right Shed, I shouldn’t have suggested it, and I apologize. Cheers.
Wood,
what’s your e-mail? thanks.
After reading this thread, I get the sudden urge to throw a turkey sandwich at someone.
Fly, good sir, might I suggest that instead you eat that turkey sandwich?
Enjoy.
Chivas, go to my old blog, trade while working, and click on the “about me” link.
Quite a contrast to the Fly’s traditional magnanimity, isn’t it?
Score one for the Yankees…
__
Feel free to throw on of those 2.5in steaks my way instead.
I’m willing to share a strategy, I don’t give a fuck.
Here’s my request Shed, if you have the bottle for it:
Backtest this strategy going back only YTD, the last 8 months.
–Short the market close on the third up day in a row
–Double down on the fourth up day in a row, if it occurs.
25,000 per trade, with 50,000 being the max.
No stop.
cover 5 days later.
Please post the results in the comments section.
I’m pretty sure it’s 13/14.
Danny,
Your strategy isn’t much different than the RSI(2) strategy, you’re just using a different way of determining when the market is overbought, i.e. using price action vs. using an indicator. Two sides of the same coin, basically.
Danny, if I can figure out how to code it as you have described, I will most certainly post the results for you.
The PPT will be able to execute backtesting strategies such as the one Danny just mentioned.
God willing, it will have a much more user-friendly interface.
Shedder,
I have some “buy three days down and sell two days up” code I can modify for StockFetcher if you can’t get it to work in TradeStation. Let me know.
dogwood – but, my strategy is manly and ungay, unlike the homo RSI2 trade.
Just kidding.
I didn’t quite think of it those terms, price action v. indicator but that’s a good way to describe it.
I still like it more conceptually than RSI because markets can stay overbought, but they don’t advance much after three days up/four days up.
That’s more about RSI in general than RSI2, which is a good trade.
——
That would be tits Jeremy.
lmao at shameless plug’s comment “a much more user-friendly interface.”
Dogwood, I have some similar code for stockfetcher, although I might still need what you have. I’m going to try and do it in TS, as then I’ll be able to have it double down on the fourth up day, and perform the money management that Boone has asked.
I’m going to have to search the TS archives and forums to find some starter code. As the beer drinking and meat-grilling has begun here in Woodshedder’s neighborhood, this will have to wait for tomorrow.
shed, you know what’s funny?
based on the dates provided I could deduce all the symbols you traded. Of course that doesn’t reveal the strategy, but I bet one could get an idea.
Essentially, this is a blackmail, you know my paypal, my vig is 20, so that’s .2(930) = 186 bucks please.
Dogwood, I have some similar code for stockfetcher, although I might still need what you have.
What cheek!
As the beer drinking and meat-grilling has begun here in Woodshedder’s neighborhood, this will have to wait for tomorrow.
Hmmm… Midlothian MILF ministrating?
We’ve been working on the bocce court all day… probably have it ready for the inaugural tourney tomorrow.
Danny, I altered the dates, just in case. The time between the entry and exits is correct in number of days, but the dates aren’t.
So solly.
Jake, Bocce for us will be tomorrow. We are letting the grass get one more day on it and the ground to firm up a bit before I mow her down low.
Shed, you cunning bastard.
Interesting results on my backtests. around 80% accurate, pf between 1.4-1.8 on various jiggerings of cover after 3 days, 5 days, etc.
Shed, do you have a box/court? Or do you just play on flat lawn, a la lawn bowling?
We built the 60′ x 10′ box, gravel, sand, brick dust, drainage, a roller… the whole thing.
It should be sweet.
__
Danny, do you want me to try and run the tests? Why don’t you do a blog post?
I think that short strategies are the hardest to develop.
Jake, a lawn that wraps around the house, with some interesting topography.
I think that short strategies are the hardest to develop.
I agree. Other than using RSI(2) for shorting overbought markets, I have only been able to create one other consistently profitable swing trading system for shorting stocks.
Experiments with trend following systems that go long and short always produce losses on the short side.
you are using Tradestation these days, right? how come you are paying so much for commission then?
at the sizes described you should be averaging around $2 commission per trade.
Born2- I have two accounts. The TS account is trading another strategy.
Who gives a man negative karma for this comment?
WTF? Fuck you negative karma fuckers!
“Woodshedder Says:
Danny, do you want me to try and run the tests? Why don’t you do a blog post?
I think that short strategies are the hardest to develop.
Jake, a lawn that wraps around the house, with some interesting topography.”
Wood, don’t sweat it, the sight is filled with haters.
Go check out my battle with the Obaminoids over at KoPG. They’ve gone quiet, deciding “negative karma” is a salutory defense of “the Mastah,” in lieu of a distinct lack of argument, which runs to “why are you criticizing hiiiiiiimmmm?”
This site needs a “no whining” sticker.
__
You spelled “site” wrong Jake. No whining ’bout it neither.
What are you mouthing off about now, Pup? I never whine. And the spelling is correct.
__
Aside– switched over from the Florida massacring today to NBC or some other net and they had half pipe skateboarding. Some “old dude” — Bart, maybe? — was ripping it up with 540’s and shit. Some of that stuff is pretty wild to watch, I must admit… like the fuckers are defying gravity or something.
It’s nice to see you have as much confidence in your spelling as your politics, but you did spell ‘site’ egregiously wrong.
It’s web SITE, not web sight. So sorry.
True dat!
Shed, forget the dates and the equities.
Your details provide way too much information for someone who knows how to code an inference engine:
http://en.wikipedia.org/wiki/Inference_engine#List_of_Engines
For the unaware, an inference engine takes data and constructs a set of rules based on the data (using multivariate analysis – ambidextrous cousin to the linear regression).
Easier than that though, I suspect that anyone around with Excel and a working knowledge of the correlation function should be able to put your entry price and exit price and generate out the formula or a reasonable copy of it.
And if you screwed around with your reported entry/exit prices as long as the difference between entry and exit remains the same – it’s still inferable.
As my childhood hero would say, “Elementary”
Actually – the easier way is to use the forecast function:
http://jldexcel.blogspot.com/2006/04/excel-forecast-function-predicting.html
Happy Labour Day
Laughing Crow- usually I would defer to your vast and obviously well-developed experience in this arena. This time, I think there is some truth to what you say, but I do not think it would be as easy as you have outlined. Maybe I’m wrong.
I went to the blogspot link and I do not at all see how that forcast function could help.
Entry Exit Forecast Error Error %
52.08 54.77 53.89 0.88 1.61%
17.63 18.12 18.4 0.28 1.55%
28.13 29.41 29.02 0.39 1.32%
35 35.65 35.82 0.17 0.46%
43.92 44.56 44.79 0.23 0.52%
31.89 32.47 32.86 0.39 1.21%
31.33 32.59 32.48 0.11 0.35%
15.28 15.94 15.94 0 0.00%
33.24 34.34 34.34 0 0.00%
This was derived by taking all of the listed winning trades and using the forecast function as such: =FORECAST(A2,B2:B10,A2:A10) in column C.
Here’s the how to:
Take all the listed data
Remove all losing trades
Put the entry price in column A
Put the exit price in column B
Enter the formula for column C and copy and paste the rest
The Error formula is calculated as =ABS(C2-B2)
The Error percentage is D2/B2
The standard deviation of the errors is .27
Thanks Cuervo. What I’m clueless about is how this gets anyone near the rules or code for the system?
Well, I can project what your system will do.
For instance, build the spreadsheet as I noted before.
Now add at G1 “Entry” and H1 “Exit”
In G1:G10 put 10,20,30,40,50,60,70,80,90
In G2 put =FORECAST(G2,B$2:B$10,A$2:A$10) and copy and paste though to G10.
You’ll end up with:
Entry Exit
10 10.19
20 20.58
30 30.96
40 41.34
50 51.73
60 62.11
70 72.49
80 82.88
90 93.26
Which tells me what your projected earnings are and what your exit is going to be. While I don’t know what your stop points, your selection criteria, and your specific rationale are, I can reasonably (+- .27) predict your exit points.
Searching through a database of the last month’s data (assuming you used August, you may have used January) would narrow down the list of candidates if needed. And the point at which someone could identify which equities/etfs/bonds/commodities you traded then it will be game over because reverse engineering your selection criteria would be child’s play.
Cuervo, I’m sure you are right, but still something doesn’t seem right.
For example, a good number of my exits are different than what one would deduce from the exit strategy, due to a number of factors, with slippage being a major one.
Also, because I trade on EOD data and market orders at the open, the exit signal may hit at any point in the day, but not be acted on until the next morning.
Also, as the exit is not based on a profit target, the exit will be different for every equity traded.
Am I making any sense?
Sure and I’m saying that there is a way to infer from your data a strategy by focusing on the winning trades.
The losing trades would give one an idea of the risk of loss.
Combining them both, and using a MonteCarlo analysis, one could project the range of your potential wins/losses.
LOL, if anyone can figure that out, they deserve it.
I’m shorting VZ T and DE. DE at 72.5 based on what I don’t know. and VZ and T, ( at recent highs) just because I feel lucky.
Just kidding. There are some T/A and Book value reasons.
BTW, MFE went against me then for me, Out with a small profit, and If up Tuesday, will short again. ( would still prefer $41.6 entry, but MFE is not co-operating)
oh yeah, UNG usually bottoms in SEPT, and then takes off for winter time. I might start looking at HES and DVN, soon.
FRO is gapped down. I think they have a SEPT 4 dividend date, I thought stocks went up going into dividend date, payable Sept 12th. I’ll have to double check those dates as I am trying to pull them out of memory.