 |
Trader's Roundtable
A forum for mechanical system traders.
|
 |
| View previous topic :: View next topic |
| Author |
Message |
Tigerline
Full Member
Joined: 23 Jun 2003
Posts: 10
Location: Cumbria, UK
|
 Posted: Sat Aug 16, 2003 9:28 am Post subject: Going Live |
 |
|
Hi,
I became interested in trading in 1997 after buying a course. Then the best and worst thing happened I made money. Worst because I thought I knew how to trade, best because it kept me interested in trading long enough to realise that I had no trading system at all.
I can still remember those days of fluctuating emotions as cocoa, coffee, wheat or whatever I happened to be trading, would rise and fall intraday. I have traded on and off since More off thank goodness as I realised my system needed changing.
This process has involved me looking at myself more than I would have liked initially. But I needed to and it has made a difference. My whole life has changed internally as well as outwardly. I am a doctor of medicine. I also established a property company 2 years ago to raise funds for trading. I am about to go live with £200,000 with my newly established trading company.
My system is my own. A simple Moving Average crossover system. With the addition of money management and position sizing, I have paper traded from Jan 03. Using 1% of equity and a volatility model, the system has an expectancy of 52. There have been 391 trades, 60% winners and 40% losers, average gain 0.7R , average loss 0.7R. Gain to date is 66% since Jan03.
As I have manually traded this system for 8m, I have a good feel for it and it fits my personality.
I have been following the discussions here for a few months. Fascinating stuff. Any comments or advice before I go live would be gratefully received.
thanks
Tigerline
|
|
| Back to top |
|
 |
Sir G
Moderator
Joined: 15 Apr 2003
Posts: 243
Location: Salt Lake City, Utah
|
| Posted: Sat Aug 16, 2003 12:34 pm Post subject: |
|
|
Hi Tigerline-
Sounds like you are on a much healthier path now.
What is settled in your mind if you start to trade and you find out that you are not 60% winners and 40% losers but 40% winners and 60% losers?
Gordon
_________________
There is no such thing as bad weather, just a poor choice of clothing. |
|
| Back to top |
|
|
Chuck B
Roundtable Knight
Joined: 17 Apr 2003
Posts: 130
|
| Posted: Sat Aug 16, 2003 1:28 pm Post subject: |
|
|
Tigerline,
If I was in your position, I would assume that my paper trading experience so far this year represents at best perhaps 2-3% of what I will see in the future. Expect your future returns to be 1/2 of historical test results, your future max dd to be at least 50% higher than past (i.e. if max was 30%dd, then assume 45% is on the way at some point), and the length of the longest dd to be perhaps twice that found in historical testing. If you can stand these bounds, emotionally, and continue to trade the system, then you're at least on your way to making it happen in real time. When the "larger than any past DD" comes, it will feel a lot different than your experience year-to-date  .
Good luck!
Cheers,
Chuck
|
|
| Back to top |
|
|
Kiwi
Roundtable Knight
Joined: 15 Apr 2003
Posts: 505
Location: Gone Too, Ciao.
|
| Posted: Sat Aug 16, 2003 4:08 pm Post subject: |
|
|
However if you hold the discipline during the first big drawdown you'll feel better about yourself. And the second drawdown wont feel nearly as bad.
If you like moving average systems you might also look at developing a second system which buys when you get pull backs to the moving average and then some form of forward movement.
Good Luck,
John
(read Mark Douglas, as tedious a writer as he is hes quite good on psych)
|
|
| Back to top |
|
|
DrHendricks
Roundtable Fellow
Joined: 12 May 2003
Posts: 83
|
| Posted: Sat Aug 16, 2003 6:22 pm Post subject: |
|
|
Hail fellow well met. As a fellow MD, just let me say- may The Force be with you. If you are not pursuing a pyramiding strategy, and there is no serial correlation in your equity curve you may be able to get some idea of the range of drawdown with a Monte Carlo simulation using your year to date trades. But I'm not sure I should be giving you advice. I am approx. 8 months behind your schedule and on the other side of the pond. 
David
_________________
It beats billing Medicare |
|
| Back to top |
|
|
Tigerline
Full Member
Joined: 23 Jun 2003
Posts: 10
Location: Cumbria, UK
|
| Posted: Sun Aug 17, 2003 8:04 am Post subject: |
|
|
Hi
Thanks for the comments and advice
Gordon.
I appreciate that the figures will be different when I trade real money. What I intend a bigger account to do is to reduce the emotional factors as you are trading a smaller % and in theory more detached from any single trade. If % winners change then so be it. Time will tell if it still profitable!
Chuck
I agree. 8m is so short a time to test. It can't really prepare me for the extremes of situations that will happen in the future. Max dd to date is 8%. I think this is small so quite prepared for it to be double this. Max length of dd to date is 6w, so ditto for this. Chuck, you sound like you have a lot of experience trading. What does it feel like when you get the max dd. I know it will come, of course, it is just a matter of time. How do you cope?
John
Yes. My system also uses a signal I call a "bounce," which is when one moving average bounces off another. I think this is similar to what you are saying but using 2 MA instead of actual price and a MA.
David
Hello fellow medic. I think we like to expand our horizons and push ourselves to new highs in self development with trading don't we
Thanks for the Monte Carlo simulation idea. According to some stats posted here a while ago, a 60:40 W:L system where there are 40 trades a month, should give you 7 straight loses in a row every 30 months. I have had 7 loses in a row already so i feel privileged.
Thanks again to you all for your help and advice. I'll provide an update good or bad when I get going.
Tigerline
|
|
| Back to top |
|
|
Chuck B
Roundtable Knight
Joined: 17 Apr 2003
Posts: 130
|
| Posted: Sun Aug 17, 2003 11:53 am Post subject: |
|
|
| Tigerline wrote: |
Chuck
I agree. 8m is so short a time to test. It can't really prepare me for the extremes of situations that will happen in the future. Max dd to date is 8%. I think this is small so quite prepared for it to be double this. Max length of dd to date is 6w, so ditto for this. Chuck, you sound like you have a lot of experience trading. What does it feel like when you get the max dd. I know it will come, of course, it is just a matter of time. How do you cope?
|
I was really referring to your historical test results -- apply those factors of safety to them. One thing you can do with the historical results is to go back and, using your best creative abilities to fantasize, "live" the equity curve and "live" a lot of individual trades. There are many time periods that may stand out as more useful to "live through". For example, look at strings in the equity curve such as number of days in a row of winning days, losing days and almost flat periods. When you experience one of these strings in real trading, it's easy for the mind to attach some significance to them. For example, the first time you ride a large winning streak to a significant new equity high, it's natural to think and feel vindication about all the hard work that went into the system, the discipline you used to reach this point, etc. Of course, those natural feelings are meaningless to the system/markets, and naturally, it's possible that a long dd is just around the corner. Anyway, those "streak" periods in the historical data are great areas to go deeper into and develop mental fantasies and rehearsals related to them by "mentally associating" into how you would feel in real trading with those outcomes. Be creative and watch for how sly the mind can be at wanting to gloss over stuff that creeps up  . Those will be the areas you most need to work on...the one's you don't want to fool with.
If you are trading a system that has a large number of days for length of the average winning trade, then you will have ample opportunity in future real trading to become enamoured with those trades. Any attachment to it can easily result in inadvertent exiting (i.e. 'it can't get any better than this'), the "call of the countertrend". On the other hand, if you can stay detached, when you are in one of those whoppers you'll stumble upon "clues" that help you know your risk/reward is now poor, and it may be time to "violate" system exit rules. I'm more speaking of those times when you get a hugely, multi-market correlated winning streak (i.e. Oct 1998 was a BIG one for my portfolio). These clues for me are almost always from the environment...when it becomes "news", it's almost always time to take risk off the table. But I digress here...
All the mental rehearsal in the world is still no substitute for actually doing it of course. My first big DD trading my long term system came in the summer of 1998. I had started the system at the beginning of 1995, and had positive returns for 95, 96, and 97. However, by July of '98 I was in the mid 20% range in DD. I think it bottomed near the end of July around -27%. At that point it had been over a year since the last equity high. I can recall thinking all sorts of thoughts like, "damn, I've done everything exactly as planned, maintained and traded this system by working every single night to generate its orders and place them, and WHAT do I have to show for it?" It felt like the biggest waste of time in the world at that point.
Fortunately, in retrospect, I was traveling for most of late July and early August with my family, running my system at night each night...keeping it going. I stuck with it as nothing was out of line with my expected outcome (other than my realizing that "one of those" periods was here). Placing those orders each night got to feel like punishment though ...I felt like Eeyore, the Winnie the Pooh donkey..."here we go again...call 'em in...run the equity curve...yep...just what I thought...another down day...oh well...what do you expect...(head down, tail between legs)"  I think the main thing helping me though is that this was just one part of my investment activity as I had nice returns in utility stocks going amongst other stuff at the time.
Well, to make a long story short here (sorry), the DD bottomed on August 20th 1998, and on Sept 30 1998 I was at a new equity high (just looked it up). For some reason Eeyore was replaced with Tigger .
This was my first major DD experience with this mechanical system in real trading, and I learned an enormous amount from it. Those internal lessons are deep within me now and very helpful on a continual basis. The "Tigger" feeling at the new equity highs is just as useful to learn from also, but that would result in another long story .
Hey, let's ask Gordon to reply to this subject with his experience as he's got tons of it (and a MUCH better equity curve outcome than me!). Gordon, what was your experience going into your first big DD? I'm guessing your trough-to-new-equity-high times are so compressed compared to my long term system that it wasn't as big of a deal?
|
|
| Back to top |
|
|
Sir G
Moderator
Joined: 15 Apr 2003
Posts: 243
Location: Salt Lake City, Utah
|
| Posted: Sun Aug 17, 2003 3:28 pm Post subject: |
|
|
Hey Chuck-
| Quote: | | Hey, let's ask Gordon to reply to this subject with his experience as he's got tons of it (and a MUCH better equity curve outcome than me!). Gordon, what was your experience going into your first big DD? I'm guessing your trough-to-equity times are so compressed compared to my long term system that it wasn't as big of a deal? |
I think everything is a function of being prepared by accepting things instead of trying to domineer things. You wrote about living certain moments and playing them in your mind until they have meaning. My interpretation of that is a very good way for you to accept those micro moments.
Acceptance is the key. I do my best to accept the macro system. Systems are the funniest of creatures, you have complete and utter control over them when you are designing them and when they are done you need to then relinquish all your control on it, in a sense, you now need to take your orders from it.
In my life in general, I have always tried to do the right thing, regardless of the repercussions of my actions. I just try to keep true to myself by doing the right thing. My job and focus is always on my actions. So the profits/loss are results and I for whatever reason don’t concern myself with them. But I try to concern myself with being sure the data is clean, the systems & programs are running properly, the brokers are offering me the best service at fair pricing, the trades are all reconciled with the clearing firm, I have backup power, computers, Hard drives and phones in case something goes down. Trades are trades and if the systems are based upon robust logic then my job is to focus on the environment to let those systems function. You gotta believe in them & respect them. Your control needs to be moved to creating and keeping an environment that will let the system function. Chuck, you placing the orders on your family holiday is a good example of keeping your environments in balance.
Think of each system as a manager that you just hired. Ideally, the manager is competent and doesn’t need you looking over his shoulder. If the manager isn’t competent then he shouldn’t be hired in the first place. If you believe enough in the manager to hire him, then let him do what he is hired to do!
I believe there was never a minute when I didn’t know the P&L.. but being concerned about it isn’t my job. My job was to keep the path clear so the P&L could live on tomorrow and the next day. Also, I seem to be a detail kind of guy, so when I did my research, it was really right on with what the realtime track record was, so there wasn’t much to be concerned with. There are also so many bigger issues in life to concern yourself with, family, loved ones, health & happiness. The better you do in those areas then better the trading should be. Another way to phrase that is, what is the reason to all this … “why do you trade?” If it is to help your family and loved ones, then there isn’t a conflict and the reason can support your goals. But if you don’t know why you are trading, it might be a reason that doesn't support what you beleive are your goals.Then it might be easy to muck up the system and results with your actions.
Prior to me getting my trading act together, I bought a bunch of books and read them and tried to trade that way, needless to say it was ugly and the emotions would run with the equity. Because at that moment, my idea of trading was just placing the orders and seeing the results of those orders. I didn’t “own” the logic of the system, nor did I really even believe in the logic. So I’m sure I was conflicted from the get go. And then was there was nothing else for me to focus on or believe in but the results of my actions. The moment that I realized that I need to develop the system to my strengths was the moment things changed. Everyone getting involved with trading now-a-days is very lucky as there is software to support the testing of their ideas. Back then when I started SystemWriter from Omega, which was an end of day platform wasn’t even out yet! The initial system logic that was later the backbone of my intraday systems, was actually a monthly data system that I had designed in SystemWriter. I built the monthly system because back then I had a $5,000 386 computer and I couldn’t stand waiting hours for the daily systems to test out, so I moved it to monthly data! In the early 90’s when TradeStation came out with the ability to test intraday data, I retooled it and that was the beginning of the journey.
After I began developing systems that played to my belief systems and played to my strengths, I don’t think trading ever caused me stress. Don’t get me wrong, I have stress like everyone else, but trading has always been a dear trusting friend to me, and me to it.
I think also, I would have some issues if my system was a long term system, as all the ebbs and flows that you see in the equity curve is simply a 1:1 move in the data. They all represent opportunities that can be exploited if the system is geared towards those ebbs and flows. Also, a question that everyone of us should ask is why do you trust the system, or why SHOULD you trust the system? I trust them because of the logic that is the base of it and because I can see thousands of occurrences of the logic in action.
I don’t think I answered your question well..... but I think it comes down to we all have true goals and perceived goals, fortunately both of mine were in the same direction and the weapon of choice that I employed was to try to do things the "right" way, define my actions before hand, then to focus on my actions and to let the results do what they will.
Gordon
_________________
There is no such thing as bad weather, just a poor choice of clothing. |
|
| Back to top |
|
|
damian
Roundtable Knight
Joined: 15 Apr 2003
Posts: 818
Location: dusseldorf
|
| Posted: Mon Aug 18, 2003 5:34 am Post subject: |
|
|
| Chuck B wrote: | | ..."here we go again...call 'em in...run the equity curve...yep...just what I thought...another down day...oh well...what do you expect...(head down, tail between legs)" |
I am glad someone bigger than I wrote this. It was good to read it.
Recently I found myself sarcastically but calmly talking to my equity curve, vis:
"of course you have gone down again today, that is all you are good at. Well, good on you. Try doing something a little different for a change, even my dog knows more than one trick".
My terse words did not produce any improvement over the ensuing days, so now my equity curve and I are no longer talking.
To me there is nothing actually that bad about a drawdown, besides the fact that:
a) you are not making any money
b) as is my case, you are heading directly back to where you were 2 years ago
c) you lose precious diversification power.
Neither of these facts will take me out of the game, and the name of the game is to stay in the game.
Point (c) is probably the one that hits me the hardest. In any given month I do not really care if I am in draw-up or draw-down, nor how much money I have, so long as I have enough money to implement my system properly and so long as I have enough money to benefit from the full offerings of diversification. Obviously I would prefer to not be in a drawdown, but that is not reality, so I do not bother with such dreaming.
So TigerLine, when your big DD comes, don't bother talking to the curve 'cause it won't listen. There really is not anything to be done about it, in fact that is the key: do nothing. Keep following your rules and marvel at how slowly time seems to pass when the curve is heading south. In a way I actually enjoy the DD periods as DD's are very valuable experiences in trading.
Did I sound convincing?
cheers
damian
|
|
| Back to top |
|
|
bagherra
Senior Member
Joined: 25 Apr 2003
Posts: 33
Location: Charlotte, NC
|
| Posted: Wed Aug 20, 2003 10:14 am Post subject: Drawdown and Return Expectations |
|
|
Chuck:
Could you reiterate why you should expect the drawdown to be 50% greater than historical test results, and why you should expect returns to be half historical test results?
This assertion has been made several times by various posters, but I don't recall it being fully explained.
I can think of no logical reason why historical results should materially differ from actual results.
|
|
| Back to top |
|
|
Chuck B
Roundtable Knight
Joined: 17 Apr 2003
Posts: 130
|
| Posted: Wed Aug 20, 2003 11:51 am Post subject: Re: Drawdown and Return Expectations |
|
|
| bagherra wrote: | Chuck:
Could you reiterate why you should expect the drawdown to be 50% greater than historical test results, and why you should expect returns to be half historical test results?
This assertion has been made several times by various posters, but I don't recall it being fully explained.
I can think of no logical reason why historical results should materially differ from actual results. |
Perhaps it's best then to consider the whole concept of historical testing to be illogical  . I'm not saying this in jest really either. The corollary would be to assume that your historical testing period has captured the worst possible drawdown that can occur, while also capturing the best possible ROI that can occur, in the future. In other words, you need a "factor of safety" on the output of your testing. In the engineering world, there are a whole set of "factors of safety" that are applied to stress calculations, for example. These take into account all sorts of things that the design engineer couldn't dream of at the time something is designed and also account for slightly imperfect workmanship and/or materials. These safety factors are applied on top of "exact" engineering analysis -- by that I mean you are dealing with something that meets laws of nature, you can run the calculations and then test it in real life, and it responds just like your analysis, every time.
Unfortunately for us, a system/markets combination does not possess this quality. Years ago, I believed that it was possible to use engineering and math to "figure out" the markets. If it did, we would only have to run a few tests and then crank open the money machine for life as we would "know" the future would respond just like the past. Fortunately, I realized that as much as some would like to sell you their latest "cycle analysis", there is hardly anything in the physical world that even remotely approaches the true difficulty of capital market system design. Perhaps trying to obtain an exact solution to the path of a fluid molecule in a turbulent flow field would be similar . In other words, it ain't happening.
I view the parameters I mentioned as an extremely conservative view. If you assume those parameters will occur in the future (and you have a substantially tested and robust method applied to liquid markets), and the whole system/markets combination still makes sense, then "perhaps" it will be worthwhile to trade into the future in real life.
Interestingly, this goes both ways of course. Measured in terms of R-multiples®, the first trade my system took in real life in the yen was a bigger winner than any in the entire history of the yen trading. This was on the long side in yen futures in early 1995. Even more interesting was that the second trade it took in the yen was an even bigger winner  (and this was on the short side this time -- it held the short from something like 4 months or more as I recall). Anyway, out of the entire history of yen analysis, trades the size of these two did not exist until I had started the system. They stand today as the biggest R trades for the yen with my system as their size hasn't been exceeded since that time.
Last edited by Chuck B on Wed Aug 20, 2003 6:11 pm; edited 1 time in total |
|
| Back to top |
|
|
gbos
Senior Member
Joined: 21 May 2003
Posts: 26
Location: Athens Greece
|
| Posted: Wed Aug 20, 2003 2:27 pm Post subject: |
|
|
Bagherra
Some of the reasons are:
1. you make uncertain (not risky) bets and you don’t know the odds (in risky bets you know the odds)
2. you can not infer accurate (if any) statistics for a generator process of trades by a sample of trades because the generator might not be stationary (stable over time) or might be left skewed (rare negative events may have tremendous impact on results)
3. maximum observed drawdown depends on the length of time that we observe the process
Unfortunately in many of the cases a margin of safety is not adequate to protect us from a wrong assumption.
|
|
| Back to top |
|
|
Javelin
Senior Member
Joined: 11 Jul 2003
Posts: 41
Location: Georgia
|
| Posted: Wed Aug 20, 2003 8:30 pm Post subject: |
|
|
| So what do you do to protect yourself from a wrong assumption? What parameters are followed to determine if your maxx DD is just that and not indicative of system failure?
|
|
| Back to top |
|
|
ForrestGump
Senior Member
Joined: 09 Jun 2003
Posts: 32
Location: Sydney Australia
|
| Posted: Thu Aug 21, 2003 4:48 pm Post subject: |
|
|
G'day,
Apologies that I am not actually adding anything constructive (or even destructive ) to this thread but I really needed to say - What a fantastic thread - Thanks.
Taking the liberty of speaking for those of us that are only at the beginning of this adventure these examinations and discussions of getting the mindset in place are invaluable.
Thanks.
Fforrest
_________________
Peace
Fforrest
The trend is your friend
till it bends
at the end |
|
| Back to top |
|
|
Kiwi
Roundtable Knight
Joined: 15 Apr 2003
Posts: 505
Location: Gone Too, Ciao.
|
| Posted: Thu Aug 21, 2003 7:27 pm Post subject: |
|
|
Javelin,
You might want to find the answer, such as can be found, by looking at these threads:
Changing system rules: www.tradingblox.com/forum/viewtopic.php?p=2523
How close do you want to get to blowing up: www.tradingblox.com/forum/viewtopic.php?t=177
Montecarlo simulations to protect against ruin: www.tradingblox.com/forum/viewtopic.php?t=139
Thoughts about the fact that your tested sequence of trades is not the only possible sequence - and what happens if the next one is worse or better: www.tradingblox.com/forum/viewtopic.php?t=266
Psychologically robust systems: www.tradingblox.com/forum/viewtopic.php?p=1729
Black Swans: www.tradingblox.com/forum/viewtopic.php?p=1584
My personal approach is to test in depth, broaden the view with Monte Carlo testing, and then to run multiple systems with very different ways of taking money. I also want to understand "why" a system should take money from the markets. Then I trade the systems and will take them past their historical drawdown in comfort -- unless I perceive that the market factors they take advantage of have changed and are likely to stay changed. Because I trade multiple systems on different markets the failure of one won't be catastrophic.
If you change too quickly you fall prey to regression to the mean. Look at what has happened with market volatility dropping off into Aug 2003. If you stopped trading short term systems, which need big days to be profitable, now then you'd be out next week when it started to pick up again (if you get what I mean). Actually that might be a counter example in that you could just wait for an improvement in volatility before restarting. A better example would be trend followers which are suffering from poor trendiness in the currencies - if you stop now you miss the start of the post volatility trends.
John
|
|
| Back to top |
|
|
bagherra
Senior Member
Joined: 25 Apr 2003
Posts: 33
Location: Charlotte, NC
|
| Posted: Thu Aug 21, 2003 9:45 pm Post subject: Expectations for Drawdowns and Returns |
|
|
This is a great thread. Thanks to all of you.
Chuck:
Your explanation demonstrates very well why I might not have seen my biggest drawdown. But doesn't cutting historical returns in half lead one to understate what one should expect? After all, one big drawdown is one drawdown, but returns occur all the time and so the average returns of a twenty year period seemingly wouldn't vary so much between historical and future returns. (I recognize that slumps can go on for a long time, as you described.)
Have I missed something in your argument?
|
|
| Back to top |
|
|
Chuck B
Roundtable Knight
Joined: 17 Apr 2003
Posts: 130
|
| Posted: Fri Aug 22, 2003 11:13 am Post subject: |
|
|
| No, I don't think you have. I just threw out some extremely conservative guidelines that hopefully would undershoot future performance. Over the very long term chances are you haven't seen the biggest drawdown nor have you seen the biggest drawup. Use whatever factor of safety you feel comfortable with as there is no science involved here...just one person (me) guessing at some, hopefully, very conservative numbers .
|
|
| Back to top |
|
|
Kiwi
Roundtable Knight
Joined: 15 Apr 2003
Posts: 505
Location: Gone Too, Ciao.
|
| Posted: Fri Aug 22, 2003 4:51 pm Post subject: |
|
|
Javelin,
Further to last nights posting: another thread that looked at system robustness was:
http://www.tradingblox.com/forum/viewtopic.php?t=287&start=0
Also, adding a couple of points to Chuck's post, some reasons you might expect future returns to be less than tested are:
- testing doesnt allow for the occasional very high slippage events that occur.
- testing doesnt allow for degradation of those characteristics of the market that your system is taking advantage of or the "cyclic" variations that take place over multiyear timeframes.
John
|
|
| Back to top |
|
|
Javelin
Senior Member
Joined: 11 Jul 2003
Posts: 41
Location: Georgia
|
| Posted: Fri Aug 22, 2003 8:50 pm Post subject: |
|
|
Kiwi,
Thanks for the links! Some of them I had read, but some I had not found. I greatly appreciate the effort and willingnes to share knowledge. This is indeed a great thread/site.
Thanks!
|
|
| Back to top |
|
|
gbos
Senior Member
Joined: 21 May 2003
Posts: 26
Location: Athens Greece
|
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum
|
|
FAQ
Search
Usergroups
Register
Profile
Log in to check your private messages
Log in
.