my journal 2

This is a discussion on my journal 2 within the Trading Journals forums, part of the Reception category; It's only fair that just as losses were an obsession in my discretionary trading before I quit, now the systems' ...

Closed Thread
 
LinkBack Thread Tools Search this Thread
Old Mar 29, 2010, 1:56pm   #571
 
travis's Avatar
Joined Mar 2003
more thinking about my systems' drawdown

travis started this thread It's only fair that just as losses were an obsession in my discretionary trading before I quit, now the systems' drawdown (which is the same as losses) are an obsession in my automated trading. And maybe that's why I failed as a discretiionary trader: because I could never see my losses as part of a natural drawdown, and instead I took them personally, each one as a personal failure.

Now it's easier to rationally accept my losses, as part of a natural drawdown. It is so, because I have the back-testing and forward-testing results, and they constantly tell me that every once in a while there will be a loss followed by more wins. The loss is followed by wins because it does not affect the system emotionally and doesn't cause it to modify the strategy. With my discretionary trading instead, losses affected me emotionally, made me upset and caused me to alter my strategy, and would therefore be followed by even bigger losses. The interesting thing is that the same happened with wins, which made me euphoric and overconfident and, because of that, altered my strategy and were followed, once again, by losses and, what's worse, losses that were out of control and blew out my account. This was the pattern:

1) loss
2) get upset, revenge trading... strategy altered (bigger leverage, more risk, less patience)
3) bigger loss ensues (because of leverage)
4) blow out account

1) win
2) get confident, euphoric... strategy altered (less patience)
3) loss ensues (see above for continuation)

With automated trading, these problems are all behind me. After each loss, the strategy stays exactly the same.

Now, back to my drawdown obsession. I tried to make things simpler and summarize them, and here's the situation.

The premise is that I want to diversify, because it's good in itself and it reduces the drawdown. Some systems are better than others, but I want to use them all. Of course, if I can only use half of them, I will use the best ones.

The general trend, is that if I trade a large number of systems, including the ones on the CL, the drawdown increases and so does the profit. But as I said, such a drawdown will increase less than if I trade more contracts on just one system, no matter how good that system is. Obviously if that system never failed, it would make sense to just trade that system and forever increase the number of contracts traded, but that's far from being the case, even for the best system.

Now, here comes yet another simplification of the drawdown situation, this time in terms of monthly gain / drawdown. This is the ultimate simplification of the drawdown tendencies of my systems (and maybe every system).

THIS IS IT

Given my preference for diversification and all other parameters being optimal, the more drawdown I accept, the more it will make each month.

Examples.

Say I trade the best combination of 20 systems, with a total of 30 contracts. I will potentially have... say a maximum drawdown that will last a month and will cause me a potential loss of 5000 dollars, and it will happen one month per year. On the other 11 months I will make 5000 per month. This is more or less the situation.

Now, say I trade a less strict selection of systems with more contracts. I would have a drawdown of 10k, that will happen for one month, and I would make 10k every month for the rest of the year.

Now, all this would not be an issue and I would not have brought it up IF I needed more capital to trade the 10k, and couldn't trade it now. But the dilemma is: since I could start immediately trading the 10k system, which is obviously better (since it will make me 10k per month rather than 5k per month), what's really there to stop me from doing what the most convenient thing is?

The only thing stopping me is that if I incur the drawdown right after I start, I go broke (I have 14k now). If I don't and I take this gamble, then at the end of the year I will have 110k rather than just 55k.

EUREKA!

But I guess I can already see the answer to my own question (it's very useful to think out loud as i did here). The answer is that I should only start trading the 10k as soon as the 5k will bring me to a point where the 10k drawdown won't blow out my account and won't affect my ability to keep on trading. Right. Because starting off with the 5k won't keep me from moving on to the 10k as soon as I have 20k.

So that my equity curve (provided I don't encounter any drawdown) would look like this, month by month:

14k (using 5k combination of sytems/contracts)
20k
25k (I can start trading the 10k combination)
35k
45k (DRAWDOWN)
35k

By using the 10k combination right from the start I simply run the risk of blowing out my account, and gain only one month if I get lucky and the drawdown doesn't happen, since I'll make 10k in one month, rather than in 2 months (which the 5k system would require).

Is it worth it to risk blowing out my account (10% probability of it happening) just in order to anticipate retirement by one month? The last time by pushing my risk and trying to anticipate retirement, I blew out my account and postponed it by one year.

MEMO to myself:
Disable the goddamn CL_ID future until I have enough capital to comfortably withstand its drawdown, which is about 20k.

I guess the reasoning is pretty simple here, as they point out on wikipedia on the kelly criterion entry:
http://en.wikipedia.org/wiki/Kelly_criterion
Quote:
A natural assumption is that taking more risk increases the probability of both very good and very bad outcomes. One of the most important ideas in Kelly is that betting more than the Kelly amount decreases the probability of very good results, while still increasing the probability of very bad results. Since in reality we seldom know the precise probabilities and payoffs, and since overbetting is worse than underbetting, it makes sense to err on the side of caution and bet less than the Kelly amount.
If I exceed in my bet, I will risk blowing out. If I put too little, I will simply risk not maximizing my returns, but certainly I will not risk blowing out my account. As a consequence, I should try to be on the safe side. In a way, I should try to be as much on the safe side as possible, and only try to increase my equity as slowly as I'll need the money. How much money do I need right now? None, because I have a job! So why I am overbetting and hoping to get lucky? I guess it's the same restlessness and thirst for action that lead me to blowing out my account over thirty times as a discretionary trader.
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 29, 2010 at 3:17pm.
travis is offline  
Old Mar 29, 2010, 2:06pm   #572
 
Adamus's Avatar
Joined Mar 2008
Re: drawdowns destabilize you and have ripple effects

I just read somewhere and I think it was one of your quotes here where somebody says "no system is fully automated - there is always a discretionary part: whether you trade or not, how much leverage you use....." or something like that. Sounds like you're proving that.

Another quote, about luck, out of MWs: "You work hard, you make your own luck". That was Tom Baldwin. Not a system trader, but that is my favourite interview in the whole 2 books.
__________________
What matters most is how well you walk through the fire.
Adamus is offline  
Thanks! The following members like this post: travis
Old Mar 29, 2010, 8:44pm   #573
 
travis's Avatar
Joined Mar 2003
self-control update

travis started this thread This is tragic. I came home and the CL system, sure enough, was incurring one of its worst losses in weeks. That system is now at its worst in months.

But that's not it. I wasn't prepared for this loss (now certainly I won't disable it), so I did the usual: discretionary trading and interfering with system. I doubled up and all that. Consequence: I lost 300 more than the system lost.

Since the compulsive trading wasn't as bad as usual (even though really bad for an automated trader), I didn't rate it as a 10, but only as a "5" self-control issue (the maximum is 10, whereas all other values are on a scale of 0 to 1).

snap1.jpg

The trouble seems to be the usual: I take losses personally even if they are caused by a system, and even if they're within its drawdown. Especially the big ones like today. I was going to disable it, as I had written in the post above, because I knew that its losses were too big. Now I can't disable it anymore, because it would not make sense to only have it enabled during its losses. You can't just disable a system as soon as you incur a loss, or you'll miss all the wins.

So I am stuck with this big drawdown system, which trades twice a week on average, and can either cause a 2000 dollar win or a 2000 dollar loss (on average the wins are bigger than the losses, and they're more frequent). Now it's at its worst, and hopefully it won't fail me again in the next few days.

I am really pissed off about today's loss, but in a certain way, I've still got the situation under control, because my discretionary trading fortunately only caused me an extra loss of 250 dollars, rather than the potential disaster of wiping out my account.

Anyway, hopefully I'll stick around long enough with my capital to reap the benefits of what Larry Hite's quote talks about. Hopefully my systems will run long enough to see profit happen.

Hopefully I'll see some profit from the other systems in the next 24 hours, or I'll start feeling pretty depressed. This CL system was the second best one I had. It didn't incur such a big drawdown since October. Since I started forward-testing this system, this is its second worst drawdown. The worst one was 5000. This time we've already exceed 3000. From this point this system could pretty much blow out my account only if doubled its worst (forward-testing) drawdown, and lost another 7000. It would have seemed impossible this morning, when I enabled it, but after seeing it take a 1700 dollars loss, it doesn't seem that impossible anymore.

I really hope I'll be too busy in the future to suffer from the systems' losses like today (and be driven to trade discretionary to make it back), and too busy to fantasize about stellar performances and push my risk management to dangerous levels by allowing trading by systems that can potentially blow out my account by slightly exceeding their biggest drawdown, like I did this time with this CL system. In the weekend I felt restless, and I felt the need to push my risks higher so I'd make money a little faster. And here I am: -2000.

And now I'll go to smoke my last cigarette, which I anticipated on my excel table as another 0.5 off my self-control daily ratings.

It's incredible how frustrated I became in just 24 hours and all due to my overachieving tendencies. Instead of trying to overachieve in terms of risk control, I underachieved at risk control by trying to increase my potential gains. CL_ID is indeed my second best system in terms of ROA and my number one system in terms of absolute gains, but I still sucked. Once again I ****ed up. All thanks to myself. My inner urges. My inner self-destructive overachieving compulsive tendencies...

Quote:
Try to realize it's all within yourself
No one else can make you change


May this crappy song hypnotize me and keep me hypnotized until the drawdown is over - hopefully it is already over. So far the fear of losing has lead me to trade discretionary and lose another 300. Same old story. I don't feel good, I don't feel good, I don't feel good...
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 29, 2010 at 9:35pm.
travis is offline  
Old Mar 29, 2010, 10:56pm   #574
 
travis's Avatar
Joined Mar 2003
sid and nancy

travis started this thread I am watching a movie, to stay distracted from trading and drawdown.

http://www.letmewatchthis.com/watch-3581-Sid-and-Nancy

It's quite interesting. As I watch, I do research. Previously, months ago, I had watched the documentary on Joe Strummer ("The Future Is Unwritten"). The Sex Pistols like The Clash couldn't play for ****, but were hugely successful. They lasted even less, and with Sid Vicious only stayed with them for a year, before the breakup. It's amazing how famous they became for doing nothing good and for so little time, particularly Sid Vicious: one year of this crap and he got so famous. The power of... I don't know. I don't know how this could ever happen. Even madonna is more talented than they are.

Anyway, this is all right:




And these movies on these worthless musicians (maybe "the power of content", maybe they had content) are quite good: the documentary on Joe Strummer, despite his music sucking so bad, makes his story very interesting. Also Sid & Nancy is good.

However, from wikipedia, here's John Lydon's opinion:

Quote:
Lydon commented on the movie in his 1994 autobiography, Rotten: No Irish, No Blacks, No Dogs:

"I cannot understand why anyone would want to put out a movie like Sid and Nancy and not bother to speak to me; Alex Cox, the director, didn’t. He used as his point of reference - of all the people on this earth - Joe Strummer! That guttural singer from The Clash? What the **** did he know about Sid and Nancy? That’s probably all he could find, which was really scraping the bottom of the barrel. The only time Alex Cox made any approach toward me was when he sent the chap who was playing me over to New York where I was. This actor told me he wanted to talk about the script. During the two days he was there, he told me that the film had already been completed. The whole thing was a sham. It was a ploy to get my name used in connection with the film, in order to support it."
"To me this movie is the lowest form of life. I honestly believe that it celebrates heroin addiction. It definitely glorifies it at the end when that stupid taxi drives off into the sky. That's such nonsense. The squalid New York hotel scenes were fine, except they needed to be even more squalid. All of the scenes in London with the Pistols were nonsense. None bore any sense of reality. The chap who played Sid, Gary Oldman, I thought was quite good. But even he only played the stage persona as opposed to the real person. I don’t consider that Gary Oldman's fault because he’s a bloody good actor. If only he had the opportunity to speak to someone who knew the man. I don’t think they ever had the intent to research properly in order to make a seriously accurate movie. It was all just for money, wasn't it? To humiliate somebody’s life like that - and very successfully - was very annoying to me. The final irony is that I still get asked questions about it. I have to explain that it's all wrong. It was all someone else’s ****ing fantasy, some Oxford graduate who missed the punk rock era. The *******."
"When I got back to London, they invited me to a screening. So I went to see it and was utterly appalled. I told Alex Cox, which was the first time I met him, that he should be shot, and he was quite lucky I didn't shoot him. I still hold him in the lowest light. Will the real Sid please stand up?"
"As for how I was portrayed, well, there's no offense in that. It was so off and ridiculous. It was absurd. Champagne and baked beans for breakfast? Sorry. I don't drink champagne. He didn't even speak like me. He had a Scouse accent. Worse, there's a slur implied in the movie that I was jealous of Nancy, which I find particularly loathsome. There is that implication that I feel was definitely put there. I guess that’s Alex Cox showing his middle class twittery. It’s all too glib, it’s all too easy."[11]
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 30, 2010 at 5:04pm.
travis is offline  
Old Mar 30, 2010, 12:18am   #575
 
travis's Avatar
Joined Mar 2003
re: my journal 2

travis started this thread Going to sleep in heaven, going to sleep late I mean. After finishing that movie about crappy music, I had to listen to something pleasant.



"I'll find my way through night and day...". This is music, poetry.

Not the crap of the people above.

I will go to sleep. It was a hard day, all thanks to my own enemies from within: impatience, mainly. What awaits me now is just one big question mark: after losing 3000, will the CL_ID lose more and even exceed its maximum drawdown, dragging me down to blowing out my account? We'll know the answer within a couple of weeks. I can even take another 2000 dollars loss, but not two certainly.
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 30, 2010 at 12:26am.
travis is offline  
Old Mar 30, 2010, 9:22am   #576
 
travis's Avatar
Joined Mar 2003
re: my journal 2

travis started this thread Not feeling good. I am feeling restless, nervous, fearful. Obviously I am overbetting. When you're constantly afraid that something will go wrong... on the other hand, if just one trade went right on the CL_ID, I could trade it from here on without worrying about anything else, since its ratio of wins to losses is about 2 wins every one loss. This is a gamble, that will keep me worried for a whole two weeks, which is the time the CL_ID will take to trade 3 more times, during which I should get about 2 wins and one loss, especially considering the last 3 trades were all losses.

Luckily I am at work, and will be for the next two weeks, so the only danger of interfering is for the hours I will be at home, from 16.00 to 23.00 (CET), which is still a lot of time.
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.
travis is offline  
Old Mar 30, 2010, 10:08am   #577
 
travis's Avatar
Joined Mar 2003
if we go up today, we'll go up/down tomorrow/tonight...

travis started this thread I was thinking of this yesterday. That friend of mine, who once, a couple of weeks ago, told me that OIL could not go higher than 83, because of some OPEC reason, whatever it is. Anyway, it partly affected my choices of yesterday of doubling up my short position and so on. But here's what I should have been aware and should have told this guy, fundamental analyst moron: how do you know your predictions are right or even likely to be right? How can you prove it?

I would have gotten no answer. Because his answer are neither back-tested, nor, in his case, backed by previous successful choices. Fundamental analysts should not even be listened to.

But this is not it: this applies also to all technical analysts, who didn't back-test their theories about the market.

And now I finally get to what I wanted to talk about: the technical theories about the market. If I look at all my systems, and at what theories they're based on, theories all verified with back-testing, I see that mainly there's two ideas:

1) if we go up now, we'll go up later (trend)
2) if we go up now, we'll go down later (reversal)

I've also got systems based on the opening gap theory, which is another thing. But all systems have some features based on these two concepts: continuation and reversal. And generally I can define moments (the difference is mainly that of day vs night) of the day where one thing or the other is true: if we have gone up we will continue to go up, and if we have gone up we will start going down.

Let's look at all the categories of my systems, in (decreasing) order by profit produced, and how they relate to these two different phenomena of rise after rise vs fall after rise, which is another way of saying trend vs reversal. When does trend take place and when does the reversal take place?

WITH ID trend: during the day, whatever the trend it continues
ON bounce: at night, whatever the trend, it reverses
overstretched: after an exaggerated trend, it reverses
Weekday Bias: in certain days of the week, the trend moves faster
Volatility Breakout: when it speeds up, the trend continues
Opening Gap: when trend is in favor, gaps get filled

In other words, as far as my systems, 4 groups out of 6 groups, are based totally or at least partially, on the concept of "trend", a term I never liked. What it basically means is that a move will be followed by another move in the same direction.

The other 2 groups of systems are based on the opposite concept, "reversal": the concept that a move will be followed by another move in the opposite direction. That tends to happen during the night and when the previous move was too long, or, as I call it, "overstretched". For example, if in one day the CL moves by 3 points, that definitely qualifies as "overstretched" to me, and that's where the "trend" principle stops working, and the "reversal" principle starts working. That's what statistics say as far as my futures. But I would be quite confident that it happens in every market, even those I don't know.
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 30, 2010 at 10:27am.
travis is offline  
Old Mar 30, 2010, 4:09pm   #578
 
travis's Avatar
Joined Mar 2003
self-control update

travis started this thread Ok, finally home. Plenty of trades open. Some of them making money. I am not touching anything. Still... feel bored, restless. Maybe I will start swimming at the swimming pool. But first I want to see some real profit, like 2000 dollars more on my account. Otherwise I don't feel good about spending money for the cab, swimming pool ticket and so on. I am punishing myself until my systems make some money. I will force myself into lethargy, to save money.

Today it might be a very good day in terms of self-control, partly simply because i didn't meet any challenges to my self-control and partly because of my acquired good habits.

The good habits are: I've pretty much acquired, since I've started this monitoring, the good habit of not scratching my head. Also, I've quit altogether discretionary trading, except for rare relapses, like yesterday, when faced with unexpected drawdowns. Also, rarer eating binges.

The challenges I didn't meet are things such as: boss was in a good mood and was kind to me and didn't even send me any emails with work to do (I was on my own all day long). Taxi drivers on the way to work and on the way back were both very honest and friendly. No conflicts with people met at work, and actually good interaction with those I met. This is pretty much it. Oh, and I didn't lose any money with my systems today. I am breaking even right now. So these are things that today didn't test me, and as a consequence I didn't feel the need for any urges to smoke/eat/scratch... to vent out my frustration, since there was no frustration. So far this is how my self-control table and chart look like:

snap1.jpg

Now I am faced with a major problem: finding something to do for the next few hours. Otherwise boredom will lead to frustration. Frustration will lead to urges to engage in negative activities (scratching, smoking, eating, trading), and in turn those negative activities will cause me more frustrations and more urges... until death.
__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.

Last edited by travis; Mar 30, 2010 at 4:26pm.
travis is offline  
Old Mar 30, 2010, 5:33pm   #579
 
travis's Avatar
Joined Mar 2003
re: my journal 2

travis started this thread http://en.wikipedia.org/wiki/The_500...gs_of_All_Time
http://en.wikipedia.org/wiki/Like_a_Rolling_Stone



http://en.wikipedia.org/wiki/Bob_Dylan

__________________
Read: E.P. Chan, Cogneau - Hubner, Sewell, Tverberg. Search: expected shortfall, Monte Carlo VaR, extreme value theory. Trade.
travis is offline  
Old Mar 30, 2010, 7:03pm   #580
 
Adamus's Avatar
Joined Mar 2008
Re: self-control update

Quote:
Originally Posted by travis View Post
Oh, and I didn't lose any money with my systems today. I am breaking even right now.
I have to tell you this: you won't be break-even until you make up the money you lost with that ****-up yesterday.

You need to get the first chapter of "Disciplined Trader" tatooed across your forehead.

Otherwise, well done. You're trading. Possibly the start of your new career. I wish I was there too.
__________________
What matters most is how well you walk through the fire.
Adamus is offline  
Closed Thread

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Similar Threads
Thread Thread Starter Forum Replies Last Post
say hello to my little journal gozinsky Trading Journals 13 Sep 27, 2011 10:37pm
my journal travis Trading Journals 2334 Jan 18, 2010 12:57pm
The EA Journal! Victor90 Trading Journals 1 Oct 28, 2008 1:47am

LinkBacks (?)
LinkBack to this Thread: http://www.trade2win.com/boards/trading-journals/85510-my-journal-2-a.html
Posted By For Type Date
Science Forum - why Standard deviation is better than Avarage Deviation This thread Refback May 22, 2010 3:34am

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)