my journal 3

travis said:
Damn. I have to admit it's really getting to me. The other day I said I wouldn't care about the ratings but now there's an ongoing war between me and the mastermind of evil. Every day I come online and see that someone has given me a few more one-stars. Obviously this is not just the consequence of regular readers, but someone doing this on purpose. I know the stats of this from my previous journals.

So, it was ok as long as I had just all one-stars and a perfect score of 1. Then some good guy (or maybe it was the evil guy again) rated me a few five-stars, and so I got into this sick mania of wanting a good score again. That's when I rated myself five-stars, not just with my main user id, but also with my other nicks, of which two have now been banned.

But then the evil asshole kept on rating me down, so I've been registering again and again, trying to give myself a higher rating. But this maniac is more of a maniac than I am, so he keeps on dropping by, probably with new nicks (or he has a lot of hyenas-friends), and lowering my rating.

So I don't know what I'll do. Also because t2w got smarter and now I don't know if it's by ip or by computer id, but they basically ban you from the start. You don't even have the time to log in the first time and you're already banned.

I'll probably stop caring or if I still care and cannot manage to achieve a higher rating, I might get discouraged altogether and stop posting, and this asshole will be happy because he won in the end and made me stop posting here.

I wish this guy just died, whoever it is. Even if it's a child.
Click to expand...
..
 
Last edited by a moderator:
Back from being banned.

I've done a few math exercises in the meanwhile:
Snap1.jpg

92% of work done. 20 exercises to go.

Great web application on this exercise:
Derivative intuition | Khan Academy

Great links on solving systems of equations:
System Solver
Advanced Solve Equations :: QuickMath.com - Automatic Math Solutions

These last two links above can solve these problems:
Rate problems 1 | Khan Academy

 
Last edited:
I rated you one star for teh lulz.

You really shouldn't get hung up on such utter and meaningless sh*t. You know how many 5 star threads filled with poison and tripe are on this site?

Good luck with your trading :)
 
weekly update (forward-testing)

Thanks for the welcome back.

Finally the big surprise I was hoping for:
Snap1.jpg

We stopped trading on September 26th, and that, so far, was the lowest point ever reached. Damn, doesn't it suck for me.

After that, the selection of systems traded went nowhere for two and a half months, and now it resumed its rise (or so it seems at this time).

As I have always thought, the systems were not unprofitable. We simply made big mistakes in the assessment of the drawdown, which is why I am studying math right now, so to be able to tackle portfolio theory one day.

I've been monitoring that equity line because i want to know how it would have ended had we kept going, and not because it's the best combination of systems, because it's not. It simply is the one combination that worked best in the past in terms of maximum historical drawdown vs. historical return and i chose it by testing hundreds of different combinations in terms of return/drawdown. Choosing what worked best is curve-fitting. That's where we made a mistake.

I must instead find a way to appraise what is most likely to work best, and only then find out how it worked in the past (and base on it my calculation of capital required). Going the other way around is like abusing the data, cheating, and ignoring the out-of-sample methodology, which I've been applying for creating the individual systems.

If anything, rather than curve-fitting the portfolio combination as we did, it would be better to simply trade all the good systems with 1 contract, after making sure there's enough capital to withstand their historical relativized drawdown times two:

Snap2.jpg

Once I will have saved a few more hundreds dollars, maybe in two months, I will resume my trading by applying this simple methodology with the three or four systems that have emerged in the chart above and that I can afford (I need systems with high sharpe ratio, low margin requirements, and low drawdown), and they'll be chosen based on their individual performance, rather than by testing hundreds of combinations of systems and choosing the ones producing the best combined performance.

It won't be much capital nor much profit, but at least i will feel the pleasure of doing something right, and making money with it. Regularly, and without worries, for the first time. It will pay for the server and it will build up capital. And if it won't work, it's better if it fails with little money. If it works, before the capital will grow, I will have the time to learn the portfolio theory I will need for combinations of more than four systems. One system doesn't need portfolio theory. Two systems with futures of equal size don't need it either. As the number of systems/futures grows portfolio theory is needed more and more. But even if I never figured out a portfolio theory which is good enough, I now know that no portfolio theory and trading 1 contract per good system (those with the highest sharpe ratio) is better than a wrong portfolio theory.
 
Last edited:
Travis, I don't judge you. I just try to look at the facts and you get no recommendation from me, because:

On the positive side, YOU

- killed that fat Boss Hogg wannabe in Godfather II, and introduced olive oil to the US
- saved an innocent girl in Taxi Driver
- scared the sh*t out of Billy Crystal in those 'analyze whatever' movies
- scared the sh*t out of Ben Stiller in those 'meet whoever' movies
- did all you could for your pals in 'Nam

BUT, on the negative side, YOU

- were an out and out unrepentant, uncontrollable as**ole in Mean Streets
- raped a woman in Once Upon a Time in America
- let wee Joe Pesci do wtf he wanted in NY
- let wee Joe Pesci do wtf he wanted in Vegas
- killed Bambi, or at least you got tooled up and tried your best

So, it's in the balance, but you weren't in The Sopranos, and you let wee Joe Pesci push you around in Rocky, so, it's a no, capisci?

Merry Christmas
 
Yes, and good job watching all those good movies. Merry Christmas.

 
Last edited:
Re: weekly update (forward-testing)

travis said:
Finally the big surprise I was hoping for:
View attachment 126340

We stopped trading on September 26th, and that, so far, was the lowest point ever reached. Damn, doesn't it suck for me.

After that, the selection of systems traded went nowhere for two and a half months, and now it resumed its rise (or so it seems at this time).

As I have always thought, the systems were not unprofitable. We simply made big mistakes in the assessment of the drawdown, which is why I am studying math right now, so to be able to tackle portfolio theory one day.
Click to expand...

Good to see you back.
FWIW, I think the only mistake made was quadrupling risk in the middle of the euro sh1tstorm.

My solution to that kind of event is discretionary shutdown.
I can appreaciate that is not what you want, so hats off to you for looking for
a solution that allows permanent always on automation.

My own setup is coming along and still on track, forward sim testing since beginning of november - results
closely mirror backtest results of same period, which means the backtest is looking like a realistic pointer (still early days though).
I did manage to reduce trade frequency as well, and backtested from 2004 instead of 2007.

I've gone the same route as you with server as well, looked into remote dedicated servers and the quality (tech support and uptime)
was generally not as good as I could do myself.
Zero downtime or tech hitches so far.
Server and connection are only needed at 2 specific times each day, so tech issues less important anyway as compared to 24/7 server.
No need to reply to any of above (just me waffling :) )
 
Last edited:
Yes, good point on saying that I want "permanent automation". Congratulations on your efforts and progress.

Regarding the server, we found the perfect solution, cheap and effective, even though by "cheap" I mean 300 dollars per month, because it's not at my place, but it would be much cheaper if it were done at your own place, and it would be almost as effective (the only problem might be when you travel, given that it's not safe to leave a computer running at home).

What we used is a laptop (Toshiba). With that, you solve power outages effectively, because UPS are extremely expensive (over 1000 dollars) if you are looking for one that provides power to a desktop computer (or similar) for several hours. The laptop has a battery that lasts several hours of power outage (which can only last 2 hours at most), and it costs less than 400 dollars. On top of it, it uses up much less space, and it has a screen of course, so, all in all, it uses up 20% of the space used by desktop + keyboard + screen + UPS. And you end up spending less than 400 dollars instead of over 1500, which is from the 500 for the desktop plus the 1000 or more for the UPS, and have a much more effective solution (less space, more battery hours).

For the connection problems, you get a secondary connection that takes over if the first one fails. With this setup we have not had any problems at all ever since we implemented it in early August of this year (so it's been 4 months).
 
Last edited:
2 exercises done, 12 exercises to go:

Snap1.jpg

Once I'll be done with everything, I will need to go over the cartesian representation/equations for circle, ellipse, hyperbola and parabola. Right now the objective is to finish 100% of exercises, and keep moving at the pace of at least one exercise/subject per day. Then I'll spend a few weeks reviewing everything, and clarifying all the implications of all material covered. I know that if I stop, I might never resume. I just need to keep going for 12 more days (or less) in terms of quantity, and then I'll focus on quality.

Beethoven. Symphony No.7 in A Major, Op.92 - II. Allegretto - YouTube
 
Last edited:
preparing to resume trading

I'm gonna restart with these four:
Snap1.jpg

And I am going to expect a profit of about 1000 per month. I can't restart yet, and I will wait one more month, for two reasons:

1) I don't have the 3000 (not willing to wire it) to withstand the potential drawdown.
2) I want to get used -- I need to get used -- for one more month to save money and know what it feels to count every cent.

And 3) i don't want to rush things. I need to plan and make myself wait, just in order to see if it's a long term plan or just something on the spur of the moment. For example, those systems above, I mentioned them several weeks ago, so I know it's well thought out.

After I resume trading at the start of February, I don't expect much. Just to grow capital and at the same time pay for the server. There cannot be any scaling up for a long time, otherwise sooner or later the drawdown will hit me.

In the meanwhile, I will study math, and then, when I know enough of it, I will tackle portfolio theory. When in doubt, I will have to underbet, as they say here:
Kelly criterion - Wikipedia, the free encyclopedia
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...
Click to expand...

The quote is about the kelly criterion, but it works for any risk management theory. That's why I didn't quote the whole thing. If you don't push your risk, you will make money more slowly. If you push it, you will lose money. So it's better to risk making money slowly than to risk losing it.

Today I took the subway for the first time in a long time. If I take the subway instead of a taxi to save a few euros, it's going to be harder to waste my money not just by going to a restaurant, but also via trading. I already know the effect of a saving/stingy attitude on restaurants. This girl asked me to take her out, as usual, and this time I replied "no way, I am walking to the taxi station to save money, so I am not going to waste it at a restaurant". Now the question is if it will work for trading as well. But I think it will. The whole population could be divided between people who check their balance and people who don't. And for the first time, and at least at the moment, I am someone who checks his balance, and I have realized that money is not unlimited. And that only so much money can be made with trading, and if you try to make more, that's when you start losing it, by overbetting. So the first step to underbetting is to not need more money than you can make, or that need will compel you to overbet.
 
Last edited:
equation of an ellipse

Good links on the ellipse:
Ellipse - Wikipedia, the free encyclopedia
Ellipse -- from Wolfram MathWorld
Equation of an Ellipse in standard form and how it relates to the graph of the Ellipse.
((y+11)^2)/4+((x+6)^2)/64=1 - Wolfram|Alpha

I am going to do this today.

This is even simpler:
(y^2)/9+(x^2)=1 - Wolfram|Alpha

The problem is the scale of the chart on wolfram alpha, which doesn't show you it's an ellipse.

Three more graphing calculators (I need to add them to my math links):
Graphing Calculator
Free Online Graphing Calculator-Graph functions, find intersections, table of values, pan, zoom, & export as image
Desmos Graphing Calculator

Ok, this is perfect, as it doesn't use a log scale for the ellipse:
https://www.desmos.com/calculator/

Snap1.jpg

The cartesian coordinate system is new to me and probably I will benefit somewhat from studying it, in that it has equations in it, and it has charts, both of which I somehow need for portfolio theory and probability. New world to me and fascinating:
Cartesian coordinate system - Wikipedia, the free encyclopedia
The invention of Cartesian coordinates in the 17th century by René Descartes (Latinized name: Cartesius) revolutionized mathematics by providing the first systematic link between Euclidean geometry and algebra. Using the Cartesian coordinate system, geometric shapes (such as curves) can be described by Cartesian equations: algebraic equations involving the coordinates of the points lying on the shape. For example, a circle of radius 2 may be described
Click to expand...

The child next door is crying again. She's really not that good at making him stop. He's been crying several hours per day throughout his 3 year-long life. They must be doing something wrong. Actually this child cries about 33% of the time. Considering he sleeps one third of the time, he cries 50% of the time he's awake.

http://en.wikipedia.org/wiki/Ellipse
Circles are special cases of ellipses, obtained when the cutting plane is orthogonal to the cone's axis.
Click to expand...

[...]

Today, I was tired and instead of taking the subway, I took a taxi all the way to home (without getting off early, at a bridge on the Tiber, like I've been doing lately). It cost me 15 euros. This showed me that time is money and also sleep is money, because when I am rested I can save on cabs. So if you really want to be saving effectively, you also have to sleep well. Another limited resource, in an inverse sense nowadays, is, rather than food, space in your body to allocate food, otherwise you get fat, and that's like wasting health and other resources. There's a whole universe of areas where one should be stingy, but the main ones are: eating (which comprises eating and drinking), sleeping and money. Also, health should be mentioned, but that's too broad because it could include everything else. But for example smoking makes you waste your health, money and time at once.
 
Last edited:
Ok, done for the day. They added 4 new exercises, but I did 5 of them. 96% of work done, and 11 exercises to go:

Snap1.jpg

This is the best online graphing calculator I've found in months of web browsing:
https://www.desmos.com/calculator/

Snap2.jpg

Its best feature is that as you adjust the values, you see the chart change. Here's how it works:

exploring sine - YouTube

I added it to my autohotkey shortcuts, so I can practice with it some cartesian skills, which I really lack.
 
Last edited:
You must log in or register to reply here.

Similar threads

Yamato
  • Locked
my journal 2
169 170 171
Replies
3K
Views
465K
timsk
timsk
J
  • Poll
Trading journals: general discussion
Replies
5
Views
2K
athvia
A
N
Ninja's Journal
Replies
5
Views
1K
Ninja377
N
G
need help with Journal template and macros
Replies
0
Views
2K
grego99
G
T
Trading Journals and No Trading Journals
Replies
14
Views
3K
hhiusa
hhiusa
Top