my journal 3

Nice newspapers titles on the Pope here:
Foto Papa Francesco: le prime pagine nel mondo / 1 - Repubblica.it

This is a nice one, for example:
143341968-20e7ac38-3eb6-4a41-a14d-69a07d2c8f8d.jpg

If it weren't for the fact that I don't believe in god, with a pope like this, I might even turn catholic again.

Remember The Mission?

The Mission (1986 film) - Wikipedia, the free encyclopedia
The film is set in the 1750s and involves Spanish Jesuit priest Father Gabriel (Jeremy Irons) who enters the South American jungle to build a mission and convert a Guaraní community to Christianity. The Guaraní community above the perilous Iguazu Falls has tied a priest to a cross and sent him over the falls to his death. Father Gabriel travels to the falls, climbs to the top, and plays his oboe. The Guaraní warriors, captivated by the music, allow him to live...
Click to expand...

This pope reminded me of that movie, because of his being a Jesuit from South America.

The Mission - Gabriel's Oboe - YouTube

Morricone: Gabriel's Oboe (The Mission) and Main Theme from Cinema Paradiso - YouTube
 
Last edited:
Ok, today capital briefly touched 35k, so I am officially in terra incognita.

However, my goal, with all these positions still open, was to at least go above 40k, so I'll definitely wait a little longer before closing any of them. Also, because they all have so much reversal potential, and yet I do not know which one will explode to the upside ("downside" for GBL), I need to keep them all open. Also none of them is overbought yet (oversold for GBL). So I will just wait until the account balance reaches 40k. I mean, it only makes sense after suffering for a whole week when the reversal wasn't there yet.

I guess they're all doing really good, but what I am still waiting for is a few green candles on JPY and a few more short candles on GBL.

But if today my combined trades brought me to 39k, I would probably close them, at least GBP and ZW. If they don't, then I'll wait up to another week, on all four trades.

...

In the meanwhile the gorillas in the room next to mine, are making the following noises: one is roaring, one is stamping his feet relentlessly, and another one is whistling. This is people. This is the people that I want to avoid. I am glad I am not in their room. As a rule, people are trouble.

Gorilla Communication
Researchers have identified 25 different sounds that gorillas made and what they mean. There are still many more though that we don’t fully understand yet. They make chirps, grunts, roars, growls, and even hooting like an owl at times. They certainly can be funny too, laughing in a way that seems like they are making fun of others around them and even the famous sticking their tongue out.
Click to expand...

Yeah, that's it. That's my colleagues.
 
Last edited:
Yo, I drank some hot chocolate, ate a chocolate chip cookie as planned, came home, and I engaged in no compulsive gambling today. I will keep testing this thing advised by Adamus to see if it really works, or if it is a coincidence. So far it was effortless, right after drinking chocolate and eating sweets.

Capital is now at about 35k. I will have to wait another week to see it go beyond 40k.

...

Systems lost about 100 dollars this week. It certainly seems hard to keep trading automated when I am having a week of 1500 dollars, then another week with -100... not fun. In the meanwhile, this week I've seen my capital rise 5000 dollars thanks to my discretionary trading... I am very tempted, if I get out of this huge trade alive, and above 40k, to keep going with discretionary trading. If I do close this trade with capital above 40k, I can indeed claim that 75% of my profit came from my discretionary trading. I mean, whether I suck or not, whether I've been lucky or not, this is a reality that cannot be ignored. Maybe I can fine-tune it and keep at it. I think I am starting to be quite good at it. But in my own way, a way that abhors stoplosses.
 
Last edited:
I've been working with my math tutor, in my 4 hour long class yesterday on a new money management methodology. It works as follows.

You assess each of the best 30 systems (out of the 120 I have) by standard deviation.

You treat them all equally as far as performance (including margin required), in order to maximize diversification, and to simplify.

Furthermore, for simplification purposes, you assume they're not correlated.

Last, and most importantly, you don't take into account their frequency of trading, which is very debatable, and maybe should be changed.

All you take into account is their standard deviation.

This came from Adamus' suggestion "dude, don't make it too complicated: just invest 1% on every trade".

This, applied to systems, translates into the standard deviation of the trades by a system.

At this point... idea2develop, let's say you have ZN_ON_02 which has a standard deviation of 200 and NG_ID_02 which has standard deviation of 400.

You will accordingly allow 2 ZN contracts every NG contract you allocate.

A possible doubt and objection is that, given that ZN trades once a month and NG trades 4 times per month, shouldn't we weigh their contracts accordingly, and allow four times as many contracts on NG?

This in turn has a counter-objection. If the standard deviation is 200, and you allocate 2 contracts so you make it equal to NG, what happens if you multiply those 2 contracts by four (given the lesser frequency of trading) is that you get 8 contracts, which, when the loss happens, will cause you a huge drawdown.

You follow me?

While it is true that NG could incur 4 losses in the span of a month, when ZN only trades once, it is also true that each individual trade should be treated as a trade, regardless where it comes from, and only weighing the standard deviation.

In other words, we are focusing on the trade (produced by the system) vs. the characteristics of the specific system.

Given that there are so many systems, it seems to make more sense to focus on the risk coming from the next trade (coming from whatever system), than to focus on the risk coming from ALL the trades from a specific system.

It's very complex and I get lost just talking about it, but in brief, it makes more sense to forget which system the trades are coming from (provided they're all profitable), and only focus on their standard deviation, totally forgetting about its performance and the frequency of its trading.

This is not complete, but the sharpe ratio, as far as I could use it, has even more problems and risks of not being used properly, and leading to bigger mistakes than this method.

We're in the business of simplifying things. My investors didn't get it right, despite being knowledgeable in this field. They got it very wrong, and back then, I did as well.

Only now I grasp the complexity of this subject, even without having conquered it completely.

Recapitulating this simplification goes like this:

1. You select your best systems and forget about the others
2. You assume, simplifying, that all systems are identical in frequency of trading, margin required, performance except for...
3. ...you do analyze their standard deviation and adapt the contracts so that each system's standard deviation is the same
4. you adopt a fixed fractional approach otherwise, allowing systems with an increasing standard deviation to trade.

Fixed fractional adapts to standard deviation as follows.

When you have 10k of capital, you only allow systems with 200 of standard deviation to trade.

When you have 20k of capital, you can allow systems with 400 of standard deviation to trade. At this point you will therefore enable 2 contracts for those systems that have a 200 dollars standard deviation.

When you have 40k of capital, you will allow systems that have a standard deviation of 800 dollars, and will double the contracts on those that have 400, and quadruple the contracts on those that have 200.

If capital will decrease, you will go back to the previous phase of standard deviation.

Roughly speaking I could devise a completely automated formula whereby your level of capital gets divides by... 50, to tell you what standard deviation you can allow and how to adapt your contracts accordingly.

40,000 divided by 50 is 800, and therefore I could allow 1 contract for every system that produces trades that have a standard deviation of 800, and 2 contracts for those with 400, and 4 contracts for those with 200.
 
Last edited:
I am doing all right.

As expected, I am now getting tired of having open positions, just as I was getting restless about not having any open positions before.

For some reason, I am not happy without action, but this also means that I am not happy with open positions for too long, because they mean no action either. I am happy trading. I am trigger happy.

This math is great in that it keeps me busy, and he even gave me some assignments this time, to run on my blender (cfr. previous posts on the "blender").

I am doing them now.

Yeah, why not, I am gonna post them. I've been discussing them in the previous post anyway. Some of it is in Italian, because we speak Italian. Of course you can always use google translate, provided anyone is interested enough to download my excel file:
View attachment paragoni_money_management.xls
 
Last edited:
Ok, I've been sleeping on it, and today I have some doubts, changes in mind.

My approach only considers standard deviation. What if I were to consider the standard deviation of losses?

That might produce some unexpected errors, due to my lack of familiarity with formulas. Ok, change discarded.

Up to now I was using average loss, and then I was multiplying it by the frequency. However, this was discarded, because I said that what counts is the specific trade.

idea2develop

What if I were to retain the concept of average loss and use it to replace standard deviation?

Let us see how correlated the two are...

Holy cow, 0.93.

They're pretty much the same exact thing.

With one big difference though. The concept of average loss is much simpler to me, and simpler is always better (all other things being equal).

Another idea that I was entertaining, and that I will probably discard, is this: if I have a system that wins all the time, this money management shouldn't apply to it. But I might discard this, because I do not have such a system, nor one that is so close to winning all the time, that it can be treated as such.
 
More thinking.

idea2develop

Shouldn't the systems have all the same weight and how do you go about measuring this?

Certainly, you can't use margin, nor can you use the size of the specific trade, can you?

Can you use total profit by system?

I remember Adamus, as many other traders say, advised to use a risk of 1% of your capital per trade.

This is tough to estimate for stock trading, but it might be a little easier for me, given that I have all the stats.

Then, provided we assume that it makes sense to approach each trade in the same way, regardless of what system they come from (except for the relevant statistics of that system, that tell us the potential risk), I would still have to address the following question:

how do we assess that 1% risk? Or whatever x % we mean to use.

Would that be the maximum loss by a given system?

Would that be the standard deviation?

Of all the possible measures of risk I am opting for the average loss, because you see, the maximum loss could be the consequence of chance... much more than the average loss could be.

Love Story(1970) - Theme From Love Story (Finale) - YouTube

Regarding my automated trading, I am at about 35k, and if I can hold out until the second part of next week, I should be able to go above 40k.

Snap2.jpg

If I can finally go above 40k, then this money management will prove to be absolutely necessary, because I'll be trading a pretty large number of systems/contracts. Everything will then have to be placed in the hands of my systems, because mentally I cannot handle such a big capital (by my standards).

...

Part of this holding out for a few days, of boredom, will consist of not being home: going to movies all day long, swimming before that, and then the math lesson on Tuesday, it will be tough.
 
Last edited:
travis said:
how do we assess that 1% risk? Or whatever x % we mean to use.

Would that be the maximum loss by a given system?

Would that be the standard deviation?

Of all the possible measures of risk I am opting for the average loss, because you see, the maximum loss could be the consequence of chance... much more than the average loss could be.
Click to expand...
Personally I base it on max loss per trade, but then I do have a hard stop,
so its easy for me to calculate (slippage aside, even then I am aware of likely extreme worst case 5-6R)
My average loss is lower than max loss, as hard stop is just max failsafe risk.
Its not usually hit.
I can see what you are saying, for me, average loss would magnify drawdown,
or else, risk per trade would have to be reduced to compensate.
Thats why I prefer max hard stop loss as its easy to quantify.
Thats only my personal view as it applies to me.

travis said:
Regarding my automated trading, I am at about 35k, and if I can hold out until the second part of next week, I should be able to go above 40k.

If I can finally go above 40k, then this money management will prove to be absolutely necessary, because I'll be trading a pretty large number of systems/contracts. Everything will then have to be placed in the hands of my systems, because mentally I cannot handle such a big capital (by my standards).
Click to expand...

(y)
 
Thanks for the feedback, quite detailed.

I see your point, but in my case the max loss has the following problem: some systems have only produced 12 trades, and using max loss on those might give me a number that is too low. Other systems have traded 300 times, and using max loss on those might give me a number that is too high. In other words, average loss seems a much more balanced and adequate value to use, in that it also has almost symmetrical matching (using excel's correl() function) with standard deviation of all the trades.

As I added later in my post, the biggest task right now, for next week, will be to keep busy and not come back home before 8 pm, because otherwise I'll incur my usual restlessness in the face of profit. When I lose, I find it impossible to close the trades. When I win, I find it impossible to let profits run.

So I must now try to not even come home, for an entire week.
 
travis said:
Thanks for the feedback, quite detailed.

I see your point, but in my case the max loss has the following problem: some systems have only produced 12 trades, and using max loss on those might give me a number that is too low. Other systems have traded 300 times, and using max loss on those might give me a number that is too high. In other words, average loss seems a much more balanced and adequate value to use, in that it also has almost symmetrical matching (using excel's correl() function) with standard deviation of all the trades.

As I added later in my post, the biggest task right now, for next week, will be to keep busy and not come back home before 8 pm, because otherwise I'll incur my usual restlessness in the face of profit. When I lose, I find it impossible to close the trades. When I win, I find it impossible to let profits run.

So I must now try to not even come home, for an entire week.
Click to expand...

True, average loss will maximise profitability.
I don't use it simply due to increased exposure to outliers.
I spose its ultimately a question of personal risk profile, I'd rather keep
risk as low as I can even if profitability suffers.
Basically I have a low risk tolerance.
I'm always thinking of the worst case scenario.
That doesn't mean its the right answer for you, only you
can decide that :)
Its just my personal approach.
 
CONFISCATION: Panicked Europeans Rush ATMs as Leaders Move To Seize Funds Directly From Bank Account Holders
For those who don’t believe the government is prepared to take extreme measures that may include the seizing of retirement accounts, cash savings or even gold, look no further than Cyprus, the latest recipient of bank bailouts.

As of right now, citizens of Cyprus are scrambling to withdraw funds from their bank accounts after the EU, with agreement from the Cypriot government, announced they will decimate funds held in personal bank accounts to the tune of up to 10% of existing deposits.
Click to expand...
 
I need some more reasoning out loud. I'd need others to quiz me on the possible implications of this money management I am implementing.

A big part of this came from Adamus' suggestion, so I hope to get some feedback from him.

Let's go ahead with some brainstorming.

I have my 30 systems, and they trade different futures, using different margins, producing different trades, with a different frequency, and have a different performance, inevitably.

Given all these characteristics, I must implement some simplification, because I am unable to come up with a formula that takes into account the dozens of characteristics of the dozens of systems that I am using.

So far so good.

Can a simplification lead to an efficient formula, or should I give up and proceed "by eye", as we say in Italian, by guesstimates?

Well, if I am taking math, and delving into portfolio formulas, it's because my objective has been to create a formula, so let us assume that it is worth doing.

We are also assuming that a simplification is needed, and therefore we must discard some or even most of the characteristics of all these systems.

Discarding performance
We're discarding performance and favoring diversification, by analyzing all the systems that have an excellent performance, such as a sharpe ratio above 1.5. At this point, we consider them all equal.

Discarding profitability based on margin
This is part of discarding performance, in that two identical systems requiring different margin do not have the same profitability, but we have to ignore this, too.

Discarding frequency
We're discarding the frequency feature of performance. Two systems with a sharpe ratio of 1.5 could produce, all other things being equal, different total profit, merely based on the frequency of their trading. This could be considered as well as part of discarding performance, but it also plays a role in risk, because a system that loses 200 at worst and it trades once a year, doesn't implicate the same risk as a system losing 200 at worst and trading once a day. But I'll have to ignore this, too.

I will be discarding everything except the average loss produced by a system.

Then I'll allow... as many contracts for each system, as result from dividing my capital by 100 (or another coefficient of my choice) and then by the average loss of that system.

So, say I have 50k and I divide that by 100. I obtain 500 dollars of loss tolerance. If a system has an average loss of 250, it will be allowed to trade 2 contracts. If it produces a loss of 400, it will be allowed 1 contract. If it has a loss of 600 dollars, it won't be allowed to trade.

Now up to you to quiz me and find the flaws of this methodology.

One thing that I do notice is that this formula is extremely simple to calculate, to use, to understand and that it's the closest thing to my allocation of contracts by eye, by guesstimates which usually take as long as month of reasoning, whereas this is automatic.

...

I am starting to think that, although "standard deviation" is not as clear and direct as "average loss" in its relationship to potential losses, it may be more accurate in representing the potential risk by each system.

I noticed this from the extreme value of contracts allocated to AUD_ID_01, which has traded only 12 times, and seems to have too many contracts according to average loss, but this seems to be corrected by using standard deviation.

...

Bingo!

Yes, indeed.

This is the old debate about sharpe ratio vs. sortino ratio and I believe sharpe ratio still wins, not only over sortino ratio but also over my concept of average loss. And it wins for the same reason and the reason is as follows.

A loss, or better, a potential loss is better represented and forecasted by monitoring all variability of a system/asset rather than just its downside variability. Indeed, the list of trades by a system/asset is nothing but the range of movements after you start the trades, and that range of variability, whether it turns out to be a "loss" or a "win", is a good measure of potential losses more than just measuring the ranges where the trade turns out to be a "loss". In other words, sharpe ratio and standard deviation are better than sortino ratio and my concept of average loss, in that they use a larger sample, by not mistakenly discarding that part of cases that turn out to be "losses".
 
Last edited:
This is where I'm off piste with portfolio management.
I specifically avoided it because of these issues, in fact that was the
primary factor in deciding to focus on one instrument for me.
Backtest and forward test sample size is larger for the former,
and quicker to build for the latter.

Although individual strategy costs are reduced, the cumulative portfolio
costs are similar, but without the sample size per strategy.
I'm not saying I disagree with your approach, just that it is much harder to
empirically test, so I chose not to.
Due to that, I can't really add any constructive input on portfolio weighting.

If I was in that position, I guess I'd whittle down the number of strategies
to the 5-10 best performing and largest sample size strats and give them
all equal weighting (once basic risk level, av. or max loss taken into account).
That would free up capital from lower sample size strats to put into
the higher sample size strats.

If I did then adjust weighting I'd probably do it based on drawdown.
I'm not a massive fan of sharpe or other metrics of that type due to
the flaws they contain in the calculation.
http://www.ilukacg.com/articles/All Hail the Sharpe Ratio.pdf
http://www.bus.lsu.edu/academics/fi...h/ThreeMethodsforImprovingYourSharpeRatio.pdf
If you take account of those flaws, no big deal :)
Like I said at the beginning, portfolio management is not an area
I have a great deal of knowledge in, so all I've just said may be no use to you :)
 
Replying as I read.

I know the feeling of being off-piste, having been spending years on this.

Yes, I could not avoid it because I started by creating the systems, and I came to money management only at the end.

Yes, my burden as far as portfolio theory is much harder than with one strategy, but I have chosen my penance and it doesn't hurt as much. To compensate for this burden, I have the advantage of diversification. My strategies are probably not as good as yours, and I need many of them to compensate for quality, lack of programming skills, and the drawback is that I need much more work on portfolio theory.

By all means I would not discard any more strategies, because these 30 I have selected are quite good, from all points of view.

Oh, capital is not a concern, because my limits on investing is more from the potential drawdown than from lack of margin. Given that these strategies all trade at different times and rarely overlap (less than 25% of them overlap), I will never have any problems of margin.

Furthermore, when I speak of 12 trades, mind you, I am not talking about back-tested sample, which is over 10 years for all 120 systems (including an out-of-sample). I am referring to the forward-tested period, which is at least 2 years for all systems, but sometimes this only means 12 trades, as it is for AUD_ID_01.

Drawdown is a concept that I consider flawed, because it is a function of chance, and the fact that a series of losses piled up at a given time doesn't mean the same sequence will happen in the future (maybe worse, maybe better). I consider the distribution of trades (as measured by the standard deviation) to be much more reliable. In other words, I still consider all trades, but not their sequence.

This could be flawed as well, and it does miss some useful aspects of looking at drawdown, but it consider drawdown's disadvantages to be more.

Thanks for the articles against the sharpe ratio. I will read them, although standard deviation is not exactly the same as the sharpe ratio, and it may still be the most effective measure of the distribution of trades. Also, I disliked it for a long time, but I keep seeing its effectiveness.

It was all very useful. Thank you very much.
 
travis said:
Yes, my burden as far as portfolio theory is much harder than with one strategy, but I have chosen my penance and it doesn't hurt as much. To compensate for this burden, I have the advantage of diversification. My strategies are probably not as good as yours, and I need many of them to compensate for quality, lack of programming skills, and the drawback is that I need much more work on portfolio theory.
Click to expand...
I wouldn't necessarily say, that, I am more exposed to outliers with
just one instrument and strategy as well, although I plan on rectifying that
in the future.
On the quality issue, who can say, I can't say that for sure, I don't know what
you are doing, and I myself don't have the automated live trade history you have.
Anyway, simplicity is a form of quality in my book.
You know yourself, its much more to do with the thought processes rather
than coding skill, which can be overcome in various ways.

travis said:
Drawdown is a concept that I consider flawed, because it is a function of chance, and the fact that a series of losses piled up at a given time doesn't mean the same sequence will happen in the future (maybe worse, maybe better). I consider the distribution of trades (as measured by the standard deviation) to be much more reliable. In other words, I still consider all trades, but not their sequence.
Click to expand...
That is a fair and valid point, can't disagree with that.

travis said:
Thanks for the articles against the sharpe ratio. I will read them, although standard deviation is not exactly the same as the sharpe ratio, and it may still be the most effective measure of the distribution of trades. Also, I disliked it for a long time, but I keep seeing its effectiveness.
Click to expand...
Yeah, it isn't useless, and if you are aware of the flaws, no harm in using it.
Good discussion as always :)
 
Yes, I agree with your points, and I am a big fan of simplicity as well, and of being orderly. Thanks for sharing your insights.
 
travis said:
Yo, I drank some hot chocolate, ate a chocolate chip cookie as planned, came home, and I engaged in no compulsive gambling today. I will keep testing this thing advised by Adamus to see if it really works, or if it is a coincidence. So far it was effortless, right after drinking chocolate and eating sweets.

Capital is now at about 35k. I will have to wait another week to see it go beyond 40k.

...

Systems lost about 100 dollars this week. It certainly seems hard to keep trading automated when I am having a week of 1500 dollars, then another week with -100... not fun. In the meanwhile, this week I've seen my capital rise 5000 dollars thanks to my discretionary trading... I am very tempted, if I get out of this huge trade alive, and above 40k, to keep going with discretionary trading. If I do close this trade with capital above 40k, I can indeed claim that 75% of my profit came from my discretionary trading. I mean, whether I suck or not, whether I've been lucky or not, this is a reality that cannot be ignored. Maybe I can fine-tune it and keep at it. I think I am starting to be quite good at it. But in my own way, a way that abhors stoplosses.
Click to expand...

I suspect that your impulse to enter discretionary trades today would be reduced regardless of your blood sugar levels, because you are already in several trades.

The material I am reading on neuroscience and psychology is still all introductory stuff about the brain's anatomy and how it works.

The main piece of neural equipment that we need to succeed in trading, or in fact anything, is the orbito-frontal cortex, which helps us use our emotions to make good decisions. It only needs 5 seconds or so for the OFC to recognise a catalytic feeling from the subconscious - for example, a quarterback in the pocket in a game of gridiron needs only to scan his players, or a TV director who is casting a new actor and watches dozens of actors recite a few lines. The information is received subliminally and the unconscious brain processes it all with reference to inaccessible data that experience has built up into an emotional database of preferences and like and dislikes.



All this tells us that which we can apply is that to be good, we need experience, and that is accumulated on a subconscious level as well as all the conscious experiences we gain.



This is from "How We Decide" by Jonah Lehrer. There's enormous amounts of detail. I'm also reading another called "Willpower" by Baumeister, and have another one lined up called "The Chimp Paradox" which is something about taming the inner chimp which we all have.




Keeping an eye on your blood sugar levels and trying to restrict your trading sessions to times when your willpower is not depleted by earlier exercise is the first approach.

The second you are already taking: distracting yourself by occupying your time with other stuff.

When I come across more, I'll let you know.

Hope you get somewhere with the money management. One thing to remember about standard deviation is that it is based on distributions and mostly unless you decide otherwise, on 'normal' distributions. However the returns from trading systems as well as the movement of markets themselves do not fall into a normal distribution, so relying heavily on standard deviation calculations leads to severe errors. This is something that has bankrupted many organisations - e.g. Long Term Capital Management.
 
You must log in or register to reply here.

Similar threads

Yamato
  • Locked
my journal 2
169 170 171
Replies
3K
Views
476K
timsk
timsk
J
  • Poll
Trading journals: general discussion
Replies
5
Views
2K
athvia
A
N
Ninja's Journal
Replies
5
Views
2K
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