my journal 2

[Yamato]

resuming with creation of new trading systems

I am resuming from here:
http://www.trade2win.com/boards/trading-journals/85510-my-journal-2-a-66.html#post1100610

Ok, so I am done. No ZN and ZC for now, because "look elsewhere" was wrong, and I'd have to spend 1300 dollars on data, which mentally I cannot spend until I've got a million dollars, so it will be a while. So I will now build another... 3 to 6 systems per symbol, so that will give me about 20 more systems.

This step of gathering data is done. Next weekend I will proceed to the next step: backtesting the existing systems on these new 5 symbols. I'll start with the easiest part: the 3 new forex futures.

By the way, today I was saying I lose about 1 day out of 3, and today is indeed a losing day. Now that I know my stats, I can take it a little better. I'll be prepared from now on to have one losing day every three.

Now I will set up the data for tradestation. First I will rename all the files and place them in the right folder, ready to be set up on tradestation, and then I'll set one of them up. Then I'll be done for the day.

Due to not finding ZC and ZS, the symbols I'll be working on will be these:
CAD
AUD
CHF
NG
SB

Interestingly, SB is not SoyBean, but it's sugar, which is why I used to get confused.

Ok, now I'll rename the files and place them in the proper folder.

Just one more thing, let's double check the exchange's code for my two most ****ed up futures: sugar and natural gas.
https://www.theice.com/productguide/ProductDetails.shtml?specId=23
http://www.cmegroup.com/trading/energy/natural-gas/natural-gas_contract_specifications.html

Definitely SB and NG. So now I'll do the moving from folders and renaming as i said above.

TIME-STAMPING
Yet another thing to do. Double-checking the candles and the time-stamping of data. Here's info about this:

03/26/2010,1415,3.998,3.998,3.966,3.976,564,714
03/26/2010,1430,3.975,3.978,3.923,3.932,1241,1579

These above are my typical 15-minute candles. What is important to stress out here is that, unlike for ANFutures, disktrading.com stamps data just like IB with the ending time of a candle (ANFutures is the beginning time of the candle). So if it says "1415" it means that candle ended at 1415. I just checked IB's data comparing it with disktrading.com data and it matches completely.

An even earlier thing that I've checked is this: I made sure that again all time-stamps are in New York time as is the norm for disktrading.com, a very important detail for me (my systems' entries and exits are time-based). And sure enough here's what disktrading's email says:

Intraday ASCII data is time-stamped in New York time, daily data is in Exchange Time.

But now I will have to make sure of this for all 5 futures I am analyzing.

These disktrading.com guys are professionals. The data is cheap but they explain things very clearly and do their best. And are friendly. Nothing like what our "look elsewhere" user said. If I ever make any large amounts of money, I'll owe them a lot, and probably will make a donation to them.

Ok, so... the toughest *******, the NYMEX's NG is all taken care of: the data matches. Let's work on the other tough guy, Sugar on NYBOT.

Ok, sugar matches as well, perfectly. But there's something to note: its value is multiplied by 100. Instead of 0.1789 it reads 17.89. Natural Gas instead was identical in everything.

Now, an easier part, because I am more used to CME currency futures, but harder because these futures are time-stamped in New York time, yet are traded in Chicago time. So I'll have to be a little more careful.

First, the AUD. AUD is special, because initially they sent me the forex data, by mistake. Then they sent me the futures data, but I had kept the forex data, which is unethical but what the hell, I even told them. I can't just delete good data like that. I even offered to pay to keep the data.

Anyway, the CAD and CHF have the wrong forex data because it's reversed compared to the future, so I can't back-test my system on it. But the AUD is the same (quotes of 0.9 and not 1.1), so I kept the forex data, because the file is twice as big and most likely has more hours, fewer data holes, no jumping from contract to contract... so I kept the forex.

Now though, this may cause extra problems (I always have the future, just in case). So I'll have to be extra careful both for the timing of candles, and the time-stamping. Also: candles will not match because the future may be worth 100 ticks more (or less?) than the forex.

WRONG... I checked everything thoroughly and I can't get the forex to match pretty much anything, whereas the future's data matches perfectly everything on IB TWS, so I reversed my opinion, and I went back to the future.

Next I'll worry about the CAD.

As I do this stuff, I realize that it requires first of a lot of precision, attention and, as a consequence it's very tiring. You want to be a control freak when you're doing these things. If your system goes one hour fast it won't work. This work requires more attention and precision than intelligence, but it takes intelligence to realize that it requires precision. In fact stupid people are not precise because they don't know the difference between one thing and the other. But I am wrong. There's intelligent people who are clumsy, messy and all. I just wanted to tell myself that I am intelligent, whereas I know that I am very precise, rather than intelligent. All my spelling mistakes of the last few days are due to xanax, mostly.

CAD is fine, too.

Now just one more and then I can rest: CHF.

CHF... wow. I can't believe it. What the hell is this "look elsewhere" buddy of mine talking about? The data is perfect. I just checked the CHF, too, and it totally matches IB's data, like all other futures. And I paid like less than 80 dollars for all 5 symbols (they gave me a discount). Now imagine had I gone to one of the others I mentioned on this thread. By now I'd have spent about 10k on just the data of the futures I am trading. Let alone all the other data they gave me which I am not using. Long live disktrading.com.

I am done for today. Probably I won't write any more.

Make sure you watch Idiocracy, by Mike Judge, because it's a masterpiece.

 

[Adamus]

Re: resuming with creation of new trading systems

Lucky I didn't bother with that money management book. Sounds like I would have been wasting my time.

Re the possibility of drawdowns happening all at the same time on your systems if you commit enough leverage to them that it would blow your account - another risk you (we) can choose to take or not, like the risk of some massive market move you can't get out of.

It might be possible to work out the probability. Easiest would be to work out the chance of all your systems having a losing trade at the same time, since you know how often they lose. But you're talking about drawdowns, not single trades.

You could work out the average time between new equity highs or new equity lows, whichever appeals to you more. I'm thinking out loud here, but would it then be possible to say that the probability of you having a new equity low on any random day is 1 / avg time between equity lows?

If that was true you could then multiply those probabilities together for all your trading systems and you would have the probability that you experience a drawdown on all of them simultaneously.

 

[Yamato]

...Easiest would be to work out the chance of all your systems having a losing trade at the same time, since you know how often they lose...

Tell me about it. Today I had 5 trades, all losing. Of course two were on the currencies so they both lost against the dollar. Two were on OIL and Dow Jones, and one was on Bonds. They all went in the opposite direction, as they are all largely correlated (directly or inversely).

Had this happened to me (five straight losses) as a discretionary trader by now I would have blown out my account probably, because I'd have made revenge trades, which are like trying to win at poker when you don't have a good set. Because when you're upset you just don't wait for the right timing, both in entries and exits. Luckily it wasn't me trading, but someone else, and I didn't even interfere at all. Maybe that's why I wrote so much, to keep myself busy.

Thanks for your advice. Your recipe for combined drawdown sounds simple, but not simple enough for me.

For now I'd rather just look at the past 9 months and see, with any given combination, what was the worst drawdown, directly on the chart.

Of course, if you asked me to add up my 40 systems' maximum drawdowns, I can give you the sum quite easily. It would amount to about 100k (2.5k average dd per system), or maybe a bit more, 140k. So, if they all incurred their worst drawdown together, I'd be wiped out pretty quickly, since I haven't got that kind of money.

Hey, I see it this way. I've brought my account from less than 10k to about 25k three times before, and I knew nothing about money management, I didn't know anything about return of each system, there was no forward testing data whatsoever, and I was taking all the signals that came through with at least half of the capital. It was really a restless period.

Somehow I made it happen three times, then I blew the winnings mostly on discretionary trading, and some in the same way I had made them, usually less than half.

This time it should go better, because I am allowing only the best systems to trade, and am letting them trade less. And when those things happened I had half as many systems as now.

But to be honest, I don't feel safe at all. I am not gambling, but I am not playing it safely either.

I would say there's a 25% chance that I'll blow out my account. And a 75% chance that I'll go up to 30k and then I won't increase the contracts so that I can incur any drawdown and not feel it.

I think if I now went down below 10k, which is a few thousands less than I have now, I would panic, because it is all money from a god damn loan I had to take after blowing out my previous accounts, live on this forum, even as late as two months ago.

As you can see now all that has changed and I am a different person. I can be like this even for as long as one year. I am not manic-depressive. It won't happen again if I am not in a desperate situation. The only reason it happened is that I had no money. The less money you have the more desperate you get, and try to make it, and then... if I went below 10k now, I would start panicking, and i'd resume discretionary trading almost for sure. But even earlier than that. As soon as I approached 12k, i'd start panicking and resume discretionary trading, hoping to recover from a loss. Whereas it would be more rational to even sit and watch as you let your account bleed to death... rather than kill it yourself. Also because it would probably recover - but with your help it will die for sure.

When i interfere with my systems, it is close to when i had a marijuana plant growing in my house, and I wanted it to grow so much that I almost felt like watering it twice as it needed. It grew fine, and I didn't get arrested either, but it was quite stupid of me. Anyway, I found out that a plant doesn't grow faster because you're looking at it or watering twice as much. So does your capital. You just have to let a patient gardener do just what is needed, and I am not a patient gardener but my system is.

As long as the roots are not severed, all is well and all will be well in the garden.

There will be growth in the spring.

In other words, as long as the systems have probability on their side, the growth will happen. I have to accept my (one-day) winters and not panic. Usually just one day of drawdown is... was enough to make me relapse into discretionary... gambling.

Which is actually the same thing that Larry Hite says on the bottom of my post, in my quote:

What makes this business so fabulous is that, while you may not know what will happen tomorrow, you can have a very good idea what will happen over the long run.

There will be growth over the long run. As long as the roots are not severed, which means as long as the markets don't change and the systems have roots connected to the real markets.

Or maybe it means something different. Or it means also something else. My roots are also - not just the systems working - but also my capital. If the capital dies, the plants' roots are severed. As long as there's capital, there's hope for growth, and of course as long as the systems work.

I must not kill capital, and growth will follow. I must not touch it, simply.

 

[Yamato]

the correlation of my systems offsets their frequent trading

I think I understood a little bit more, just as I was sleeping on it.

Yesterday I was noting how, if you increase the frequency of trading you decrease your negative days, given that you have an ege, which means a probability of winning higher than 50% (this is just for simplification purposes: I am aware that wins could be bigger than losses and you could win with less than 50% of wins).

But I was also noticing and noting that, whereas my systems had an edge of two winning trades out of 3, and traded 8 times per day, that edge didn't play out as much as in the rand() function, which returned just a 10% of negative days, with a 66% edge and 8 trades per day: see attachment below.

And I was wondering why. And I answered myself: yes, an increasing of the frequency of trading decreases the negative days, but if the losses are distributed randomly the results are better than if the losses stick together, as with systems that are correlated, like mine are, systems which are built on similar algorithms for correlated markets (directly correlated or inversely correlated, which still makes the systems lose together).

Today I realized why this happens. You see, if the losses are random, and you have a 66% edge but you trade once a year, what happens in 10 years? You'll tend to have 6 to 7 negative years and 3 to 4 profitable years. But if you trade 10 times a year, with random losses, all years will become profitable, because each year will have a 66% of wins, which will make it profitable. So you see the pattern that if I increase the frequency, even though the profit itself may not increase (there's more trades, but they could be smaller trades on smaller timeframes), the probability of profitable days increases.

Let's go even further. If we traded 10 times a year, we'd have no negative years as I said. If we traded ten times a month, always with the same exact edge of 66% of wins, we'd tend to have no negative months (as I said earlier it's actually 10%, rather than none). If we traded 8 times a day, which we do, with that edge and losses that are randomly distributed, then we would have no negative days, which actually means we only tend to lose 10% of the time.

But then, since in my system losses stick together, like today and yesterday when I've just had my seventh straight loss, here's what happens: the frequency of trades gets offset by their correlation, and what really looks like 8 trades is actually just one big trade, with capital allocated among several correlated systems/markets.

That explains why I have 8 trades per day and an edge of 66%, but I still get 33% of losing days, whereas the rand() function gets 10%.

What this means basically... what it can be translated into is this: my systems seem to trade 8 times a day but, given their correlation, it is as if they really traded 8 times per week. And therefore their edge only ensures - like for the rand() function - that they lose every 10% of those 8 trades sets. So, that explains why that 10% of losing sets I don't experience in my daily trading is actually seen in my weekly trading, because I do have more or less a 10% of losing weeks, about 5 per year.

With a formula, it could have all been made easier, but I had to grasp it a piece at a time and kind of reinvent the wheel.

The formula - if it exists - would probably have it that for every set of 8 trades with a 66% probability of winning, you get a 10% of losing sets.

Obviously, the larger that set of trades gets, the smaller that percentage of losing sets will be.

I just tried it, just to make sure I was right about what I said. If you increase the set of trades from 8 to 14, and require a number > 7 losses to have a losing set/day, you only have a 6% of losing sets/days. If you decrease it from 8 to 6, and require losses > 3 to have a losing set, then you have 12% of losing sets.

I am sure there's a formula about this, from some swiss guy maybe, who lived in the 1700s.

Anyway, this is telling me something very clearly:

1) I must increase my systems
2) I must either make my systems or their markets non-correlated

In other words "random" is good. My systems must get as close as possible to having randomly distributed losses, given that edge of 66%.

The best and fastest way to do this is, rather than devising new systems, since I have already found systems that have an edge, to look for new, uncorrelated, markets. Like the future on soybeans market on Mars. Similar stuff. The further, the better.

I really have to explore the cmegroup web site, and maybe even develop new trading systems on new markets (even better) as needed.

Of course, to do this, to trade 100 systems or more, on 60 markets like Larry Hite did, I would need a different platform than excel, so this all will take a long long time.

I wouldn't want to do too much and do things with negligence and inefficiency, so for now I will stick to my 14 markets and to what excel can handle. In the future we will see.

For now let's remember from all I wrote that, for every 8 sets of trades I have, they're so correlated that they count as one randomly generated outcome. And therefore I cannot expect just a 10% of unprofitable days, but at best on a 10% of unprofitable weeks.

This is the way it works, very approximately.

I have about 8 trades per day, of which 4 intraday and 4 overnight.

These trades try to guess the direction of the market, but the futures I trade make it so that all 4 guesses tend to be wrong or right together because these markets I trade are all either directly or inversely correlated:

1) DIRECTLY CORRELATED MARKETS:
EUR, GBP, JPY, YM, ES, GC, CL

2) INVERSELY CORRELATED MARKETS (WITH REGARDS TO THE ABOVE MARKETS):
GBL, ZN

So basically my systems say something like:
EUR will go up
GBP will go up
ES will go up
ZN will go down...

...or any different combination of systems and markets, and if they're wrong they tend to be wrong together, despite the fact that signals are generated at different times and also based on different parameters and indicators.

Then at night, right around the New York closing time (16.00 EST), there's another bet by another (roughly) 4 systems/futures, and, once again, if they go wrong, they tend to go wrong together.

So basically it's almost (not exactly, but almost) as if there were 10 trades per week, rather than 40 trades per week. And, with a 66% edge, I can go wrong about 3.5 trades, which corresponds to almost two days per week (1.75).

Then of course things are quite different, because I simplified a lot, and wins are bigger than losses and my edge is smaller than 66%.

Another difference, against me, is that those bets don't only go wrong 4 at a time, but also several days at a time. So that I could be wrong about one entire week at a time.

It's hard... to be continued in later posts and sleepless nights.

 

[Yamato]

next week only authorize trading by non-reversing systems

I have a feeling this week will be a losing week, not just for the few systems I've used but even more for the whole 40 systems, which are the ones deciding the past week filter goes one way or ther other.

If the week is unprofitable for the 40 systems, about 30 of the 40 systems will reverse their signals next week.

On the other hand, since there are some false signals by this filter, and since I feel the EUR will not go down, but sideways at worst, this might be a false signal.

So I don't feel like allowing my systems to reverse their signals with real money. I will therefore only allow trading by the systems which don't reverse (ever) their signals, which to me are the healthiest and safest ones.

This is because I know the reversing signals produces a lot of profits and avoids a lot of losses, but I am not still sure why exactly that is.

The EUR forex has just crossed above its 40-day average, and this to me used to be the other filter and it would have told me to keep systems running as they are, without reversing signals. So, I'll do something in between and only allow the non-reversing systems to trade, but definitely I will not reverse the other systems' signals.

 

[Yamato]

awful automated week

It really tested my nerves.

I've had about 16 straight losing trades this week. This is so bad!!!

It's telling me that my systems were extremely unlucky but also that they are extremely correlated.

All in all, I've lost about 5000 dollars.

I didn't resist it and I started doing discretionary trades again, which - luckily - showed me a profit of about 1000 dollars.

I feel like Johnny Fontaine talking to Don Corleone right now:

 

[Adamus]

Re: awful automated week

I didn't resist it and I started doing discretionary trades again, which - luckily - showed me a profit of about 1000 dollars.

Please don't think that I'm encouraging you to do any discretionary trading, but what trades did you do? Are you still running them now?

I suppose I have to ask - you know I've got a tight deadline, my last fall-back option if I dont get any tradeable results from my intraday data is to start discretionary trading. I don't know what the learning curve is like though - possibly quite long.

 

[Yamato]

Nope. No discretionary trades are open any more.

In the meanwhile I did even more trades. Today one on the CL, which helped me recoup even more from my awful week (from automated trading -5000 with about 15 straight losses). Overall, with these discretionary trades, I saved my account almost 2000 dollars. Which mean everything to me at this moment, because my capital came almost entirely from a loan. If I go to zero this time, I can't trade anymore for the next few years.

The others were made yesterday, on the EUR, GBP, ZN, GBL... I can't remember exactly what time, at about 5 PM (Central European Time). I went LONG on EUR, SHORT on GBP, and LONG on ZN and GBL, all at the same time. Made money on all of them except the GBP.

It's not hard to make money for me. It happens on a majority of trades, as long as I am not addicted, which happens to me if I trade for a week on a regular basis. But the problem is that - even when I am not addicted I am not profitable in the long run, because the one time out of ten that I am not just wrong but very wrong, I just keep it open until it blows out my account. I am not a profitable discretionary trader. I've never been one. I might be profitable if I just made one (intraday) trade per month, but that's not a normal behaviour for anyone.

This week I was desperated, I broke my rules, and I made money only because I got lucky, and if I keep on going, that one very bad trade will happen and it will blow out my account, for the fiftieth time in 13 years.

So yes, the learning curve is quite long, to infinity in my case.

 

[Adamus]

So yes, the learning curve is quite long, to infinity in my case.

"Live as if you were to die tomorrow. Learn as if you were to live forever."
-Gandhi

 

[Yamato]

Which one are you applying? Not that I agree very much (the sentences contradict one another), but to judge by the title you chose for your journal, "Deadline June", you're applying "trade as if you were to die tomorrow" (no need to make money in that case). Which you confirmed by saying that if your system is not ready by June, you will start with discretionary trading. Not that my past behaviour makes a lot of sense either.

 

[Adamus]

To a certain extent you are right. Since learning is a part of living, you cannot do them in different ways.

He should have said, "Live as if you were going to die tomorrow, except when it comes to learning, in which case, behave as if you were going to live for ever."

That, I suspect, would never have made it into the quotation books.

It is also illogical, coming to think about it, since someone who is immortal would want to learn everything little thing they could, since in an eternal life you are bound to need the information at some point.

I just felt it was an appropriate point for a quotation. So I will retract that one, and come up with another:

Bessie Braddock: Winston, you are drunk, and what's more, you are disgustingly drunk.
Churchill: Bessie, my dear, you are ugly, and what's more, you are disgustingly ugly. But tomorrow I shall be sober and you will still be disgustingly ugly.

Must be Friday night.

I think I liked the similarity of an infinite learning curve and learning for ever, juxtaposed with living for the moment, or in your case perhaps trading for the moment.

 

[Yamato]

some doubts about reversing systems...

This week was highly unprofitable, so, according to my systems, I should now reverse half of them, but I am not sure about it... now that I am investing real money on this rule.

You see, i feel like next week won't go down, and that this week was an exception.

Luckily, there's almost half of them that work better without reversing, so those will just keep running as usual. Too bad they only account for less than 20% of profits. However, what's good is that they make money in both negative and positive weeks just the same, so they can balance the others, who at best will be one week behind.

The bad news is that the other half of systems actually makes more money in the reversed weeks, and with fewer trades.

The reason I don't feel comfortable with reversing them is that I feel this past week was a negative one just because the EUR went up too much on Monday, and opened with a gap up, using up all fuel on its first few hours of the week.

Now that it closed its gap up, i feel it will resume its rise.

This is what I will do most likely. I will trade the non-reversing systems. I will stay flat on the reversible ones as long as the EUR is below its 40-day moving average. Then, if it rises above it, I will consider trade everything as usual, pretending that the previous week was an exception, as I am guessing.

This is not an automated system right now, because my capital is so slow that if I had another drawdown like last week, I would be in serious trouble.

I will go even further in my discretionary approach: from now on, if the EUR doesn't look like it's in good health, I will keep all the reversible systems flat, so to avoid another week like last week. Especially if I witness another gap up like Monday's.

Here's a picture of it:
http://futuresource.quote.com/charts/charts.jsp?s=QEC 1!&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=

 

[Adamus]

Re: some doubts about reversing systems...

You can examine that idea quickly. Look at the start of the chart you gave - there's a gap down. Did that effect everything?

These gaps over the weekend are the first thing that I saw when I started looking at the tick data. In comparison to gaps when you're trading EOD, the gaps on an hourly chart look much meaner. Do you day trade, or do you hold positions over the weekends?

 

[Yamato]

I do both. Half of my systems day trade and the other half hold positions overnight (and therefore also "over-weekend", once a week).

 

[Yamato]

If it rises above the average, I will allow regular trading by reversible systems. If it stays below, I will keep them flat, but not reverse them as I should.
http://futuresource.quote.com/chart...a=D&z=800x550&d=LOW&b=CANDLE&st=MA(40,40,40);

 

[Yamato]

internet connections and geography

I've been studying internet connectivity on this web site:
http://speedtest.net/

The worst place I can reach and be reached from is this place in China (about 5000 miles away from me):


The best place of course is Rome, so I checked other interesting places, like California, expecting it to better even though farther than China (6k miles vs 5k miles):


But I would bet that even Alaska is better than China:


This "even" I just said was wrong, because it turns out Alaska is actually closer to me than California, but I was mislead by the map, which can't be tridimensional. Here's a better tool:
http://www.freemaptools.com/measure-distance.htm





Also, this link is quite good:
http://earth.watype.net/en/globe/




Anyway, connectivity has to do more with technology than with distance. Let's even check out New Zealand, the farthest place from me, more than twice as far as China, and going through oceans and stuff (not just land):


This is really amazing stuff. Anyway, New Zealand is more than twice as far, and yet more than twice as fast.

Useful links:
http://en.wikipedia.org/wiki/Bit
http://earth.watype.net/en/globe/
http://en.wikipedia.org/wiki/Svalbard
http://speedtest.net/global.php#0 (awesome link)

http://en.wikipedia.org/wiki/Eratosthenes
http://www.pinholephotography.org/SSGB Page.htm
http://www.pinholephotography.org/Solargraph instructions.htm

Ok, I am quitting this geographical quest, but one last thing. I just found out how to calculate the farthest place from me. You reverse the latitude and you reverse the longitude. Say you are at 42 latitude, you plug in -42. If your longitude is 12, you plug in -168. In google maps I mean, like this:
http://maps.google.com/maps?q=42,12
http://maps.google.com/maps?q=-42,-168

 

[Yamato]

museum of american finance

Now I am checking this out. It's quite interesting stuff:
http://www.moaf.org/

http://en.wikipedia.org/wiki/Buttonwood_Agreement
http://en.wikipedia.org/wiki/Tontine_Coffee_House

http://www.youtube.com/user/FinanceMuseum

A great collection of links from moaf.org site:
http://www.moaf.org/education/links/index

One of the links is The Stock Market Game:
http://www.smgww.org

Interesting stuff, very funny even. Here the students use fundamental analysis to trading:
http://www.pearsonfoundation.org/SMG/media/videoSMG_Hanser.html
http://www.pearsonfoundation.org/SMG/media/videoSMG_Friedman.html

Here's more on the Stock Market Game:
http://www.sifma.org/education/foundation-programs.html
http://www.sifma.org/education/smg.html

The Stock Market Game's carefully integrated curriculum lets students experience the financial markets in the classroom as they learn math, economics, social studies and the importance of saving and investing for the long term. Since its inception in 1977, nearly 12 million students in elementary, middle and high school have taken part in this real-world learning experience that enables teams of 3-5 students research, buy, sell, and trade a $100,000 portfolio of stocks, bonds and mutual funds at real-time prices, practicing teamwork, communication and leadership in the process.

These people are right. This is the way education should always be but is not. I started being interested in math only when I started trading, and the same applies to programming: I even taught myself VBA and excel in order to trade. In school, the way they presented this knowledge to me was very unappealing. The stock market game is a way to make you understand, ahead of time, that math could be useful and it's not just about doing your homework without any practical uses.

Here, in the Sue Friedman video this is much clearer than in the other video:
http://www.pearsonfoundation.org/SMG/media/videoSMG_Friedman.html

At the end of the video, one 5th grade student sums it all up:

You're learning without even knowing you're learning, and in the end, when the game is finished, you go like... "wow, I wanna play it again!".

 

[Yamato]

Cinema History Youtube Channel

I was looking for some stuff on Chicago, and here's what I stumbled across:
http://www.youtube.com/user/CinemaHistory

Some great videos, including one by Edison on Chicago in 1897:

 

[dog4]

Re: some doubts about reversing systems...

This week was highly unprofitable, so, according to my systems, I should now reverse half of them, but I am not sure about it... now that I am investing real money on this rule.

You see, i feel like next week won't go down, and that this week was an exception.

Luckily, there's almost half of them that work better without reversing, so those will just keep running as usual. Too bad they only account for less than 20% of profits. However, what's good is that they make money in both negative and positive weeks just the same, so they can balance the others, who at best will be one week behind.

The bad news is that the other half of systems actually makes more money in the reversed weeks, and with fewer trades.

The reason I don't feel comfortable with reversing them is that I feel this past week was a negative one just because the EUR went up too much on Monday, and opened with a gap up, using up all fuel on its first few hours of the week.

Now that it closed its gap up, i feel it will resume its rise.

This is what I will do most likely. I will trade the non-reversing systems. I will stay flat on the reversible ones as long as the EUR is below its 40-day moving average. Then, if it rises above it, I will consider trade everything as usual, pretending that the previous week was an exception, as I am guessing.

This is not an automated system right now, because my capital is so slow that if I had another drawdown like last week, I would be in serious trouble.

I will go even further in my discretionary approach: from now on, if the EUR doesn't look like it's in good health, I will keep all the reversible systems flat, so to avoid another week like last week. Especially if I witness another gap up like Monday's.

Here's a picture of it:
http://futuresource.quote.com/charts/charts.jsp?s=QEC 1!&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=

View attachment 81164

I think you must trade what you backtest, otherwise you are trading discretionary. If your backtests show a big improvement with reversing then that is what you must do. I don't think you should give them a chance, though I know it is painful.

If it makes you more comfortable, I subscribe to a market newsletter written by a successful futures trader (for >20 years)... he says EUR needs to get over 1.37 to be a buy. Before then the trend is down

 

[Yamato]

Thanks for the advice.