my journal 2

[Adamus]

Regarding those guys in Market Wizards, they say that, but I don't see Ed Seykota, Larry Hite, Richard Dennis, or whoever said it, providing the code or even just the principles of their systems on their web sites... what do you think?

Google for a PDF called Way of the Turtle by Curtis Faith - one of Richard Dennis's original turtles. It is the publication of the precise rules they used.

I too wouldn't worry too much about security. Set up the basics recommended for your operating system, avoid all executables from incoming spam or bit torrent downloads of pirated software etc, and you should be safe. The hackers can only get into your PC via breaking through from the outside, e.g. dictionary attacks trying to crack your password, or from the inside using executables smuggled in or 'exploits' in the OS you are using.

re: weighting or capital allocation: you are taking the same approach as Ralph Vince in Mathematics of Money Management. He always uses his system's worst loss as the basis for his optimal 'f', but in reality, your largest loss is always going to be in the future (unless you plan to quit tomorrow). I think you should look hard at the way your system trades in a particular market, go over the historical data and look to see what might have happened in a worst case scenario. i.e. if you are long the SP and an airliner flies into the WTC - if that was your market and it caught you in a trade, what could have happened?

 

[Yamato]

Oh, I do not know how much exactly I will lose in case of a terrorist attack, because I don't have a maximum loss but a time exit. To me it seems best that way for backtesting. I am afraid of overoptimizing stoplosses.

Thanks for the info about hackers (I still don't feel secure though). Thanks for letting me know about Curtis Faith and that I am on the right track as far as money management. If this Vince is doing it, it must be right.

Regarding the worst case scenario, I don't have any problems. If it happens, and I have as much capital as I am hoping to have, I can take any 5% of capital loss (according to my maximum loss rule, which is based on the previous biggest forward-tested loss), which as you pointed out is likely to be bigger in the (long-term) future, so I should be prepared to take a 10% loss, and I'm ok with that.

I'd say that I have it all figured out. The hardest part now will be to wait for weeks and months without doing anything at all. This forum will help. And so will the company of other automated traders like you and dog4.

For example, today there's a trade by a GBP system and nothing else. It will be quite tiring to wait. Also, tomorrow I have it down as a holiday, so my systems will not trade at all. Quite a boring life. I am living a more boring life than my computer, because he's the one doing all the trading. I might watch some movies again, like 10 to 20 movies, in the next few days.

 

[dog4]

Oh, I do not know how much exactly I will lose in case of a terrorist attack, because I don't have a maximum loss but a time exit. To me it seems best that way for backtesting. I am afraid of overoptimizing stoplosses.

Thanks for the info about hackers (I still don't feel secure though). Thanks for letting me know about Curtis Faith and that I am on the right track as far as money management. If this Vince is doing it, it must be right.

Regarding the worst case scenario, I don't have any problems. If it happens, and I have as much capital as I am hoping to have, I can take any 5% of capital loss (according to my maximum loss rule, which is based on the previous biggest forward-tested loss), which as you pointed out is likely to be bigger in the (long-term) future, so I should be prepared to take a 10% loss, and I'm ok with that.

I'd say that I have it all figured out. The hardest part now will be to wait for weeks and months without doing anything at all. This forum will help. And so will the company of other automated traders like you and dog4.

For example, today there's a trade by a GBP system and nothing else. It will be quite tiring to wait. Also, tomorrow I have it down as a holiday, so my systems will not trade at all. Quite a boring life. I am living a more boring life than my computer, because he's the one doing all the trading. I might watch some movies again, like 10 to 20 movies, in the next few days.

Surely not tiring! Your GBP was long right?

Another interview I read somewhere (not Mkt Wizds) was about a guy who lost loads in discretionary trading. He put his efforts into system development and going to the gym. Great results all round. I try (doing ok other than the gym!) to do the same. That way you keep busy productively rather than stare and stress. But sure, the first days or perhaps a couple of weeks, you can't help check it out all the time. Once you get more comfort, its easier to take your eyes off it.

You know all this... I think you are going through the withdrawal phase of losing, discretionary trading :cheesy:

 

[Yamato]

good formula... =IF(RAND()>0.44,1.5,-1)

Yeah, we will see if I am ready to quit discretionary trading or if, after losing for 12 years, I still haven't gotten my ass kicked long enough.

Yeah, the GBP was long, and I am glad but at this point I don't really feel like congratulating myself or the system... maybe because the trade was triggered by an idea that goes back months and months, so it's not like anyone was "brilliant" in picking this trade. But that's what is good about probabilities. I lose money by looking for brilliant trades, and the system makes money by making routine trades, based on my back-testing which simply found out what tends to happen in the markets.

Yeah, I'll manage to stop checking my systems, if my losses are not too big, due to a good money management and ultimately to luck, too, because even Ed Seykota could still lose all his money despite his big capital and his good money management. Indeed, no one can be positive that, despite all precautions, he won't lose every single trade of the next 1000 trades. It's close to impossible, but it's not impossible.

I kept on running my random tests, and got to develop a perfect pivot chart, which can be refreshed by just pushing the pivot exclamation mark, on the chart itself.



The chart shows more or less my systems, that have about 8 trades per day, with a 56% of wins (realistically), and av.win/loss ratio of 1.5 to 1 (makes 1.5, loses 1). Here's the formula I used:

=IF(RAND()>0.44,1.5,-1)

Here's the file:
View attachment system_with_56_percent_of_wins_and_8_daily_trades.xls

The differences from real trading are mainly two and they make my simulation look much better than my real trading. With these two different characteristics below one could make money and be content even with just a 51% edge:

1) The % of wins and the reward/risk ratio (ratio of av.win to av.loss) I used in the simulation are averages of all my systems, but my systems in reality vary among one another, as far as both characteristics (% of wins, and reward/risk). For example, if the CL loses, I go down 2000 as I did on Monday, and it might take me two weeks to recover. If the CL loses big and 3 times in a row, then I blow out my account (which is why maybe I should have not started trading it).

2) Random outcomes of trades as provided by the rand() function are actually better than the regular drawdown (and thus irrealistic), which tends to happen simultaneously on several systems. If my real trades outcomes were distributed so randomly it would be great, because it would mean my systems are totally uncorrelated, and also, very importantly, that they will never stop working, which might not be the case with real systems.

 

[dog4]

Re: good formula... =IF(RAND()>0.44,1.5,-1)

2) Random outcomes of trades as provided by the rand() function are actually better than the regular drawdown (and thus irrealistic), which tends to happen simultaneously on several systems. If my real trades outcomes were distributed so randomly it would be great, because it would mean my systems are totally uncorrelated, and also, very importantly, that they will never stop working, which might not be the case with real systems.

Very true. I recently had a similar issue. We know about fat tails, but these events also cluster in time. You cannot rely on the rand function for this reason.

I had to stop trading and figure out what was going wrong.

 

[Yamato]

Yes, thanks for reading the whole post! I thought I was writing notes to myself... and I usually do, but I am glad that sometimes someone else reads them.

What can I say, I was happy with finally finding a pretty concise formula that did the whole job, but you're right: it cannot be used for various reasons and mainly that randomly generated outcomes are too good to be true. And also that thing about the size of each trade is important. Unfortunately with my systems and my futures - given the small capital I have - I cannot weigh my investments the way I should. The CL has very good systems, but with its losses, if I get unlucky, it could blow out my account.

 

[Yamato]

one concept has become clear... it may be quite simple but... now that it's done

One concept, and major advantage of system (automated) trading vs discretionary trading, that has recently become very clear to me, after 12 years of discretionary trading and 5 years of automated trading (with studying trading systems since 2002), is the following:

system trading does not get affected by previous losses, no matter how often it trades

discretionary trading does, no matter how rarely it trades


I will not get tired of repeating this on my journal because of how important this concept is for everyone and particularly for me. I cannot become profitable, because - besides not having a strategy - after a loss AND after a win, my judgment is altered adversely. For example, after a loss I might become vengeful, and typically I will double up on a losing trade (instead of closing it) and bet my whole account on it, and that's a perfect recipe for blowing out my account, and sooner or later I will be in that state of mind. After a win, or after two wins, or after three wins, I will be excited and over confident, even if I wait a few days, and then I will incur a loss simply because I was too confident and didn't evaluate things properly. Essentially I could make a good trade once a month, if I really wanted to be objective. And even if that worked, my system in the meanwhile will have made 200 trades.

I have now gotten to a point where I just don't see how there can still be people, even on this forum (almost everyone actually), trying to make money with discretionary trading. I see them like people walking on a girder for a living. It's just not a good way to try to make a living.

This is how I felt when, in September, I started this journal (the previous thread, "my journal"), and wrote the details of my profile. Then I reverted back to discretionary trading for a few months, lost about 20k, and... sooner or later I had to open my eyes again. Discretionary trading is not for me, and it's not for most people. But they just don't realize that as well as I do now, because everyone, like I did, thinks they're better than average. And maybe we are better in many ways: more intelligent, more determined, but not in the way that makes you a profitable discretionary trader. Maybe one of the reasons I am not a good discretionary trader is the very reason that made me feel I would have become profitable: my determination and my drive. It didn't work the way I hoped it would. It translated into a persistence to lose and lose, without ever giving up despite the regularity of losses (never had a profitable month in 12 years of discretionary trading).

 

[Yamato]

Profit Factor: Errata and more thinking...

ERRATA CORRIGE

Some more tests and thinking about drawdown, with the help of excel and all the stats I have on my systems.

On average, they trade 8 to 9 times per day, and win 6 out of 10 times.

Now I'll also calculate how much they win vs. how much they lose, simply by a ratio of sum of wins vs sum of losses. Hold on... it wins 350k and loses 200k. Let me do some more math... damn do I suck... let's say that roughly it makes twice as much as it loses, also considering that I'd be using my best systems and not all systems. So, ok, this guy trades roughly 10 times a day, wins 3 out of 5 (or 6 out of 10), but those 3 it wins are twice as big as the 2 it loses. So that's about 3 times 2, which is 6 vs 2. So I guess I could say the system has a profit ratio of 3...

In this post (also quoted above):
http://www.trade2win.com/boards/trading-journals/85510-my-journal-2-a-60.html#post1093848

and maybe others, I kept referring to "profit ratio", but I really meant to say what tradestation reports call "profit factor".


FORMULAS

I will now try to implement such "profit factor" in my systems, so first of all, let's look at an old tradestation report and make sure we got the formula right.

It turns out it can be calculated in two different ways that return the same exact result. This is the way I thought was right:
http://fickletrader.blogspot.com/2005/10/profit-factor-walk-through-in-my.html

One of the primary statistics that I'm concerned with when looking at the results of a trading system backtest is the Profit Factor. The Profit Factor is computed as follows:

(PW * AW) / (PL * AL)

Where:
PW = Probability of a trade being a winning trade
AW = Average win size
PL = Probability of a trade being a losing trade
AL = Average loss size

But this way is easier and returns the same result, so why not do it this way from now on:
http://www.trade2win.com/boards/first-steps/23167-profit-factor.html

I calculate profit factor as the total of all winning trades divided by the total of all loosing trades...

Is generally gross profit/gross loss...

USEFULNESS

Let us think a little bit about what the profit factor tells us about a system, without reading other people's answers.

Now, let's take my GBL system. For my test period, on tradestation, it lists a profit factor of 2.19, which is the result of these two different (above-mentioned) formulas:

1) 62% times 0.17 points divided by 38% times 0.13 points, which is equal to 0.1054 divided by 0.0494, which is roughly equal to the number above, 2.19

2) easier formula: gross profit over gross loss, which 12.16/5.56 = 2.19

The second formula was so much easier!!!

Now let's look at what it measures. Say it didn't lose anything, ever. It returns an impossible result, because we can't divide by zero. Now let's say it lost 0.01 point, which is one tick on the GBL. We'd then get a profit ratio of 1216, which is everyone's dream. And it's telling us that such a system is only making money. A profit ratio of 1 means the system is not making any money, and a negative profit ratio means the system loses money.

So these were the extremes. But in between these numbers, 1216 and 1, there can be a lot of combinations. The first thing we have to make sure is that our systems have a large enough data sample, whether in forward testing or back-testing, because the profit ratio will be unreliable if we don't have dozens of trades. That's why a good filter would be overall profit, and if I had to choose just one information about a system, I would want to know its total profit (with 1 contract traded of course).


WHY I AM ALREADY DISCARDING THE PROFIT FACTOR: MY ROA INDEX IS BETTER

Now let's say that my system had a profit ratio of 12. That would not be enough, and I would want to know how many trades it made, and how big was his biggest loss. After all this reasoning, I am sensing that my measurement of ROA is inclusive of everything that the profit factor could tell me. So I might decide to not even figure out the profit factors for my forward-testing, because I'd rather keep things simple, and if something is not necessary, i will definitely not want it to distract me in any way.

My formula for ROA is inclusive of profit factor, in many ways, even though not by measuring the same exact things, because it is as follows (my invention):

=-G2/AVERAGE(H2,I2,2*K2)

This is just copied from one of the cells. G2 is the total profit of the system, and it gets divided by the average of these negative values (that's why I added a minus, so it turns out to be a positive value):

H2 is the forward drawdown: I need this because I want to know how badly my system came down in real-time testing

I2 is the back-tested drawdown: I decided I needed to use this, because some of my systems didn't lose much in forward-testing due to having made very few trades.

K2 is the forward-tested maximum loss: I doubled this because the value tends to be much lower than the drawdowns and this would cause it to count too little in the index, so I doubled it to equate its weight to the other two. I used the maximum loss because I want to know how much my system can make money compared to how deeply it can go down in just one trade.

The end result is an index that tells me exactly the quality and potential of my systems in terms of Return On Account, that is Gains over Losses (measured in terms of dd and max loss). It is not a good-looking formula but it tells me more than the simpler and better-looking profit factor formula.

But there is one thing that this profit factor which is definitely a close relative of my "ROA index" formula: I didn't incorporate in my formula the gross loss of my systems. Forget about the gross profit. But is it worth it to add this value and complicate things?


WOULD THE GROSS LOSS ADD ANYTHING TO MY "ROA INDEX" FORMULA?

How much is the Gross Loss really telling me that my drawdown and maximum loss didn't tell me? I don't think i will add it, because by using in my formula the Net Profit as a numerator, I am already counting the two ingredients from which it was built: gross profit minus gross loss. If the gross loss was huge, I'd have a lower net profit, so it is already in my formula.


NO, I WILL DISCARD IT THEN

Am I repeating in such a way any other ingredients of my formula? I don't think so, because max loss is not the same as max drawdown. You could have huge losses but very rare, and that would be almost ok. Or you could have small losses, but that last years, and that would not be ok at all. So I need to monitor both. Finally, I do need to use the back-tested drawdown, too, because the forward-tested data could be so little that the drawdown doesn't exist. But then one might say, why not use the back-tested max loss as well? First of all, I don't trust those values as much. Second of all, if the system has not incurred losses, it means it hasn't traded much, and if hasn't, then it hasn't made much profit so its ROA index value will be low because of that. Maybe there's still flaws in my reasoning, because I am not a math guy...

...but what matters to me is to keep things as simple as I can understand them completely or at least enough to use them profitably. So I'd rather keep a simple formula, which is within my grasp, than use the famous "sharpe ratio", which seems too complicated to me. Same applies to all other formulas you can find a tradestation report. I'd rather do little and do it very well, than do a lot poorly. I will only increase quantity when my quality in taking care of that quantity will be very high. Until then... I am done with formulas for now.

Different is the talk about systems. In that case, quantity is better than quality. Because the search for quality in a system encourages you to over-optimize it, and then you don't have great quality and little quality, but only zero quality and little quantity, so you might as well focus on quantity from the start, as Larry Hite says in Market Wizards:

There are many money managers who use trend-following systems—and quite a few of those don't second-guess their own systems. What makes Mint different? How have you been able to achieve return/risk ratios far above the industry average?

Because we know that we don't know. No matter what information you have, no matter what you are doing, you can be wrong. I have a friend who has amassed a fortune in excess of $100 million. He taught me two basic lessons. First...
[...]
The third thing we do to reduce risk is diversify. We diversify in two ways. First, we probably trade more markets worldwide than any other money manager. Second, we don't just use a single best system. To provide balance, we use lots of different systems ranging from short term to long term. Some of these systems may not be that good by themselves, but we really don't care; that is not what they are there for.
[...]

In summary, when you work your ass off as an automated trader, you're better off worrying about diversification (quantity of systems) than worrying about % wins, risk/reward and profit factor (quality of systems), because you may otherwise waste years over-optimizing, and may never be able to create the excellent system you wanted to create, whereas you could have used those years to build 100 pretty good systems (as long as they're profitable, they're all good), and build up your capital in the meanwhile.

 

[Yamato]

profit factor reconsidered and now enabled

I kept thinking about it, and did some tests, and found out that profit factor gave me quite different rankings for my systems than the ROA index. So I am now using, as money management rule, an average of profit factor and the previous ROA index, which I already explained in the previous post. I am quite satsfied.

 

[Yamato]

bringing my systems from 40 to 60

Adamus was talking about currency futures and he made me realize, together with Larry Hite's interview on Market Wizards, that I should expand more and bring my systems from 40 to 60, thanks to some more currency futures, which i will now analyze in terms of volume.

I will now list the futures I already trade and those I still don't trade, in terms of volume, as provided by futuresource.com, which is the same as the charts provided by CME, but which sometimes don't work. Or maybe you have to access CME charts, for them to work, from these pages (from home, the CME charts are faster than futuresource.com, but their links do not work):
http://www.cmegroup.com/trading/fx/g10/euro-fx-british-pound_quotes_globex.html


ALREADY TRADED BY MY SYSTEMS

EUR: 300k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

GBP: 100k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

JPY: 100k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);


NOT YET TRADED

AUD: 80k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

CHF: 70k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

CAD: 70k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

No wonder I wasn't trading them - I had discarded them due to lower volume. But still - they're good enough for diversification purposes, especially since my trades last many hours.

I will now build 10 to 20 more systems on these 3 guys (AUD, CHF, CAD).

By "now" I mean "in the next two months". So let's take it easy for now.

At the end of this final feat, I will have over 50 systems, based on 6 currencies, 2 commodities, 2 bonds, 2 stock indexes.

Let's study the population and GDP of these currency futures, and understand why for example the AUD has such big volumes (whereas the RUB has very low volumes)...

In order of volume (thousands):
EUR, 300
JPY, 100
GBP, 100
AUD, 80
CAD, 70
CHF, 70

GDP from here (2008 List by the World Bank, in millions):
http://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal)

EUR, 13,565,4793
JPY, 4,909,272
GBP, 2,645,593
AUD, 1,015,217
CAD, 1,400,091
CHF, 488,470

Wow, great explanation. I have to admit that the only discrepancy here is that the AUD has more volume than the CAD. Yet also Russia and Mexico have around the same GDP as Canada, whereas their currency futures don't have volume. I guess it has to do with the fact they don't trade much with the US? Maybe so.

Let's now look at the population, and later I want to find out why China doesn't have a currency future:

Population from here:
EUR, 328,597,348
JPY, 127,380,000
GBP, 62,041,708
AUD, 22,211,000
CAD, 34,056,000
CHF, 7,782,900

Here everything also looks quite normal. Once again, Australia seems to have a higher volume than its population would justify, and so does switzerland, but of course we know why.

Now, on with my quest to understand what the **** is the matter with China and its currency.

Speaking of population, this is really something:
http://www.census.gov/population/www/popclockus.html



Now, on with my search on the...
http://en.wikipedia.org/wiki/Chinese_yuan

While I was studying the Chinese Currency, I came across a link of major importance as far as currencies and volume:
http://en.wikipedia.org/wiki/Forex#Trading_characteristics

It lists the most traded currencies in this order:
 United States dollar
 Euro
 Japanese yen
 Pound sterling
 Swiss franc
 Australian dollar
 Canadian dollar
 Swedish krona
 Hong Kong dollar
 Norwegian krone
 New Zealand dollar
 Mexican peso
 Singapore dollar
 South Korean won

It's amazing, because - except for CHF being a bit higher - they're exactly in the same order I listed for CME currency futures. One more reason to expand my systems to cover those three extra currencies. Besides, there's a big gap in volume between the last one of them and the next one, the Swedish Krona. Everything is just perfect.

These extra symbols/systems will simply add profit, without increasing the drawdown as much as they increase profit, which is all that counts. Because, of course, if they didn't do that, and didn't do the job done by diversification, then it would simply make more sense to increase the number of contracts on the other systems. Let me reason more about this. If they were directly and completely correlated to the other systems (which is impossible), they would do no harm, but simply be useless (it would be the same as trading more contracts of the other systems). If they were inversely correlated, and won when the others lost, they would actually decrease the drawdown (right?). If they were something in between, somewhat correlated and somewhat on their own, they would still benefit me, by increasing profit more than they increase drawdown. So, once again, it makes sense to use them, as they will increase the Return On Account of the systems as a whole.

The AUD-CAD-CHF project , which will bring my systems to about 60, is a long term project, which I could worry about once I quit my job, and not immediately in the next few months.

Anyway, back to the Chinese currency and to why the hell it doesn't have a future.

Here it is listed with the other currencies:

At OANDA's Currency Converter, they have it, too:




Wow, man... OANDA even lets you trade it:
http://fxtrade.oanda.com/forex_trading/why_trade_with_oanda/spreads/all_spreads




So I guess I'll have to go to the CME web site to figure out why they don't have a future on it.

Ok, I found a .pdf on their web site about the "Chinese Renminbi Futures":
http://www.cmegroup.com/rulebook/CME/III/250/270/270.pdf

Chapter 270
Chinese Renminbi Futures
27000. SCOPE OF CHAPTER
This chapter is limited in application to futures trading in Chinese renminbi versus the U.S. dollar.
The procedures for trading, clearing, delivery, settlement and any other matters not specifically
contained herein shall be governed by the rules of the Exchange...

Ok, then, they talk about it as if it were a product, but why aren't they listing it as a product?

Here's a thread from elitetrader.com which might explain why. I'll quote the parts that say something useful:
CME Renminbi futures

06-20-06 01:50 PM

Is anyone planning on trading the new Renminbi futures on the CME?

Here's the email I got from the CME:

On Monday, June 19, 2006, CME announced today its plans to list new futures contracts and options on futures on the Chinese Renminbi against the U.S. dollar, Euro and Japanese yen. The new contracts, CME Chinese Renminbi futures contract, CME Chinese Renminbi/Euro futures contract and CME Chinese Renminbi/Japanese Yen futures contract, as well as options on these futures, will trade exclusively on the CME Globex® electronic trading platform and are currently scheduled to begin trading on Sunday, August 27, 2006, for the trade date of Monday, August 28.

CME foreign exchange (FX) markets reflect notional value of more than $60 billion a day, making CME the largest global FX marketplace outside the interbank market. These new Renminbi contracts will lead to increased market participation by current CME FX customers as well as introduce CME FX markets to new market participants.

For more information on CME Chinese Renminbi futures and options, please visit www.cme.com/renminbi.

Ok, what the hell... that link doesn't work, but I found the official page on the Renminbi that works:
http://www.cmegroup.com/trading/fx/emerging-market/chinese-renminbi.html

So, why are they hiding it from their main links? Damn it, what a moron I am! They're not hiding it, it's right here:



I was looking for the word CNY or similar, as for "Chinese Yuan", but they used RMB, which is so far from it!

So now I'll just have to look at volume and see if it's worth trading. But I am pretty confident that it will not be.

Wow, can you believe this? Who would have thought that the future on the currency of the biggest country in the world by population would have traded 2 contracts per day? And yet this is the case.



I know nothing about finance, but what this is telling me is something like the States telling to China: you're big but we're not playing with you because we don't trust you, and we want nothing to do with you.

 

[dog4]

Re: bringing my systems from 40 to 60

Adamus was talking about currency futures and he made me realize, together with Larry Hite's interview on Market Wizards, that I should expand more and bring my systems from 40 to 60, thanks to some more currency futures, which i will now analyze in terms of volume.

I will now list the futures I already trade and those I still don't trade, in terms of volume, as provided by futuresource.com, which is the same as the charts provided by CME, but which sometimes don't work. Or maybe you have to access CME charts, for them to work, from these pages (from home, the CME charts are faster than futuresource.com, but their links do not work):
http://www.cmegroup.com/trading/fx/g10/euro-fx-british-pound_quotes_globex.html


ALREADY TRADED BY MY SYSTEMS

EUR: 300k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

GBP: 100k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

JPY: 100k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);


NOT YET TRADED

AUD: 80k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

CHF: 70k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

CAD: 70k daily contracts
http://futuresource.quote.com/chart...&o=&a=D&z=800x550&d=LOW&b=CANDLE&st=VOI(1,1);

No wonder I wasn't trading them - I had discarded them due to lower volume. But still - they're good enough for diversification purposes, especially since my trades last many hours.

I will now build 10 to 20 more systems on these 3 guys (AUD, CHF, CAD).

By "now" I mean "in the next two months". So let's take it easy for now.

At the end of this final feat, I will have over 50 systems, based on 6 currencies, 2 commodities, 2 bonds, 2 stock indexes.

Let's study the population and GDP of these currency futures, and understand why for example the AUD has such big volumes (whereas the RUB has very low volumes)...

In order of volume (thousands):
EUR, 300
JPY, 100
GBP, 100
AUD, 80
CAD, 70
CHF, 70

GDP from here (2008 List by the World Bank, in millions):
http://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal)

EUR, 13,565,4793
JPY, 4,909,272
GBP, 2,645,593
AUD, 1,015,217
CAD, 1,400,091
CHF, 488,470

Wow, great explanation. I have to admit that the only discrepancy here is that the AUD has more volume than the CAD. Yet also Russia and Mexico have around the same GDP as Canada, whereas their currency futures don't have volume. I guess it has to do with the fact they don't trade much with the US? Maybe so.

Let's now look at the population, and later I want to find out why China doesn't have a currency future:

Population from here:
EUR, 328,597,348
JPY, 127,380,000
GBP, 62,041,708
AUD, 22,211,000
CAD, 34,056,000
CHF, 7,782,900

Here everything also looks quite normal. Once again, Australia seems to have a higher volume than its population would justify, and so does switzerland, but of course we know why.

Now, on with my quest to understand what the **** is the matter with China and its currency.

Speaking of population, this is really something:
http://www.census.gov/population/www/popclockus.html

View attachment 80194

Now, on with my search on the...
http://en.wikipedia.org/wiki/Chinese_yuan

While I was studying the Chinese Currency, I came across a link of major importance as far as currencies and volume:
http://en.wikipedia.org/wiki/Forex#Trading_characteristics

It lists the most traded currencies in this order:
 United States dollar
 Euro
 Japanese yen
 Pound sterling
 Swiss franc
 Australian dollar
 Canadian dollar
 Swedish krona
 Hong Kong dollar
 Norwegian krone
 New Zealand dollar
 Mexican peso
 Singapore dollar
 South Korean won

It's amazing, because - except for CHF being a bit higher - they're exactly in the same order I listed for CME currency futures. One more reason to expand my systems to cover those three extra currencies. Besides, there's a big gap in volume between the last one of them and the next one, the Swedish Krona. Everything is just perfect.

These extra symbols/systems will simply add profit, without increasing the drawdown as much as they increase profit, which is all that counts. Because, of course, if they didn't do that, and didn't do the job done by diversification, then it would simply make more sense to increase the number of contracts on the other systems. Let me reason more about this. If they were directly and completely correlated to the other systems (which is impossible), they would do no harm, but simply be useless (it would be the same as trading more contracts of the other systems). If they were inversely correlated, and won when the others lost, they would actually decrease the drawdown (right?). If they were something in between, somewhat correlated and somewhat on their own, they would still benefit me, by increasing profit more than they increase drawdown. So, once again, it makes sense to use them, as they will increase the Return On Account of the systems as a whole.

The AUD-CAD-CHF project , which will bring my systems to about 60, is a long term project, which I could worry about once I quit my job, and not immediately in the next few months.

Anyway, back to the Chinese currency and to why the hell it doesn't have a future.

Here it is listed with the other currencies:

At OANDA's Currency Converter, they have it, too:

View attachment 80196


Wow, man... OANDA even lets you trade it:
http://fxtrade.oanda.com/forex_trading/why_trade_with_oanda/spreads/all_spreads

View attachment 80198


So I guess I'll have to go to the CME web site to figure out why they don't have a future on it.

Ok, I found a .pdf on their web site about the "Chinese Renminbi Futures":
http://www.cmegroup.com/rulebook/CME/III/250/270/270.pdf



Ok, then, they talk about it as if it were a product, but why aren't they listing it as a product?

Here's a thread from elitetrader.com which might explain why. I'll quote the parts that say something useful:
CME Renminbi futures


Ok, what the hell... that link doesn't work, but I found the official page on the Renminbi that works:
http://www.cmegroup.com/trading/fx/emerging-market/chinese-renminbi.html

So, why are they hiding it from their main links? Damn it, what a moron I am! They're not hiding it, it's right here:

View attachment 80200

I was looking for the word CNY or similar, as for "Chinese Yuan", but they used RMB, which is so far from it!

So now I'll just have to look at volume and see if it's worth trading. But I am pretty confident that it will not be.

Wow, can you believe this? Who would have thought that the future on the currency of the biggest country in the world by population would have traded 2 contracts per day? And yet this is the case.

View attachment 80202

I know nothing about finance, but what this is telling me is something like the States telling to China: you're big but we're not playing with you because we don't trust you, and we want nothing to do with you.

Because its pegged to the USD. So people trade that instead, which is liquid.

One day all this will change. But people have been saying this for 5 years

 

[Yamato]

Let me check what "pegged" means:
http://en.wikipedia.org/wiki/Pegged

Oh, I see. Thanks for the info.

 

[Yamato]

AUD-CAD-CHF project part 2: getting the data

I am pretty satisfied that I narrowed it down to just 3 more futures. Let's celebrate with a song, before moving on to the next step.

The next step is finding out if I have the data to do my back-tests.

I couldn't really figure out if I had the data (too many symbols), but for 20 dollars per symbol, I just wrote an email to the usual data vendor, disktrading.com, and asked them if I can buy those three symbols. Besides, the previous data stopped at 2008, and also now they might have better data. Once I'll get done with buying the data, I will worry about preparing it for backtesting.

Unfortunately the "look elsewhere" expert didn't return to tell us where we could find better data at the same price, despite being asked about 10 times here. Let me quote him once more:

...many more vendors exist for intraday data with simliar pricing. I would just look elsewhere.

It's as if you asked someone for directions, and he replied "nope, you've been going in the wrong direction. You gotta go... in a different direction...".

 

[Yamato]

reasons for my failure at discretionary trading

How my personality adversely affects my discretionary trading
If anyone, after 3000 posts on t2w, should still wonder why travis has been unprofitable for 12 years (i get asked that a lot), I will stress out once again some facts that I see as evidence in my life. Just in case anyone (even just me) still cares to hear more explanations. I like to explain things as clearly as possible, even just to myself.


Not enough patience to stick to a manual strategy, even if entirely mechanical
In summary these facts (I will list some later) show I have a compulsive, addictive personality, and I believe this personality doesn't mix at all with discretionary trading. It doesn't have much to do with not having a profitable strategy. I can't even develop one, and when I have one, I can't implement it. Even on this journal, I had come up with several profitable discretionary strategies (e.g.: the "Sniper" EUR "Protrend Discretionary System"), which were different from my automated systems (because I could have copied one of those as well). And guess what: with one reason or another I always ended up abandoning them, mainly due to a lack of patience and an excess of compulsiveness.

In much the same way that I:
1) put the milk on the stove and often take it off the stove before it even warms up enough, because I just can't be patient enough to wait. Yes, it's unbelievable.

2) put the bread in the microwave and more than once I didn't wait long enough for it to unfreeze, because I couldn't stand to stand up there, waiting for the microwave to do its thing.

3) I scratch my nose compulsively, and until recently, even my scalp. When I was a teenager I used to bite my fingernails. And even earlier I used to suck on my thumb.

4) I've "befriended" every single user I found reading my journal and now I have 157 "friends".

5) I've written 2500 long posts all on my journal.

6)...

I don't need to list any more behaviours. Some of them are even on my self-control table and chart, which I posted several times. You get my point.

A constantly impatient behaviour which I rarely realize
You just can't place a person like me in front of a trading platform and expect them to behave and effectively control themselves in a wide variety of ways, necessary to consistently produce profitable trades. I can't even stay rational with the chart game. After a while I start trading compulsively, and take losses personally even at that. If you gave me an entirely mechanical method, I could probably trade it, but then again, it would require a lot of patience, and now that I think of it, I already had come up with such methods, and I just didn't follow them. If I have to stare at charts for 3 hours in a row, I will follow the method, and pretty soon, start placing other trades not related to the method. I will not tell you "I am breaking my rules", I will present it to myself as "look, I have come up with a second, better method". But the pattern will always be that of changing the rules and not following the mechanical method.

Never been good at it, and always kept trying = forever unprofitable
Once you have understood the above reasons for my failure at discretionary trading, it is also pretty easy to see why such a persistent researcher and hard worker like me, has been unprofitable until now. It's because I never stopped doing discretionary trading since 1997, and this has caused me consistently unprofitable months for 12 years. I can't succeed at it, because I can't change my personality, and I cannot become profitable until I quit trying to succeed at it. I think I quit it about a month ago, but I won't be sure about it until I'll have put years between my last discretionary trade and the present. In the past two years, it eroded all the profits I made in system trading.

This journal has damaged me at times, but overall it was a good thing
It's funny that on this very useful journal for me, there's been people who offered to help me in order to finally make me a profitable discretionary trader. And, needless to say, I don't feel it was a good thing, despite their good faith of course. I remember i even told them "please don't go there, because you'll cause me further disasters". I ended up trying one more time. When I came here I was sober, then I relapsed. Now I am sober again.

In the meanwhile, many people who were following this journal left it, maybe because they've become bored by it, or disappointed by my behaviour. But this journal has been useful. It witnesses to me and others what my weaknesses and my strengths are. What I do when I am affected by my compulsive gambling behaviours, what I say, how I feel, how i justify it. How dangerous i can be to my account. This journal has helped me, in that it has helped me document and observe and analyze my behaviour, and better identify the enemies within myself.

How my compulsive personality adversely/favorably affects my discretionary/automated trading
After all, the compulsiveness that hurts me so much while I do discretionary trading, that extra energy that causes me to trade while I shouldn't, is the same energy that makes me double-check macros, formulas, percentages, bugs, and clearly whereas in one field it hurts me by making me double-trade, in the other field, this double-checking benefits my systems, where the more I double-check, the more things will work smoothly.

I cannot change my personality head-on
To sum up what I've realized until now, I would say that, for better or for worse, I cannot change my personality head-on and it would be a waste of time to keep on trying. I can only trick and manipulate my personality. But I can't change it head-on, as I can't stop a truck by standing in front of it. I can shoot the driver, etc., you know what I mean.

I can't say "let's stand in front of the platform and learn to control myself and be patient". For a variety of reasons I am unable to achieve that. What I can do and should do, is to not stand in front of the screen and get used to not trading at all. That I can do and should do. It's just like my self-control updates. I've been able to cut back on a lot of compulsive negative behaviours, such as scratching my head and eating, but I can't eliminate all of them at once and permanently. But whereas it won't kill me if I don't do that, with discretionary trading, unless you eliminate all of them at once and permanently, your account will not grow. Therefore I cannot do discretionary trading, simple as that.

On a side note, I am better at controlling the negative behaviours, if that mother ****er and the goddamn maid don't fill up the refrigerator with sweets as they've been doing lately. Today I ate the last two remaining sweets and hopefully they won't buy any more, or I'll throw them away, flush them down the toilet. It's hard enough to control myself without people placing temptations in the house... that's why I could never have a gun with bullets, because I'd feel it has to be used and I'd compulsively start shooting people until i've used all the bullets.

 

[Yamato]

back to work: profit factors

Now let's get back to work and analyze what the newly implemented profit factors tell us about the systems.

I want to do two things. First see what the profit factor tells me about my systems. Second, see if there's any discrepancies between that and what the ROA index was telling me.

1) What the profit factor is telling me

First of all, i must say that a profit factor below 1.5 is not good enough, unless it's due to few trades having been made (since it's based on the forward-tested data).

So let's look at how many systems I've got below 1.5 and see if they're all bad.



I've shaded those systems in yellow. There is a total of 22 bad mother ****ers, quite a lot actually. A majority of my systems seems to suck in terms of profit factor, and such value rarely lies.

Let's see how many of them have a low value because they have under 20 trades. We can forgive these:

CL_ON_4 (12 trades and 1.4 profit factor)
All opening gap systems get forgiven because they trade rarely but rarely miss, I am positive about this.
GBP_ID_4 is good enough (I guess 1.5 would be acceptable)
...

...hmm, this analysis of profit factors has made me realize that I really do not have much of an edge, and I can't afford to discard profit factors of below 1.5, but should be happy to discard just those below 1.25. Because most of my edge stays with systems that have about 1.5 of profit factor. I am glad I didn't discard this measurement as I was going to do a couple of days ago.

Essentially, due to profit factor, I know that I should discard those 7 systems that have traded enough and that still have a pf lower than 1.25.

2) discrepancies between profit factor and ROA index?

Just as profit factor is not complete, the other index could also be lacking, and they indicate conflicting ratings for my systems. For example, if a system traded 2000 times, and each time either made 11 dollars or lost 10 dollars, such a system would suck, and have a profit factor of 1.1. On the other hand, in terms of ROA, this system would score very high, because its maximum loss and drawdown could be as low as 1 dollar. Furthermore, this system would make 1 dollar every two trades, and the 1000 profit would get divided by 1 dollar loss and drawdown, giving me a stellar ROA rating. I have such systems. And they're not worth trading, as they keep my capital busy for no profit. I was alerted as to that by seeing a very low average gain per trade, which in many ways is an indicator similar to the profit factor. Now let's see if profit factor demolishes any high-scoring ROA index systems.

Not really, I was even wrong about that ZN hypothesis (that scored high on one ratio and low on the other). I guess right now the PF is in line with the ROA. Actually some PF values are telling that some systems are better than what the ROA values would have had me believe.

From this little study, I'll retain that it was a good idea to compute an average index of ROA and PF. And that systems with PF lower than 1.25 are to be discarded.

 

[Yamato]

I thought I was done, but...

I thought i was done writing, but I've just realized i have nothing else to do, so I'll write more.

First of all about edge.

yes, I do have a small edge, those 22 systems with less than 1.5 show it quite clearly. But let's think about what this small edge can do. Let's say I get... hey I already knew i had a small edge when I knew i had an average 60% of wins.

That basically means... doing some approximate estimates... hey! 6 wins and 4 losses. That is exactly a 1.5 profit factor. What a moron.

That is my edge to begin with.

Anyway, if I make 10 trades per day, that gives me 2 net wins. And if those trades are about 200 dollars each, I'll have a gain of 400 per day, with a monthly gain of 8000 dollars, on a capital of about 20k. Let alone that some systems are better than risk/reward of 1. But wait... profit factor already incorporates both % wins and average win / average loss. Ok, now it all makes sense. A profit factor of 1.5 is not bad at all, but mine is higher (1.75), because, besides having a 60% of wins, I have bigger wins than losses. This is all quite good, provided that you have enough systems. Yet those morons on forums (especially those on elitetrader.com) keep saying you need better than that, thereby encouraging you to over-optimize just one single system, you're left with nothing in your hands... give up... and we know the rest. You should instead build something simple with 1.5 profit factor and then move on to building more systems.

A small edge is powerful if applied on a large number of systems.

Moving on to... profit factor by groups

There you have it. I just did my favorite thing: a pivot table. Here it is:

Average of Profit Factor
type of system - average
ON bounce 1.37
Opening Gap 0.41
overstretched 2.90
Volat.Breakout 1.56
WeekDay Bias 2.17
WITH ID trend 1.79
Grand Total 1.75

Ok, Opening Gap didn't trade pretty much at all, so let's forget about those guys. And let's even remove them from the picture, and see what i get.

Average of Profit Factor
type of system - average
ON bounce 1.37
overstretched 2.90
Volat.Breakout 1.56
WeekDay Bias 2.17
WITH ID trend 1.79
Grand Total 1.89

Ah ah, what do you know... those mother ****ers saying that top and bottom picking is bad, and yet the best systems are exactly the top and bottom picking systems, with a profit factor of about 3.

Second best systems, weekday bias, the simplest ones. They say that on given days of the week, a market will tend to go up/down more frequently. But these guys are dangerous because if you are wrong, they have potentially a big drawdown. So right now I am not trading them.

WITH ID trend trade very often and make the most money, because they are pretty good. So far everything is clear.

The volatility group is good as well. Nothing to say about this group.

The last group kind of sucks, the ON group, so let's look at them one by one.

system Total
CL_ON 1.19
CL_ON_2 1.76
ES_ON 1.36
EUR_ON 1.27
GBL_ON 1.07
GBP_ON 0.62
GC_ON 1.47
JPY_ON 1.45
YM_ON 2.12
ZN_ON 1.39
Grand Total 1.37

Now this is a group where some do really poorly and I don't even let them trade, but it is true that the edge is not really a good one. If I had to discard ideas, this would be the one I'd discard first. That is, if I had so many systems that I didn't know what to do with them, which is not the case.

So I'd discard... after discarding the 4 worst systems, I am left with a pretty reasonable group, but I have to hope that those 4 are plain bad systems and aren't going through a bad period, because otherwise I'd be screwed by discarding them and keeping others that will go through a bad period next.

system Total
CL_ON_2 1.76
ES_ON 1.36
GC_ON 1.47
JPY_ON 1.45
YM_ON 2.12
ZN_ON 1.39
Grand Total 1.59

On the other hand, let's now look at how often they all trade, to make sure that the bad ones aren't bad simply because they didn't trade long enough.

system - trades - profit factor
CL_ON 64 1.19
CL_ON_2 76 1.76
ES_ON 58 1.36
EUR_ON 22 1.27
GBL_ON 70 1.07
GBP_ON 12 0.62
GC_ON 51 1.47
JPY_ON 79 1.45
YM_ON 62 2.12
ZN_ON 132 1.39
Grand Total 1.37

Of the 4 systems I am getting rid of, the EUR and GBP have traded very little so they could still recover quite quickly. The CL and the GBL instead definitely suck because their sample is large enough.

 

[dog4]

Re: back to work: profit factors

Now let's get back to work and analyze what the newly implemented profit factors tell us about the systems.

I want to do two things. First see what the profit factor tells me about my systems. Second, see if there's any discrepancies between that and what the ROA index was telling me.

1) What the profit factor is telling me

First of all, i must say that a profit factor below 1.5 is not good enough, unless it's due to few trades having been made (since it's based on the forward-tested data).

So let's look at how many systems I've got below 1.5 and see if they're all bad.

View attachment 80214

I've shaded those systems in yellow. There is a total of 22 bad mother ****ers, quite a lot actually. A majority of my systems seems to suck in terms of profit factor, and such value rarely lies.

Let's see how many of them have a low value because they have under 20 trades. We can forgive these:

CL_ON_4 (12 trades and 1.4 profit factor)
All opening gap systems get forgiven because they trade rarely but rarely miss, I am positive about this.
GBP_ID_4 is good enough (I guess 1.5 would be acceptable)
...

...hmm, this analysis of profit factors has made me realize that I really do not have much of an edge, and I can't afford to discard profit factors of below 1.5, but should be happy to discard just those below 1.25. Because most of my edge stays with systems that have about 1.5 of profit factor. I am glad I didn't discard this measurement as I was going to do a couple of days ago.

Essentially, due to profit factor, I know that I should discard those 7 systems that have traded enough and that still have a pf lower than 1.25.

2) discrepancies between profit factor and ROA index?

Just as profit factor is not complete, the other index could also be lacking, and they indicate conflicting ratings for my systems. For example, if a system traded 2000 times, and each time either made 11 dollars or lost 10 dollars, such a system would suck, and have a profit factor of 1.1. On the other hand, in terms of ROA, this system would score very high, because its maximum loss and drawdown could be as low as 1 dollar. Furthermore, this system would make 1 dollar every two trades, and the 1000 profit would get divided by 1 dollar loss and drawdown, giving me a stellar ROA rating. I have such systems. And they're not worth trading, as they keep my capital busy for no profit. I was alerted as to that by seeing a very low average gain per trade, which in many ways is an indicator similar to the profit factor. Now let's see if profit factor demolishes any high-scoring ROA index systems.

Not really, I was even wrong about that ZN hypothesis (that scored high on one ratio and low on the other). I guess right now the PF is in line with the ROA. Actually some PF values are telling that some systems are better than what the ROA values would have had me believe.

From this little study, I'll retain that it was a good idea to compute an average index of ROA and PF. And that systems with PF lower than 1.25 are to be discarded.

I know you mentioned that you don't like the Sharpe ratio, but all it does is calculate risk-adjusted performance. In other words, enormous returns with the occasional big drawdown may do no better in terms of Shapre than a small return system that is very regular.

Instead of fretting about which system passes the ROA and/or PF tests, you should find some very interesting results if you take a "portfolio approach" to including or excluding systems. Combine the p&l of many systems and calculate the overall Sharpe... its not that hard to do. The system that looks poor, may make its money when the others suffer. You want to avoid losing systems of course, but if the equity curve is flat, or only loses a little, in the bad periods that's not too bad

Many traders have very few systems. You have loads. And they could trade many, many contracts as these are liquid markets. If you do what is suggested here (max an hour or two for an excel wizard!), you could find you get excellent results from only 5-10 systems that combine well i.e. select systems based on their contribution to the portfolio Sharpe.

 

[FXTrend240]

im lazy lol; the CME currency futures; is it spread based ( as in like retail forex) or commission based ( as in you pay like $5 RT rather than a 1 pip spread). I'm trying to conclude whether ECN spot forex is better than futures or vice versa.

 

[Yamato]

Re: back to work: profit factors

I know you mentioned that you don't like the Sharpe ratio, but all it does is calculate risk-adjusted performance. In other words, enormous returns with the occasional big drawdown may do no better in terms of Shapre than a small return system that is very regular.

Instead of fretting about which system passes the ROA and/or PF tests, you should find some very interesting results if you take a "portfolio approach" to including or excluding systems. Combine the p&l of many systems and calculate the overall Sharpe... its not that hard to do. The system that looks poor, may make its money when the others suffer. You want to avoid losing systems of course, but if the equity curve is flat, or only loses a little, in the bad periods that's not too bad

Many traders have very few systems. You have loads. And they could trade many, many contracts as these are liquid markets. If you do what is suggested here (max an hour or two for an excel wizard!), you could find you get excellent results from only 5-10 systems that combine well i.e. select systems based on their contribution to the portfolio Sharpe.

Ok, it's a deal. If you can make the Sharpe Ratio very simple and logical to me, as would be a "Sharpe Ratio for Dummies", then I'll deal with it. Feel free to ignore my post if you're too tired to reply. I am giving Sharpe Ratio a chance, only if it incorporates all the indicators of a system's performance. In other words, I am expecting it to be better than PF and ROA put together. Beware that the name "Sharpe" is frightening to me, because I feel like I could get cut dealing with this formula.

 

[Yamato]

im lazy lol; the CME currency futures; is it spread based ( as in like retail forex) or commission based ( as in you pay like $5 RT rather than a 1 pip spread). I'm trying to conclude whether ECN spot forex is better than futures or vice versa.

I would say go for the bigger market, so futures. It's harder to get ripped off in general. Go with the futures. As long as we're talking about automated trading. If we're talking about discretionary trading, based on my experience, don't trade anything at all. I'd advise anyone against discretionary trading, no matter what market. But if you have to lose or find out you're not profitable, maybe ECN is better, because you can invest less, and even diversify more as a consequence. But once again I advise against discretionary trading. If instead we're talking about automated trading, you do need futures because you need a big broker like Interactive Brokers which will make your automation much easier. Besides, as a general principle, I would go where there's more customers, more volume... because everything works better, you know, economy of scale first of all, and also in terms of customers (e.g.: if there's a technical problem, bug, you won't be the first one to notice). But it's true that FXCM is just as big... well, you know what I meant. Big is not always good, but IB is big in a good way.

To answer your question, IB does futures, and yes it's commission-based. And they have the lowest commissions.