Who's Testing Z1 or Z2 ?

I thought it would be interesting to start a thread for all the folks who are testing Zorro, either in demos or live trading. This is just a thread for us to compare notes & ideas. In all of the other threads we are asking technical questions for jcl, maybe we can start learning with & from each other too. It would just be nice to hear how things are going from everyone else.


I started a demo on May 26th running Z2nfa, on FXCM. I set the Margin at 75 and the Risk at 24. It began making trades within about a day. I've seen it trading the EUR/USD, the USD/JPY, and the USD/CHF. It's been ahead the entire time, but seems to make most of the losses with the EUR/USD. It has been hovering around $50,040 to $50055 for the last few days. If it would quit loosing on the EUR/USD it would have a lot more profit. It made a lot on the USD/CHF yesterday.

I'm wondering when Zorro will buy the USD/GBP. If it begins making better trades with the EUR/USD I think it will begin making some good profits.

So, how are your tests going? Let's all post here and share what we are seeing, and whatever we are wondering or thinking about the progress.
Have any of you been programming scripts? How has that been going? Have you managed to get them to test as well, or better than Z2? What info did you read to help decide which strategies to try?

hi GlennR,

I'm testing Z2 on a spreadbetting platform using the MT4 plugin. It has been running since 5th June, but has not taken a trade yet.

It may be issues due to the indices and commodities being provides as futures rather than CFDs and I had to disable them, or it may be due to gaps in the history data.

I wish I could test more than one broker at a time. It would be nice when testing with the spreadbetting company to be able to compare the bet sizes and lot sizes, differences in execution and data feeds between the brokers, differences in profit calculation methods, etc. But at the minute I'd settle for simply being able to see if it should be taking trades or not.

Good idea opening topic like this. Had something similar in mind...

Anyway, I've been trading Z1fx & Z2fx together on the same UK demo account for a few weeks. Started with demo $50000, been down to ca 49000 and up to ca 51500. Sometimes Z1fx leads, but Z2fx quickly loses what Z1fx earned. Other times it's vice versa. The equity is $50770 atm.

I used to up the margin & risk @ 100, 50 respectively, but got tired of moving the sliders everytime I restart Zorro, so nowadays I let it do what it wants (i.e. use the default values). It's not like I understand what those sliders do, anyway.

Anyway, Z strategies are nice and fluffy, you start them and they do their thing. But, as they're trading on the 4h timeframe, I'd say that at least a year of trading would be needed to make any useful judgements. Because running 2 strategies at the same time quite complicates things, I will stop trading Z2fx, and continue only with Z1fx. Effective this weekend. I will also commit much more time to development of my own strategies from now on. I wanted to do that earlier, but didn't have enough time, I hope it's goin' to change next few months. Famous last words.

Hi acidburn,

How do you trade them together on the same account? I get a broker interface error.

I ran the backtests on various levels of risk and margin.

I understand margin is the average margin committed to a trade. Risk factor is kind of soft limit on how far that can be multiplied for an individual trade, but it can exceed that to allow a trade take place, e.g if the min lot size is too big for the calculated lot size.

In backtesting, risk tends not be as influential as margin in Z2 at least, and for the same margin there is a sweet range for risk (10-30 seems to it for my proposed level of real investment) Beyond 30 the parameters appear to level out, below 10 overall performance seems to be poorer.

The margin is the one to adjust first until you get parameters for capital required close to what you are prepared to lose for real, then adjust risk to see if it improves the output.

I'm not sure how useful any of that is, it may be another version of overfitting, either way I'd decide the parameters and let it off, adjusting the sliders afterwards doesn't sound like it would be helpful, especially not if you are evaluating.

At last a trade was triggered on Z2 USD/JPY long 6 at 97.40 Risk 144. Not a bad looking trade on a 4hr chart to my unscientific trader eye. my settings are 50 10 0 - giving an sort of max risk per trade of €500. I think the 144 means amount risked on this trade in €.

The stop loss in metatrader is way more than that, but from another thread, I think Zorro manages its own stops.

As it is spread betting the 6 lots translates into 0.6 (6 by 10c per pip bets) spreadbetting is way simpler than calculating lot sizes as there is no currency conversion all bets are in your deposit currency, if any pair,index or commodity moves 100 pips in your favour you simply multiply the pips won by your bet per pip.

I am still wondering if Z1 and Z2 can manage this difference unmodified.

You can't on the same machine. You need 2 different machines to trade 2 strategies at the same time.



Thank you for sharing your insight. Though, I will probably stay confused for a while. And I'm not complaining, it can't be any different when an aspiring quant with a lot to learn meets black box strategy. Having backtests and some documentation helps, but I still need to spend some time with Zorro before I can even properly formulate questions to ask. One promising thing about all this is that jcl is such a nice guy, so I'm confident it's worth investing my time in Zorro.

One thing, I've seen elsewhere on the forum that max risk per trade could be approximated by risk * margin formula. But, that confuses me very much, and here's why: say you invest $5000 capital (should be enough for both Z1 & Z2) and run the strategy with default 50/10. That would mean you risk up to $500 or 10% of the capital per trade, right? But, what if you run the same default parameters on the demo account, which btw can be opened only with $50000 of starting capital. Now default parameters and max risk per trade calculated by that formula mean something very different. You're now risking only 1% of the capital. By that measure, earning 1.6% after a month or so, as I've managed to do, is quite nice. But I wonder how much would I have earned if I matched the risk of the $5k account? But, alas, I can't move the slider to that position. It stops at 50, so max risk per trade becomes $2500 which is only 5% of the capital. And yes, I started moving both sliders to the right, after I learned about the formula, but I forgot to do it so many times, that many trades ended open with default 50/10. Argh.. So, as a conclusion either the formula is wrong, or the sliders are not adequate or I got something wrong. Currently, I'm heavily betting on the third option.

Anyway, my Z2fx has also opened long 2@97.42 USD/JPY. See, only 2k, with 50k in account, such a shame. Too bad the trade won't stay alive much longer, as per my plan to stop trading Z2.

I'm skipping your commentary about SB, because I know nothing about that, sorry.

Yes jcl is unbelievably helpful.



OK - I have no insight into the black box either, so I'm only surmising from my reading of the manual and the posts on here. Black boxes make me uncomfortable, but not as much as the feeling that all the old indicators I've been using unsuccessfully are unproven rubbish and I need to get smart on a whole load of maths on signal processing and statistics.

My understanding of the strategies underlying Z1 & Z2 are that %equity is only important in the context of avoiding margin calls, i.e. minimizing risk, protecting capital at all costs, minimizing drawdowns etc.

The backtesting module is designed around maximising return from the minimum capital expended. Therefore, the systems try to analyse and "learn" where to allocate precious margin to the assets and algos that give the most robust return, i.e. steady returns with minmal drawdowns. Jcl has proven countless times how reinvesting %equity leads inevitably to blowing out an account, so I'm no longer interested in compounding %s of my equity to infinity. I'm now happy to try to eak out nice steady returns with static margin at risk.

This is where the risk margin comes in, they are sliders allowing you to more or less set the max drawdown you can tolerate while operating the strategies over a long period. The max drawdown plus a calcualtion of how much margin you need to run up 30 open positions fall out of the margin per trade you are setting here.

I am happy to run this at 50 10 while testing. Backtesting at these max margin levels per trade, means I would need € 2,658.00 to deliver a return approximating € 319.00 per month over time. I would be happy to run this live at those levels and am uninterested in the demo account equity.

I am however extremely interested to see if it performs as per the backtests and whether the strategies can somewhat replicate the backtested performance. What the real returns, profit factors, sharpe ratios, ulcer index, max drawdown are in demo trading.

If you are interested in finding out a different question, how would the z strategies perform with 50,000, then set the sliders to the right, but remember as margin at risk increases, so will the max drawdown.

The reason the sliders stop is that jcl & co, not unfairly, would like you to buy the zorro s product for more serious trading.

Also my understanding is that the sliders are set and forget unless you wish to reduce them. Increasing them is reinvesting profit, and not advised. Remember they are overall margin at risk per trade settings not risk per individual trades. Even if the sliders allowed a 50,000 per trade margin, the robot would not use this on each trade as it uses a optimised margin per algo asset model.

I hope for 2 things:

1) I actually understand correctly what I've been reading - otherwise jcl might dive in and set us all straight
2) I've made things clearer rather than muddier.

I think you guys are a little ahead of me, but we are essentially in the same boat.

I think we are all a bit unsure about how Zorro is handling the $50K of demo capital. I have the feeling that Zorro might be configured to ignore the $43,000 extra, because it is opening such relatively small trades.

I've looked at the performance graphs and it looks like Z2 is supposed to begin slowly, then make progress more rapidly. I guess it has to study some before it feels comfortable making bigger trades and trading more pairs.

I'm concerned that it will need more time than the 30 day limit that I have with my FXCM demo account. I don't know how to extend the account. Maybe there is a way Zorro can continue with a new account #.

It will be exciting to watch it work with a lot of different pairs at the same time.

I wonder if Z1 & Z2 are only able to operate in the 4 hour time frame? I haven't seen it mentioned in the manual. Btw, the manual seems "huge" as a beginner. I know it will all become more familiar as we work our way through it. I haven't printed it yet, but I think that will help me to work through it section by section.

I am surprised there are not more people using Zorro. I think a lot of folks who have been trading for a while have seen a lot of promises come and go, and most are scams.

Glenn,

The FXCM demo doesn't expire, so long as you trade every few weeks. At least hasn't over the last year.

4hr is hard coded into the scripts so adjusting the bar slider does nothing, I think.

I'd like to be able to back test different time periods, but they seem to be hard coded too.

I'm also naturally suspicious of something new, but Zorro ticks all the boxes when it comes the addressing the limitations of MT4, so it is worth a decent test.

swingtraderkk,

Good to hear FXCM demo doesn't expire.


I wonder where the other Zorro testers are? I can't believe there are only 3 of us. Hopefully it's because the results are so good that they're afraid to share.

I know, the lack of people posting along the lines of ....

"OK been testing Z1 & Z2 successfully for x months and am now taking the plunge with live trading, will keep you posted ..."

makes me wary of putting on real money.

The other issue may be that it would take a long time of forward testing to be confident enough to put real money down. the last year has not been a great time to test any system. Crazy stuff happening on the indices with the distortions of QEs. Now I think the opposite, if Z1 & Z2 can handle these volatile times then they are robust, but you need to stick with the system to see, many would lose interest on big drawdowns. I think from another post, Z2 did not have a good year in 2012, but ended up in profit, how many would have stuck with it through a year of testing?


It also takes the commitment of running zorro 24/7, not something many retail traders have the infrstructure, skills or discipline to do.

I'm testing since last Wed Z2 on a spreadbetting platform, only one trade so far, a long USDJPY that has profit locked in.

I'll try to keep updating, let's see how my discipline holds

Mine has traded the EUR/USD, USD/JPY, & USD/CHF

USD/JPY only right now.

Since it's not betting much money per trade I decided to move the Margin slider up from 75 to 100. Maybe it will place larger trades? Risk is at 24. Maybe I should slide it up too, but I'll wait to see if the Margin adjustment makes much difference for a couple of days.

Glenn,

If you are only testing, I would caution you against moving the sliders at all.

The question you are trying to answer is whether the backtests are sufficiently reliable, not curve fitted, in live trading to warrant investing real capital. You then know whether your live performance (either real or demo) is within the parameters to be expected or not. If yes, back tests are valid then continue, if not within parameters, then you need to investigate why and either adjust parameters to account for the new reality or stop.

If you change the sliders you are invalidating the question, as you are no longer evaluating like with like.

I am leaving the sliders at 50 10 and evaluating the performance live against the saved backtest. It should then deliver returns, drawdowns, max margin employed comparable to the backtest. If it does perform within parameters over a prolonged period, I'll trust it enough to try a microlot account, If the microlot also proves that the backtests are a reliable indicator of margin required for x return, I'll concern myself with correctly sizing the sliders for the account size I wish to trade.

Hmmm?

Well, I did run a series of tests using a variety of slider positions before starting the demo.

I noted the results and found that Margin-76 & Risk-24 produced the highest annual return. Z2fna produced +196% / +39807 (capital required $2277)

Margin-100 Risk-50 produced +192% / +51216 (capital required $2980)


I don't know about the parameter ratio details, that you mention. But I was just hoping to give Zorro a "little prod" to start buying larger trades. I read that it begins slowly, but I don't quite understand why.

Actually, I don't really know the $amount that is being traded. On the FXCM platform there's an "Amt" column. Now it says "1" for the USD/JPY trade that's open. I've never seen it over "3" for any trade yet. My "Usable Mr" is 99.96%, so I don't understand why Zorro doesn't use more.

I wish the demo could be set to work with $7,000 like the Zorro restriction, so I'd have a better idea of what to expect with an account that size. I'm a bit confused about how it compares.

Glenn,

I think that at the slider settings 76 - 24, the maximum at risk per trade is 76*24 = 1824 of whatever currency you are trading. There are a few caveats to that and I suggest reading the manual to clear it up. Few if any trades will actually put that amt at risk, as margin is allocated based on performance of the pair and internal algo, I think Z2 has 4-5 algos per trade. Rerun the backtest with the same settings on a new instance of zorro, and read the performance report, it will give the max margin at any one time and I'm sure it will be a small fraction of 1824.

The key figure is capital required at 2277. According to the back test you only need 2277 in your account (I think this is at the end of the period to safely continue the strategy) to get an annual return of 196%.

The bit I currently don't understand fully is whether your current balance ever feeds into the lot sizing. i.e. the 196% return is calculated on the margin you put at risk over the period, and the capital required is based on historic margin required and max drawdown. I think not, but it may be possible that there is some limit on lot size linked to balance or equity independent of your max risk per trade setting of 1824.

As I previously stated, for now all I'm interested in is the accuracy of the back tests in calculating returns and capital required. If accurate I'll be investing in an unrestricted Zorro S and upping the numbers, but remember the logic of trying to force Zorro to put on bigger trades defeats the whole idea of Z1/Z2 which is trying to manage the allocation of margin to only the most profitable asset/algo combinations without reinvestment.

Z1fx here has an open position since 2013-05-29, short USD/CHF @0.96149 size 46k. It is currently $1362 in profit. Unfortunately, I don't remeber what settings I had for margin/risk when the position was initiated. But, I often see trades of that size open. Typically accompanied with a few much smaller trades (9k, 10k).

My Z2nfa demo just isn't making any progress. It lost a lot on the GDP/USD yesterday, and is having trouble so far today with the USD/CHF. I will continue to be patient for another week, or so.



I'm in the US and need to use the "nfa" version, but it looks like the manual says I can use the "fx" versions also. Is that correct? If so, maybe I will switch to one of those.

Hiho guys,

Amazon offers a VPS for one year free. So you can run Z1/2 24/7 :-)

(Just one of the links you will find when searching the internet: link )

I by myself have running Z1 for some days. Till now only loosers... But those strategies need time i think and a winner should outperforme a looser. So i wait...


happy new week...

Excellent link Sundance thank you!

I could do with an idiots guide on how to setup and run zorro on a vps though, this stuff is hard!

Can I create multiple server instances on that free offer and run z1 in one and z2 in another all on demo accounts and not violate either the zorro terms or amazons and start to incur charges?

Today my Zorro2 finished the day at 50058. Not sure how long it should take to begin making profits, but I'll keep watching it run.


In the meantime I'll keep reading the manual & learning Lite-C.


I noticed that Z3 info has been posted, so I will probably switch to it when it becomes available.

Thanks Glenn for the head ups about Z3. (Can you show me the post? Can't find it... )
The manual says that most strategies including Z1/Z2 will start with some loosing trades. So i'am not concearned at all.

@swingtraderkk: You get an VPS with Microsft Server 2008 operating system. You can do whatever you would do with your local computer. But remember the RAM ist only 640MB. At the moment i have Zorro and one MT$ instance running. MT4 runs only cause of updating my MT4i account.

It's in the manual now. The page "Income with Z1 & Z2" has been updated. You have to go through the Zorro home page, the manual through the "help" tab doesn't get the updated version yet....or late least mine doesn't.

Ah. Okay. Thank you

Just a short reply. The Z2 running on the Amazon VPS is doing better. Made some good trades now and has now done more positive then negative pips :-)

It is good to hear that the VPS works for you. I was curious about it.

I think it will be handy for testing scripts, when I get the skills to write them....

Amazon VPS is working for me too.

Z2 was doing well for me too, but I closed it down and closed out the trades to install and run Z12 yesterday. Not doing so good so far.


I'm concerned that all this chopping and changing is ruining the test, Zorro needs to run constantly and consistently.

Wish I had an option to test more strategies simultaneously, as I'm kind of committed to testing on the spreadbetting platforms, and Amazon gets expensive with more than one VPS instance.

You are right. I will let Z2 now run for the next view weeks. Since yesterday it had some positiv trades. I have no direct solution to the problem of testing more than one strategy. I have another VPS which i bought for a monthly fee of 4,50€. It has 20GB HDD and 1GB RAM. Enough for another MT4 instance...

Do you need the VPS for the computing power to run more instances at the same time for testing different strategies? Would a more powerful & faster computer solve the problem or would you also need a better internet connection?

I remember the manual saying a fancier computer is necessary for testing a number of strategies at the same time.

Here's one that looks powerful & fast to me for cheap ( I don't know a lot about computers)

Anyway it has an 8 core processor and is under $400. The 6 core one is about $250. Maybe that would be better than the VPS?

http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=8148341&CatId=7718

I think the Intel i7 is only a 4 core.

That seems like a fair price for a small Win VPS. Would you care to share where you have ordered that VPS?



Unless you're backtesting rather complex strategies on lower timeframes (<1h) Zorro is plenty fast and shouldn't require anything better than cheap commodity hardware. Also it can't use multiple cores from what I know (it's not internally multithreaded), so the only way to fully utilize today's multicore machines like the one you mention is to run multiple instances of Zorro, and backtest different strategies at the same time.

I believe Sundance is just looking for a cheap way to demo trade multiple strategies at once, which currently can be done only if you have multiple VPS/machines. Otherwise you get 'broker interface busy' errors. IIRC, that limit is lifted in the Sponsor edition.

Stopped Z1fx trading, closed all trades, balance $50222.

Installed 1.12.0beta, started demoing Z12 FXonly, Margin 100, Risk 50.

I thought Z12 was just available for Zorro S. I didn't realize it's available for demo now. It seems Z3 and Z5 are too.

So, how does the VPS work?
I set a VPS up and use a separate one for each FXCM demo?

I tried to set up an Amazon one for free, but it required a CC# and I didn't feel comfortable giving one. So, I found vps.me/ that has free ones too. No CC# required, but "something" didn't work during registration. It's getting late, I'll try again in the morning.

Z3 looks interesting. I'd like to try it and Z12 after getting the VPS going.

Hi Glenn,

this free VPS has 348MB RAM and 5GB HDD. What will you setup there? Setup a Win OS won't do... Please give us a feedback on how its performing...

Hi Acid,

here you can get the VPS für ~4,50€/month:

LINK
This is an affiliate link. There you should see the VM or can configure it that way. The price will be a little less than normal and i will also get some cents :-) Its using cookies so don't deactivate cookies or the link won't function.


You must select the KVM vServer and at first you can only select Linux distributions. But when finished the registration you can change the OS to server 2012/Win7/Win8. There are some images you can use but you must have a windows key-code on your own.

Thats it...

I don't know if I can figure it out. Looks like I'll have to choose a Linux OS for it, which may be simple, but will complicate it for "me". I was hoping perhaps somebody here with more computer skills/knowledge would be interested in it, since it's free...and tell us "how easy" it is (or isn't!).

I've got a full "to do list" today. Maybe I'll have time for it later in the week.

I wonder if there are any free VPS that use Windows OS & do not require a CC#. (I just don't like giving it out for "free" things.)

Thanks Sundance! Will check it out later...

Instead of a VPS how difficult is it to use a different broker? That would also solve the problem of running additional demos, right?

Many roads lead to rome :-)
The advantage of a VPS is that it runs outside your home and is 24/7 online. At home in most cases you don't wan't a computer running 24/7. But surely you can buy a PC and run several instances with different brokers. But if i remember correctly i don't know if it's allowed to run Z1 and Z2 at the same time with the same broker!? There was something JCL posted somewhere.... But i can be wrong...

I can run one demo instance of zorro with the fxcm plugin and one with the mt4 bridge on the same pc or vps. So this is a max of 2 strategies being tested per PC or vps without zorro S.

The solution is zorro s or multiple vps, this is simply a cost issue. I'll have no problem committing to zorro s and a robust vps solution when I have successfully tested the strategies either z strategies or my own.

I'd never run a live robot on real money on a home pc, I trust neither my electricity supply, my broadband provider or my fellow dwellers not to mess it up.

Testing is a bit of a nightmare though, as I've limited scope to test the z strategies and my own.

I think I know the answer to this one, but I'll ask anyway, jcl why can't there be a zorro version that is limited to demo accounts only but allows multiple Demo instances to test multiple strategies on a single pc or vps?

Lol. Those guys at vps.me are really funny. I tried the free VPS but when it comes to verifying the account throug a SMS i get a 'Clickatell: ERR: 108, Invalid or missing api_id'

This is since yesterday and i wrote them a message. Also when i registered i got an email as a confirmation. This email had my given password in clear letters. This is a totally NO GO and unprofessional. What are those guys doing? When i look at there forum there is only one entry. Seems they are really new to the whole market. With the picture they have given me so far i think in three month they are gone. :-)

To make this clear, it is allowed to run Z1 and Z2 at the same time, but only on demo accounts, not on real accounts. It does not matter if that's with the same broker or with different brokers.

Ah. Okay. I was then near to the truth.

FXCM seems to be the problem with running multiple demos at the same time. I get a "broker busy" message if I try to open a second demo.

I should have mentioned that although it's allowed for demo accounts, the FXCM plugin of the free Zorro version can not trade with multiple instances. So you need to run it on several PCs or VMs for several instances.