This article is obsolete and no longer maintained.
The first thought I had about this EA was something along the lines of “omg lolz what an uninspired name, nobody trades Forex on the Wall Street”. Ok, I don’t really think in terms of “omg” or “lolz” and I am much harsher than “uninspired”, but that was the main train of thought. At first glance, I believed it to be yet another Asian session scalper and I was wondering if it’s just a copy of an existing EA or something original. But when I got to inspecting the opening times of the positions, lo and behold: I noticed that it wasn’t restricted to the Asian session; trades were being opened in the forward tests all around the clock! You don’t really see many 24/5 scalpers nowadays and given my very suspicious nature I was inclined to believe it’s some kind of a scam. Many EA sellers have been caught going as far as faking statement results and to clear my doubts I naturally proceeded to ask for a Wall Street Forex Robot review copy, request that has been granted with a very prompt email answer.
The fact that I’m now writing about it should be clear indication that I don’t consider it to be a scam in any way: it’s a truly original scalper that trades several pairs around the clock and it seems to manage scraping a nice profit off it. The pairs it runs on are EURUSD, GBPUSD, USDJPY, USDCHF and to a lesser extent USDCAD, although the last one is not even mentioned in the manual and I only happen to know about it because the author mentioned it. As many other scalpers nowadays, it runs on the M15 timeframe, so there’s nothing out of the ordinary there.
Well, you already know the important facts: Wall Street Forex Robot is a scalper, it runs 24/5, it uses the M15 timeframe and the pairs recommended by the vendor are EURUSD, GBPUSD, USDJPY and USDCHF, while USDCAD is working but not officially supported.
The manual is a bit confusing on the topic of the supported pairs as in the beginning only EURUSD, GBPUSD and USDJPY are mentioned, but later on it proceeds to mention USDCHF as well; in addition, as you will see, USDCHF is present in the backtests on the product website as well as in the official live forward test.
There are a few other things that you should know and I’ll try to briefly list them. Although you can manually configure the SL & TP for each pair, it’s probably not a good idea to do it. The EA updates its settings from the server upon a successful authentication, configuring (among other things) the preset values for each pair; the stop loss ranges from 120 pips on EURUSD and GBPUSD to as high as 160 on USDCHF, while the TP is around 25 pips, with the notable exceptions of GBPUSD where it’s 50 and USDCAD where it’s 14. The stop loss is rarely hit, though: from what I could see in the backtests, like any sensible scalper EA, it’s able to close the trades before they hit SL in most cases when the market goes against it, the average win:average loss ratio being roughly 1:2.75. Just like its ability to close trades before they hit SL, it has an ability to take profit early, before its positions hit the take profit target, when it figures that’s as many pips as the market is going to give it.
The strategy itself is not excessively complex: it uses a few of the indicators that are shipped with Metatrader in a creative way to determine entry signals. While I was taking a look at this, I also checked its order sending and I noticed it has retry loops for opening/closing orders, denoting a certain degree of experience with automated live trading. Although the DLL programming can sometimes be a problem for EAs running on multiple pairs with the same DLL, in this case it seems to be fully thread safe.
The Wallstreet Robot averages about 3.5 trades per day when running on the recommended four pairs and almost 5 trades per day when running on all five, so it qualifies as a rather frequent trading EA. Since it trades around the clock, there’s no GMT setting, so (thankfully) that’s not something that we have to worry about in this case.
The EA is fully compliant with the NFA rules: since it doesn’t open more than one trade at once on each instrument traded, there are no problems with FIFO and hedging. Careful, though: operating it together with other EAs that trade the same pairs on an account with the NFA restrictions in place might result in one of the EAs not being able to open trades.
If you don’t know it by now, I’m blunt. So, I’m sorry, mr. vendor, but I dislike the Wall Street Forex website. It’s an amalgam of classic marketing stuff with fonts of all kinds of sizes and colors and I’m inclined to believe it’s a genuine attempt to blind me. It’s probably what made me think I was dealing with a scam in the first place.
There is some interesting info, too, if your mouse wheel can take the scrollfest. Four 1999-2011 backtests can be found, but they’re performed using a large lot on a low starting balance, resulting in some rather large drawdowns, despite the steep ascending balance curve. More interesting than that, there is a live Alpari account running the EA since December 2010 and three Alpari demos, but I didn’t really understand what’s the difference between the demos other than the starting date. All of the accounts are stamped with the independent verification of myfxbook. You can scroll directly to the HTML abominations that encompass a scrolling account statement in a graphic frame (which are impossible to miss as they tend to fill the whole screen) and then scroll up just a tad to get to the live results. Come to think of it, the site is so loaded with various stuff that I felt the need to provide directions and I almost forgot that I can also link the myfxbook widget here:
There’s a scary notice at the top of the page that states only 200 copies are going to be sold but I am taking the liberty to disbelieve that. Anyway, if you’re reading this, chances are you’re not much into that marketing style either, so I ask you to disregard the website and proceed with the review. An ugly website is no reason to dismiss a good product and after all it’s just a matter of taste. Who knows, perhaps some people actually like this marketing style.
But all this rant made me almost forget a very important detail mentioned on the website: the EA is already available for Metatrader 5. Buying the MT4 version will also give you access to the MT5 version, which is available for download in the members’ section.
The only parameter that you really have to set is the AutoMM. This is basically the risk that the EA runs with and failing to configure it will make the EA trade a constant 0.1 lot size. I’m going to run my forward test with AutoMM set to 3, but I suspect many of you will set it to something like 5. Unless you want to use fixed lots, that is, in which case you’ll have to change the default of 0.1 to your desired lot size.
There is also a RecoveryMode parameter, which I strongly recommend not enabling. While it may lead to apparently higher gains, it will surely increase the drawdown dramatically. Even the manual makes a similar recommendation. There’s also another parameter for limiting the maximum risk per trade, which is configured to 20 by default and you’d probably do well to leave it there, it’s in essence a safety setting.
The Wallstreet Robot features a StealthMode parameter and you’re likely better off disabling it unless you suspect your broker of misbehaving. I’ve recently seen some very heavy losses due to a similar setting with a different EA.
Instead of using the pair defaults, any user can configure the pair parameters manually. The stop loss, take profit, secure profit and secure profit trigger can be configured, which should give many weeks of busy CPU time to the computers of the people who are into optimizations.
All the parameters are described in the manual and there’s a whole chapter dedicated to money management which explains how to correctly configure the EA to use the risk that you want, but an additional thing that you have to keep in mind is that each currency pair has its own inherent risk and while running the EA on multiple pairs usually means smoothed out drawdowns, there will be times when it also means combined drawdowns. The manual also features a backtesting chapter that I carefully disregarded and proceeded to doing it my own way (which is almost the same when it comes to history center data anyway). It also contains some really sensible broker advice and no affiliate links, which kind of surprised me, given the product website.
The EA features a HUD (a chart display, if you will) more or less in the same style as KangarooEA but with a bit less information. Most importantly, it displays the lots it will trade, the current spread and trade status along with the authentication result.
As usual, I began by running history center data backtests on the 1999-2011 time range. I chose an average spread for each pair:
EURUSD: 2.0 pips
GBPUSD: 2.0 pips
USDJPY: 2.0 pips
USDCHF: 2.5 pips
USDCAD: 3.0 pips
Sure, if your account is on an ECN broker you will have lower spreads and that’s going to affect the performance of the EA in a positive way. However, if you’re using a fixed spread broker, the spreads might be higher and in the case of the Wall Street Forex Robot it’s probably not a good idea to use such an account.
All the backtests performed on history center data were using the default settings for all parameters except the AutoMM which was set to 3. During the whole review, I’ve used a GO Markets terminal for all the backtests and FXT file creation.
GBPUSD outperformed all the others, but EURUSD, USDJPY and USDCHF are also displaying a nice, constantly ascending balance curve. USDCAD on the other hand looks quite chaotic and I guess that’s why it’s not an officially supported pair. I will definitely not include it in the forward test, but I will run the other backtests on it as well, for the sake of consistency. The reason for the USDCAD weird balance curve is probably the poor average win:loss ratio, despite its high percentage of winning trades.
To save you the trouble of clicking each balance chart, I made a small table with some relevant data for each backtest. Speaking of which, many of my readers don’t seem to know they’re clickable so here it is: clicking one of the backtest result charts will load the whole backtest statement in your browser.
I proceeded to perform some backtests on Dukascopy tick data next, using fixed spreads equal to those used with the history center data above. The EA settings were also the same: all defaults with AutoMM set to 3.
The GBPUSD and EURUSD are still the reigning pairs, just as in the previous backtest series. Surprisingly, the USDCAD looks better, but that’s still not reason enough for me to use it in forward testing, especially given the fact that it’s not an official pair and I tend to go with the out-of-the-box settings if possible. If you decide to run USDCAD even on demo, please be so kind to drop a comment with your results after a month or so.
I’m going to put together yet another table for you, this time with the tick data fixed spread results:
I also merged and analyzed the results for EURUSD, GBPUSD, USDJPY and USDCHF (in other words, I excluded USDCAD). To get relevant results, I set all the trades to lot size to 0.3, which is the value obtained from running the EA with an AutoMM of 3 on a 10k balance. As such, the chart below would’ve had a much larger ending balance with the original lot sizes, but for the purpose of our analysis, this is good enough:
The overall average monthly return is somewhere around 8.3%, while the combined maximal relative drawdown lies around 20%, which is what I was targeting as a maximum figure when I selected an AutoMM of 3. Many trades (over a third) are closed within one hour while the majority of trades (over 80%) does not last over 5 hours. Generally, the longer a trade is prolonged, the lower the chance of ending up with a profit from it. The average profit factor is 1.59, while the winning trades constitute a 78% of the total, with an average win:loss ratio of 1:2.28. I also ran a quick Monte Carlo projection, which I am only going to link here because not many people are interested.
Finally, I ran some backtests using the real Dukascopy spread and adding a commission of 0.8 pips:
The results look about the same as those obtained on a fixed spread but as expected, due to the commission and variable spread, they were slightly poorer.
I used my Average Spread EA to check out the spread of the whole data and let’s see what it looks like on the real spread files (it does not include the commission, that’s added on top of the spread):
Smoothed min spread
Smoothed max spread
When adding the commission, this is sensibly larger than the fixed spreads I used so it makes a lot of sense to have results that are slightly worse, but I guess these are more realistic. Since I made tables for the two other backtest series, I decided to create one for this as well, just in case you want to compare the results:
New pairs and settings
Recent versions of the Wallstreet Forex Robot have added support for AUDUSD and NZDUSD, together with a whole new setting for EURUSD with a tight 33 pip stoploss. This naturally prompted me to revisit the article and include the backtests of the new pairs but unfortunately I’ve been caught up in a whole bunch of other stuff and this update is coming many weeks later than it should have.
Initially, I only ran tick data backtests using the Dukascopy data with variable spread but when I noticed I mistakenly ran them with risk 2, making them not easily comparable with the previous backtests, I decided to also run some history center data backtests and some tick data backtests with risk 3, all of which you will find below.
Of the three, I would say the new EURUSD setting is the most exciting item but I won’t switch my forward test to it simply because I prefer to keep it as close to “factory settings” as possible. Wallstreet Forex Robot also performs rather nicely on AUDUSD, in line with the other pairs. However, when it comes to NZDUSD, even though the backtest seems to indicate a decent performance, the amount of trades taken is a lot lower than for the other pairs, so the backtest is not as relevant as the others. Both new pairs have been added to my forward test a while ago and I will keep them there and let the live performance speak for itself: just use the custom analysis option on myfxbook to get an idea.
Personally, I like Wall Street Forex Robot, otherwise I wouldn’t have written a review. It does have a few drawbacks, though, and when I say this I am not only referring to the style of the website, but also to the rather steep price: a copy costs $489. The whole deal is sweetened by the fact that it’s licensed to run on three live accounts and that a MT5 version is already available, but it’s still not something that I would count as a cheap product. With a $1000 account, you’d have to run it for something like half an year with risk 3 just to have it make a profit equal to the investment in the EA, so I’m guessing it’s targeting people with larger account sizes. Finally, there’s a $50 discount exclusive to eareview.net readers, bringing the price down to $439 – all you have to do is buy using this link to take advantage of it.
Sure, given the fact that it’s pretty much the only scalper that I know of that’s profitable and opens positions around the clock, I can see how the high price might be justified. The official forward test seems to be fully in line with my backtest results, so it looks very promising so far. Finally, it’s worth mentioning that there’s a 60 day refund policy, conditioned upon the EA failing to produce profit.
Due to the spread requirements, I was planning to go with a GO Markets L-Plate live account, but as it turns out there’s a limit of one per customer. As such, I set up a LiteForex floating spread live account for it. It’s running with the default settings (including recovery mode which is disabled), the only change being that I set AutoMM to 2 for all pairs. The live forward test was started on 29.03.2011.