Could anyone shade some experiences in autotrading futures systems with Open E Cry?
What kind of slippage you are experiencing?
Does C2 still charge $1.99 per side "autotrading fee"?
Are you happy with this solution overall?
Cheers,
[LINKSYSTEM_42777369]
I have had one issue where OEC locked my account for no reason and it took 3-days to get it back online. They (OEC) would have manually made trades for me, but in my case, I did not have any open positions and was willing to miss a few trades instead of trying to manually trade.
Slippage is minimal (0 ticks mostly) from C2 using market orders.
I started not using autosync and eventually turned it on b/c of a C2 trade issue. Depending on the trade frequency (real fast off, slower on) of the system you are trading, you may elect to keep it off.
Yes to the $1.99 per side. I don’t like it, but what can you do? It is not a big cost concern for system that aims to capture bigger moves (multi points), it significantly hurts 0.25-1.0 point scalping systems.
Overall, I am happy, but would like to see a price reduction from the $1.99 to $0.50-$1.00.
Jim
I’ve been using it for a while. I have no complaints overall. There are some minor issues. Sometimes the orders do take a while to “arrive”. Sometimes stops/limits seem to be canceled for no reason, and then re-placed after a while (without any signal from the vendor), or they don’t show up at all, and positions are closed by sync. Sometimes OEC Trader pops up errors with the orders. But in general, the results from autotrading match the C2 account results pretty well.
I don’t trade fast-moving, scalping kind of systems, so I don’t know how it would perform in that case, or how much the autotrading costs would impact the results. In the case of systems that place a couple of trades a week, the electricity cost of leaving the computer on 24/5 would probably be more than what C2 charges, let alone the inconvenience and possibility of failures.
Matthew, could you please comment on some of the issues with gen3 outlined in the posts above?
Thanks a lot.
Without knowing the exact details of what he is referring to, I can’t comment on the specifics of ironcito’s post. I can only say that there are cases in which signals need to be held on the server side, and not posted to the broker immediately, for a variety of reasons (time of day, other pending orders).
There are also cases in which exchanges accept only day orders, not GTC orders, and we shield the system developer (and subscriber) from these exchange-by-exchange requirements and intricacies, and instead take care of them automatically on the server side. This might require canceling and replacing orders as necessary, which I suppose could seem mysterious if it is not explained.
In any case, the main point is that Gen3 AutoTrade at this point is rock-solid, fast, and has a variety of synchronization features (both on the position and signal level) working behind-the-scenes which improve fidelity to the C2 system. It’s pretty neat.
Thanks ironcito for your reply.
Could you help understand the full commission structure for OEC?
Let’s assume S&P mini.
Commission: from OEC website there is a special offer of $0.5 per side. So $1 round turn.
Exchange fees: $2.28 round turn
Autotrading fees: $1.99 per side. So $3.98 round turn.
NFA fees: $0.01
So total round turn = 1+2.28+3.98+0.01 = $7.27
Does this match what you are charged?
Forgot to mention. OEC website quotes
"OEC API Rates. A direct connection fee of $0.15 per side will be applied to all commission rates (listed above) for trades sent directly through the OEC API not utilizing the OEC Trader Software."
Is this 0.15 included in the $1.99 fee or is it on top?
Matthew,
Sorry but I need to do some due diligence before I decide to switch from Tradebullet/TWS to OEC gen3.
Matthew, could you or anyone comment on what "fast" means?
From this post by James it seems it takes 10-30 seconds?
http://preview.collective2.com/cgi-perl/board.mpl?term=oec&want=listmsgs&boardid=9323885&msghilite=39018628&threadhilite=6506&searchstartat=6506#msgid39018628
Is this real? 10-30 seconds is a huge amount of time in futures. Why would that happen?
Thanks!
[LINKSYSTEM_42777369]
Yes, a “fast” system on C2 is making a round trip trade in minutes rather than hours. In this case, I don’t think autosync is a good idea. If the trades are once a day or twice a day and held for 30min or more, autosync should be considered, as it will ensure that you (the trader) are not holding an rouge open position. This happened to me and I had to manually close the position a day or two later, once I saw it.
Jim
Here’s what AutoSync does:
+ A vendor sends a signal. But the OEC server goes down temporarily, and misses the signal. Autosync fixes it.
+ An exchange rejects one of your orders because it was submitted outside of a strict price range for limits or stops. AutoSync fixes it.
+ Mid-stream in a trade, you change your scaling preferences because you want to trade less rather than more. AutoSync reduces your previously-open position.
+ God forbid, a C2 data center goes down during a trade when a closing signal is generated. When the backup data center comes online, AutoSync corrects it.
+ You specify in your AutoTrade setup that you only want to trade in your real account 10% of whatever the C2 system trades. But the system vendor places a series of weird, small 1-unit trades. Because you can never trade less than 1 unit of a futures contract, you find your account trading 15% of the C2 account’s equity, not 10%. AutoSync notices the problem and corrects it.
In short, AutoSync is a technology that constantly compares your actual real-world positions with what you should have in your account (based on C2 systems) and makes appropriate adjustments.
When might you not want to use it? The only time you would consider turning it off is when a system tries to go in and out for small profits – typically less than twice the bid/ask spread of the instrument. In these cases, if you miss a just-touched limit order, and AutoSync corrects the problem, you may be dinged for a tick or two of slippage.
Matthew, could you please comment on latency between C2 servers and OEC servers? I asked OEC but they told me to ask you.
I would like to know:
1. Latency in ms
2. Why according to user experiences (see James Nelson) a trade takes between 1-10 seconds from showing up on ITM to OEC account
3. Expected slippage (in ticks) for ES during market hours.
Thanks!
I do not think the forum message you referenced is currently accurate. While obviously conditions vary, and all sorts of stuff might happen, a 10-second delay between signal posting and being traded would be unusual and large. I would expect more along the lines of about one second.
I cannot comment on expected slippage, since it depends on order type, whether you are filled at the limit or stop, whether the price is market clearing, etc.
I can tell you that the prices you see on the C2 hypothetical track record, while of course hypothetical since they do not represent any single account’s performance, are adjusted to use real-life brokerage reports. In other words, if a system is being autotraded at OEC, and the average fill price in a real-life OEC account is 1234.50, then C2 will automatically change the price of fill in the C2 track record to equal the VWAP of those actual real-life fills. This means that for any system that is being autotraded, the C2 hypothetical track record provides a reasonable indication about potential slippage. Of course I must stress once again the important warning that C2 Model Account results must be regarded as hypothetical.
Today, ITM order for 1 ES at 10:09:07, OEC fill at 10:09:12 = 5sec delay between C2 ITM and OEC order fill.
1-10sec is typical.
Jim
We looked at the server audit trail for the trade you describe.
Signal id 43592423 was issued by the system developer at 10:09:08 Eastern U.S., and was submitted to your OEC account at 10:09:09.
OEC reported it as filled at 10:09:10.
So the difference between the signal-generation and the moment your broker received the order was 1 second, not 5 seconds.
Hi Matthew, I started testing my own system on OEC (demo) gen3: account DEMO150527.
From the gen3 Performance Report yesterday’s P&L was +$154.
However, the C2 hypothetical results (recently closed trades) for yesterday are -$1583.
There are also two instances on the Performance Report of closing 2 contracts whereas the closing signal was only 1 contract as correctly represented in C2 recently closed trades.
43588439 2009-10-06 13:09:55 QRBX9 Sell to Close 2
43591426 2009-10-06 13:53:53 @EMDZ9 Buy to Close 2
Could you please explain the reason for the dramatic difference in P&L between Performance Report and C2 recently closed trades as well as the issue with the two signals above (should have been 1 contract, not 2).
Thanks!
[LINKSYSTEM_42777369]
Hi, Financial:
C2’s “Performance Report” feature is relatively new. The software is nifty; it attempts to do the following: take execution reports from your broker, compare them to C2 system signals, and piece together which execution reports from your brokers correspond to which C2 signals, and then finally pieces together a system-by-system P/L report based on reported brokerage fills.
In most cases it works pretty well, but not always. There are a lot of complexities that aren’t immediately obvious. Like, for instance, broker reports aren’t always delivered in a way that makes sense to C2. To take one example (which you point out), an order to reverse a long position is represented at C2 like this:
Sell to Close 1
Sell to Open 1
But this is reported to us by the broker as Sell 2, which - unless it is properly handled - can throw the P/L reporting into a tizzy.
We should handle this gracefully, but I examined your C2-generated Performance Report and agree that the P/L calcs are not correct, for reasons such as these. Someone here will work on this, perhaps as early as today.
The good news is, if you look at your broker’s execution reports, and compare them to the C2 system you are autotrading, you’ll see your actual broker trades correspond to your C2 system almost perfectly. You’ll also notice that your OEC fill prices and C2’s reported fill prices are nearly identical. (They won’t be completely identical, because you are trading with a demo account at OEC, not a live account, and thus C2 does not automatically use your broker fill prices to adjust its own hypothetical prices, which is what we do in cases of live broker accounts).
Thanks for pointing out the problem with the Performance Reports page P/L calcs. We’ll try to look at it shortly.
Thanks for checking. I pulled my data from the ITM page on C2, that indicated 10:09:07 and my OEC account that indicated a fill at 10:09:11.8, but unless everyone’s clock is sync’ed to within 1sec, I am sure it is all the same. Fill time, may be 1-2sec from order receipt. 1-10sec is acceptable in my opinion (looking at C2 to OEC fill), the order process is likely faster (as you indicate).
Let me point out that yesterday’s trade was perfect, no problems. I guess the real question here is how often is it perfect and how often not so perfect? When it fails, what are the likely outcomes/delays? 1sec 95% of the time, 2 sec 3.5% of the time, etc.
Jim
Matthew, while you’re at it, my performance report is also wrong. I think the mistake is in a trade from [LINKSYSTEM_29987971] which shows a $19,100 profit in a single trade (I wish!). The trade was for @SMF9, closed on 2008.12.24. It has no signal ID next to it, but I think the opening signal ID was 37217410. If you could check into it, it’d be great. I’m not trading that system anymore (it’s dead actually), but it would be nice to see my grand Gen3 total correctly. Thanks in advance!
Hi Matthew, today another problem with OEC demo autotrading.
My system only sends 1 contract not 2. So the first trade is not correct (should have been 1, not 2), the second trade on the other hand was never sent. The third must be autosynch.
35948000 Buy 2 GRBX9 1.7431 4433242 9:10:54
35948006 Buy 1 GRBX9 1.7431 4433273 9:11:04
35948199 Sell 3 GRBX9 1.7316 4433777 9:15:52
Major issue here for today:
P&L C2 historical hypo trades = -1702
P&L OEC demo gen3 = -2650
P&L TWS-TB gen1 = -1667
So gen1 worked perfectly and with a P&L for the day in line with C2 hypo fills. Gen3 on the other hand failed today and lost $950 more the C2 hypo.
What is causing this?
[LINKSYSTEM_42777369]
Okay, we spent a lot of time tracking this one down.
Here’s what we found: This was a problem with the OpenECry simulated demo account. We placed several market orders into your sim account, but OEC’s simulator did not acknowledge them as filled for several minutes (which of course would not happen in a live account). This threw off quantity calculations.
In the 4+ years we’ve been doing this, we have never seen this issue in live accounts. We have seen it a few time in demo accounts, but even those incidents are pretty rare.
I can assure you this issue is limited only to the market simulation used by OEC.
