Autotrade 2 fx signallers?

"Why would you want to? In order to do the same thing discussed in my scenarios above, and allow the broker to automatically handle the closing trades (Limit and Stop-Loss, one-cancels-other). "



A bit of further clarification: If trading two systems, the trades for those systems would NOT be entered on an OCO basis. They would both be entered, and stand alone.



Look at Jules’ sequence of trades again.



Hans.

Dan - This is not how Gen 2 auto-trade works. It is not an OptionsXpress limitation, as C2 auto-trade never attempts to do that.



On another note, a big problem I’m having with C2 auto-trade as applied to OptionsXpress is the fact that all limit orders are converted into market orders when sent to OX. Try buying or selling 100 YM contracts during the night market that way and slippage looks so bad you’ll think you’re trading banana peels. In addition, trade signals experience enough of a delay in reaching OX that it is impractical attempting to trade fast-moving systems (those which round-turn in a manner of minutes).

Lew:



Now you have me curious. I Gen2 autotrade with Open E Cry. Limit orders are sent to OEC as limits, but if they trigger on C2, then C2 tries to do a cancel/replace to market. The XRPL should (and usually does) get rejected in favor of the limit. As for delays, I find the latency between order entry on C2 to order entry on OEC to be between 2 to 5 seconds. Could this be an OptionsXpress issue?



Dan:



I thought of another example of potentially confusing multiple system integration. Suppose System A buys corn at market with the price at $4.00 with a protective sell stop at $3.80. System B has a different take on market action and goes short corn at $3.95 using a stop. You are long with System A. Price falls, triggering the System B short entry at $3.95. Your account is now FLAT. Prices fall, triggering the System A sell stop at $3.80. You are now short. System A is out of the picture entirely with a 20 cent loss, and your corn positions are now being managed by System B. System B, with corn at $3.80, has a 15 cent gain. Your account shows a closed trade with a 5 cent loss, and an open trade currently sitting BE. If traded in two separate accounts, you would show a 20 cent closed trade loss in one and a 15 cent open trade gain in the other. Traded in one account you show a net 5 cent loss altogether.



Hans.

Hans - By design, all C2 auto-trade limit/stop orders are “held” by the C2 system until it sees the limit/stop price touched (within its data feed), at which point the order is sent to OptionsXpress as a market order. There are no exceptions to this rule (for C2 auto-trade into OX). I am seeing an average of one minute delay between the vendor’s order appearing on C2 and it actually being transmitted to OX. Since I also use the OX API for my own applications, I can tell the delay is not with OX.

Lew, Matthew denied this. In the VN Forex forum thread “Important Announcement!” we had a brief discussion about stop hunting, and I suggested that this would not be possible if the stops are held at C2 and transmitted as market orders when the price is hit. Matthew responded:



“Jules: Bulldog sends stops into the customer’s account beforehand. It does not “wait” for C2 to trigger a stop, or for some price to be hit.”

OptionXPress’ (OX) API has a very severe limitation: you cannot place simultaneous stop and limit orders on the same stock. You might say, “How could a broker impose such a bizarre limitation? Doesn’t that make trading, and automation, difficult?” Yes it does, and I have no idea who determined such a limitation would be a good idea. It is one of the strangest things I have ever seen… but, well, C2 just has to work around it.



So, in consequence, C2 handles OX differently than other brokers. It essentially watches the quote feeds for each trade it needs to place. When a bid/ask is above the trigger price, it sends orders to market.



I love Lew, but I think he is smoking crack when he says there is a “one minute delay” between a vendor’s order appearing and being transmitted to OX. One minute? Now way. Surely you must have meant to write one second. That might be a more reasonable number.



I personally trade a fast-moving systems at OX, using a fairly large amount of money, and have practically zero slippage. The problem doesn’t occur specifically when a system is “fast” (i.e. goes in and out quickly, trying to capture a small price movement); the problem occurs when the instrument being traded is illiquid.



But here’s the thing that you need to keep in mind. If the instrument is illiquid, you will have the same problem at all brokers. Remember how AutoTrading works: if the vendor tries to enter a scalping order, and tries to get filled on a just-touched limit order (i.e. the price bounces down to the limit price but then quickly bounces back up), then one lucky trader might get the fill – but everyone else will convert to a market order, anyway, in order to stay in sync with the vendor.



So, effectively, for 99.99% of traders, they will be submitting market orders, too.



In summary:



1) No way is there a one-minute latency, Lew. Maybe a few seconds, but not a minute. If you have evidence otherwise, please tell me the specifics and I will look into it.



2) The OX API limitation does not create any additional slippage for highly-liquid instruments (stocks with at least modest volume, futures with good volume during day-time hours).



3) For scalping systems, in which vendors try to enter or exit on just-touched limit orders, you will effectively be converted to a market order at all brokers, not just OptionsXpress.

>I love Lew,



um…



>but I think he is smoking crack



Amen. LOL.

Jules - It would help if you would read carefully and take a moment to digest what is said, prior to responding. My post was succinct and clear. I am not talking about bulldog. I clearly stated my comment applied to C2 Gen-II trading via OptionsXpress. Please try to stay on-topic.

1) No way is there a one-minute latency, Lew. Maybe a few seconds, but not a minute. If you have evidence otherwise, please tell me the specifics and I will look into it.



In one of my “debugging” emails to you regarding OX interaction, I did make mention of this… I stated I’d see a trade execute on C2, yet not be sent to OX for up to a minute (It’s easy to see pending trades on OX). I’ve seen this time and time again now, to the tune of tens of thousands of dollars… and if you wanted further (or more concise) proof of this all you needed to do was email me, rather than suggest I’m smoking crack.



2) The OX API limitation does not create any additional slippage for highly-liquid instruments (stocks with at least modest volume, futures with good volume during day-time hours).



Again, I’d be happy to send you proof that it does… on the ES and YM instruments during the day market, which I think you’d have to agree is a fairly liquid market.



3) For scalping systems, in which vendors try to enter or exit on just-touched limit orders, you will effectively be converted to a market order at all brokers, not just OptionsXpress.



Correct… but in the case of OX, you always start off as a market order. Combined with the inherent delay of order transmission (C2 to OX), this creates a situation in which slippage prevails. Again, I’d be happy to send you concise proof… if it meant you’d reconsider engineering the OX auto-trade engine (I emailed you a proposed outline which uses limit orders a week ago).

Let’s take the conversation offline. In a private email I’ll ask to see your concise evidence of one-minute latency. It’s really not possible, unless there was some kind of server downtime. I’ll send you an email in a moment. - MK

Lew: Actually the thread was about FXCM at first, so I thought you were referring to OX as an example for all. But when I read it again I see that this is not the case.