Matthew, pls find enclosed trades from ITM and from OEC live account on Friday and related latency of the gen3 solution.
ITM timings
4:05:17 PM
9:30:01 AM
9:29:59 AM
9:23:30 AM
OEC live account timings
4:05:21 PM
9:30:06 AM
9:30:05 AM
9:23:33 AM
Delta in seconds
4
5
6
3
This average delay of 4.5 seconds is in line with reports of other users of gen3 OEC.
What is causing this delay if there is a direct server connection?
We all know very well that seconds in financial markets is a huge latency. Prop firms look at improving nano seconds via colocation.
I certainly do not expect nanoseconds but an average delay of 4.5 seconds seems just to high especially when we are charhed $3.8 per rt trade.
Again, $3.8 per rt trade on a daytrading system that trade 4-5 tmes per day adds up to $5000 a year.
This is a lot of money which I don’t have a problem spending for superior execution and low latency.
A latency of 4.5 secinds is poor in my view.
Could you please explain the reason of the delay and if you have plans to work on improving this?
Thanks a lot,
Manuel
The clocks on the servers are different, so I’m not sure you are comparing apples to apples.
But if the latency as you calculate it bothers you, then by all means stop using the software, and sit there and type in the trades yourself.
The clocks on the servers are different, so I’m not sure you are comparing apples to apples.
Regardless, the design of the AutoTrading protocol is such that latency from the C2 database to your broker is generally less than two seconds. But it’s not always. Personally, I think it’s foolish to make latency the sole criteria by which you judge the efficacy of AutoTrading.
Any trading method that depends on having orders placed within one second probably isn’t going to work well here. Before you obsess about C2-to-broker latency, let’s be realistic. The reality is that there is latency everywhere. The C2 Model – of having an open platform, onto which anyone, anywhere in the world, can pipe trading signals, and from which anyone, anywhere in the world, can take signals and pipe them to any broker, anywhere in the word – that model, while astoundingly powerful – probably isn’t built for certain types of trading.
If that bothers you, Manuel, then don’t use the software. Sit around your computer and type in the orders yourself manually. Limber up those fingers, though, before you start.
And further, we’ve actually checked the clock time stamps and your initial post is not correct.
C2 sent the signals to the OEC server at the following times:
4:05:21 (I’m puzzled that you think you saw the signal in your Instant Trade Messenger window at 4:05:17 … since the signal it wasn’t posted into the C2 database until 4:05:21)
9:30:02
9:30:02
9:23:31
So I think you have fallen victim to the fact that there are different clocks out there, and OEC’s may not be synced with C2’s. The point stands: signal latency is less than 2 seconds – in some cases less than 1 seond.
Finally, let me note that the Instant Trade Messenger (ITM) window uses a javascript seconds-counter to show the user the general time and to let the user know things are alive. It’s not meant as any sort of scientific timer, and it’s not really a good idea to use that javascript timer running locally on whatever PC you are using as some sort of arbitrator of latency measurements.
Matthew, thanks for your reply. I take your point that the clock times may not be precise enough hence it does not make sense to consider those for latency measurements.
My point was very simple: if we pay a good amount of money for a service that service should be best in class. There is not need to get defensive with statements such "If that bothers you, Manuel, then don’t use the software. Sit around your computer and type in the orders yourself manually. Limber up those fingers, though, before you start."
I am not questioning the versatility or convenience of the C2 model. The model works as proven by its success. I am just saying that C2 could be better if on top of great userfriedlyness and statistical information it strived to be best in class in autotrading execution.
Does C2 have today top notch autotrading execution? Probably not given latencies in the order of magnitude of seconds (both vendor to C2 and C2 to broker). Could C2 get better at this by constantly revising and updating its technlogy? For sure.
At the end the success or failure of C2 will only depend on its ability to generate profits for subscribers. In today´s environment where 50%+ of trading volume is generated by algorithms and where all professional firms co-locate theirs servers with the Exchanges the ability for an individual investor to make money in daytrading are less and less.
http://www.eurexchange.com/institutional/it/co_location_services_en.html
I hope you will aways try to make C2 autotrading better and faster as great information and convenience is not enough for the long term viability of the model (unless the business model is purely based on “trial” and not on “repeat”). Trading profits in real subscribers accounts and low slippage are also required for people to keep using C2 on a long term basis.
Cheers,
Manuel
scalping systems here on C2 that depend on being in and out in sub-minute times have never done well on C2.
In fact, scalping rarely works well for amateurs, given that you pay the same commissions, fees and other costs as people who hold for hours to days. Scalper usually delude themselves into thinking it is les risky, when in reality it is often unprofitable.
The concept of near flash trading from something like C2 will probably never exist. There are existing facilities for that. But trying to enter flash trades and expect subscribers to benefit is unrealistic.
For the normal C2 type trade, a second is of little importance, In fact, the market movement is such that it probably all averages out in the end.
Nobody talked about scalping. Many traders on C2 do not fully consider the effect of slippage. Even 1 tick per trade would diminish the average P&L per unit by about $20, depending on the tradable.
There are many systems on C2 with an average P&L per unit well below $40. A slippage of 1 tick would diminish profitability by mor than 50%.
The intent of my post is broader than just scalping and I think has relevance for all daytraders.
You are confusing flash trading and high frequency trading. They are not the same. Flash trading is in the process of being abolished, in any case.
We are talking precisely about scalping. And I did not talk at all about slippage. I talked about the fretting about not getting in lightning fast over C2. You started off convinced the delays were on the order of 5 seconds or so, which MK addresses as correct.
Worrying about a second on a lengthy trade has little relevance for position trades, as I already said. It averages out in the end. Only scalpers are concerned about subsecond execution, because the advantage they perceive can evaporate for scalping. And as I said, scalping has not proven to work on C2 anyway.
Slippage has nothing to do with the conversation.
Flash trading was a play on the concern Manuel has over his perception that C2 is off by multiple seconds. Flash trading is an advantage by those who see prices a tiny fraction before other traders. There was no confusion.
High frequency trading is not really the correct term anyway. Scalping is about LENGTH of time and size of profit sought more than frequency of execution. A person might only scalp a few times a day with large sums. They are looking to get in and out rapidly seeking the desired few ticks. For C2 it is a pipe dream. Many have tried, and none really have made a successful C2 system on this concept, having decent stats.
Flash trading is an advantage by those who see prices a tiny fraction before other traders.
This is why it was irrelevant to the conversation. Manuel didn’t introduce the idea of flash trading, you did. In any case, C2 isn’t capable of flash trading, just as retail traders aren’t capable of flash trading. And the SEC and exchanges are winding down their flash trading in any case.
High frequency trading is not really the correct term anyway. Scalping is about LENGTH of time and size of profit sought more than frequency of execution.
Believe it or not, HFT does refer to the length of time of the trade and not solely the frequency of execution. “Computer-driven, algorithm-based trading at speeds measured in millionths of a second,” as defined by the WSJ.
In the end, Matthew Klein concluded that C2 is not suitable for the kind of trading that Manuel is looking to do and that he should look elsewhere. That’s certainly clear enough.
so the entire point of your posts here, has been to show you have subscribed to Investopedia.com Word of the Day?
No, the entire point of my responses to you are to show:
1) You don’t know basic technical trading terms
2) You don’t know basic rules of risk management
3) You don’t know how to build or evaluate a successful system
4) You have more sock-puppet accounts than a classroom full of kindergartners
In short, to expose you for the fraud that you are.
funny, all of those descriptions seem above you. Seriously, you are number 10-20 in a long line of trading duffers who is going to champion the cause of big-boy wannabes everywhere and stand up to regular posters and try and prove his knowledge. You dropped the name High Frequency Trading and will keep responding until you hope the other person gives up. You did not grasp the humour in Flash Trading when used with Manuel and were try to quick to correct what was correct, and are trying to score brownie points by telling me what I already know.
1) You don’t know basic technical trading terms - Why don’t you give me a sensible discussion of all the greeks (without looking it up in your Idiot’s Guide to Trying to Look Clever on a trading forum)? BTW, I don’t particularly care, as I know you can find it on WSJ. But it will keep your cut and paste adventure going for a little while
2) You don’t know basic rules of risk management - why don’t you tell me about the YOUR basic rules of risk management? So far, all I see is someone who I am trying not to chuckle over.
3) You don’t know how to build or evaluate a successful system - why don’t you tell me about YOUR successful systems you have posted here?
4) You have more sock-puppet accounts than a classroom full of kindergartners - how many sock-puppet accounts did you have in kindergarten again?
In short, to expose you for the fraud that you are. - don’t dribble while you are speaking. Gee, wanna “put up our dukes?” Wanna compare lengths in the 8th grader’s boys room? Are you going to cry if you don’t win?
Future posts of similar paucity will only get an LOL. If I am in the mood. I will give you something to get eggsited aboute - mispelings so ewe kan draft annuther responds…
