Ceasar2 signal

[ScottFX]

I have been running two Caesar2 signal subscriptions to two ICM accounts on the same server and the one account always outperforms the other.

[timo]

In theory that is absolutely correct chris ...

But in practice there seems to be another factor which is impacting results. An unknown factor at the moment.

I've been running from the UK - Manchester VPS to ICM for a while now, and consistently getting much worse than the master, and worse than others on ICM and NY.

So there is possibly a gremlin somewhere ... snacking on pips ...

[chris0212]

Per my understand and confirmed with simpletrader.net, the work flow is start from signal server of simpletrader.net in UK-Manchester --> VPS Server ---Simpletrader.net's server---> broker server. Then, for each trade the simpletrader.net's server will send a updating signal to monitor the status. Therefore, if the vps for IC market is located in NY, the signal will send from UK-Manchester ---> NY VPS---> UK Mechester ---> IC's server in NY. Since the signal need pass through the Atlantic Ocean two times, the ping at least delay 500ms. In short, the most fastest vps would be located at UK-Manchester near the simpletrader.net even though the broker server in NY.

[timo]

Thats a very respectable result littlemax

3-4 pips behind the master is very good. I'd be extremely happy with that.

On this basket 3 pips would represent about 10% less than the master. Whereas I was around 22% less.

I think once you starting getting over 10% difference from the master it begins to affect the stats / performance quite heavily.

Thanks for your details ... interesting that you are on NY VPS also ...

[littlemax]

My account - - ICM on CNS NY server 30.2 pips (MT4) - interesting comparison with your ST VPS on ICM, might have been lucky this time, although usually 3-4 pips behind master, sometimes more on a fast moving closing candle.

[timo]

Your test accounts did pretty well however.

The MT4 almost matched the master, and cTrader not far off.

That was unexpected, as the MT4 is ICM + STVPS UK as well right?

Me (MT4 ICM + STVPS UK) - 25.8 pips
cTrader - 29.8 pips
MT4 - 32.40 pips
C2 Master - 33.1 pips

[timo]

How was that recent basket for you Will?

My 4 trades all said >5 secs closure time. With the following slippage:

-2.1
-2.9
-1.8
-0.6

ICM + STVPS UK

This is reasonably indicative of most baskets for me. Except this basket grabbed a few more pips profit which was good, but usually the total pip count is less.

p.s. thanks for investigating this in detail btw. I know you're busy, and we appreciate it

[WillT]

Hey guys - we are still waiting for a big closure to see something. The closes today executed in a nice 1 second or so.

[ralf]

Hello,

the trading of the second basket last Friday (08/15) was far from being executed as desired. Here are some numbers:

- Master account closed the 7 trades of the basket from 17:41:40 to 17:41:43
- My account closed these trades from 17:43:42 to 17:53:44.

This is a delay between 01:59 and 12:01 (appr. 2 - 12 minutes). I am wondering if this delay can be explained with the lack of liquidity because of the flood of orders - I have my doubts.

Question is where the delays are coming from? In the list below I inserted all time stamps I could collect for the one trade with the biggest delay in the basket. The statements with the question marks are in the hand of SimpleTrader :]

MT4 trade ID: 3940466

Master account closed trade: 17:41:40
Trade signal reached ST server in UK: ?
ST server passed trade to copier EA: ?
Trade reached terminal, terminal sent request to broker: ?
Broker executed the trade and sent confirmation back to terminal: 17:53:43.917 - 17:53:44.347
Terminal reports back to ST server: ?

I really would like to know what the log files of ST report regarding the missing time stamps.

Regards
Ralf

[VictorTous]

Thanks for looking into it Will!!

[WillT]

And another post

An example of cTrader closing out a trade recieved from the EA.

ctraderclosing.png

As you can see - The EA knows to close the trade at 22:25:06.533 - all the calculations and checks are done within the same millisecond - but it takes a further 400ms to actually fully close the trade.

This is in an ideal environment on a demo cTrader I'm using for development.

[WillT]

So just to keep going on this a little - The SLOWEST time in a basket - would actually be the entire time it took to close the whole basket.

slowest.png

So as you can see from this image - the slowest closure was 4.32 seconds - however this was the first trade to close in the basket and as such will be the one that was reported to our system last. Hence the entire basket took approx 4.32 seconds to close.

Now when we are talking 7 trades in an active market, 4.32 seconds to close the basket of 7 trades from start to finish isn't bad going. The EA has all the information to close all 7 trades, and on a normal execution of a trade we would probably say a broker takes approximately 0.5-1second to close a trade from transmission to closure:

- Trade sent to NY IC Markets trade server
- Trade sent to "integral"
- Liqudity is found
- Trade is closed

[WillT]

One thing we need to bare in mind here which is seriously screwing the stats. And I don't think I've said anything about this as I only just thought of it now.

The cycle in which the EA's for both MT4 and cTrader would go in would be:

- 13 Trades to close
- EA loops through each of the 13 trades and closes each one
- EA THEN tells our servers they have closed.

So in this instance the reporting is a little screwed because it doesn't get told about the closure of each of hte trades until the very end of the final closure.

This works perfectly with single trades, but as soon as you go multiple baskets, you head into issues where the reports are going to show slower closures than were happening in real time.

The reason we do this is because the reports are the least important aspect in terms of copying trades. Yes its good we can look over the reports, but we all want the trades to close out as fast as possible.

I'm sorry I hadn't picked up on this before but I was just going through things step by step again. We've had a chat with IC Markets and we are sure that during those last trades there were serious market movements which caused general slippage.

I've changed the way in which the cTrader EA tells me information - so upon the next basket closure, I'll be able to tell if its closing fast in the market or not.

[SpiderX]

Hi Nick,

Any chance you could have another MT4 platform hosted in NY just for comparison purposes ?

Cheers

[WhiteApe]

The real issue is that brokers can't close thousands of trades at once during high volatility, however I'm expecting an improvement considering cTrader is regarded as having much faster execution than MT4.

I suspect you are right about that Nick, although its good to test it out empirically - thanks to you & Will for doing that. (Round 1 to MT4 lol).

It does raise the issue that some of the few successful signals will reach a point of diminishing returns, which is a real shame. Lets hope the competitive nature of the market means the brokers will continue to improve their capacity to deal with high volumes of orders at once.