I have never heard of this and it does not make sense to me technically. If you have any references to documentation of this phenomenon (rather than anecdotal evidence) I would appreciate it.
Please accept my apologies regarding my latency comments to IB. I tried to verify if they had access points in Europe and didn't find any info. I was probably looking in the wrong places. I still stand by my comments regarding the speed through IB vs TT in the US. These tests were done several times over several days and are true for CBOT and CME data. TT's results were well ahead of IB as mentioned. My X-Trader is connected to a TT gateway and then directly to the corresponding exchange. The data went directly from my computer on the clearer's network to the exchange versus IB's data running over the internet. Average latency to IB was 26 ms which was deducted from the results for a more equitable comparison (private network vs. internet feed). The test with CQG was done head to head because they both run over the internet. I have no interest in promoting CQG or TT over IB. I'm an end-user trading an average of 1500 roundturns a day scalping, so my loyalty lies with whomever provides the best service at the best (not necessarily lowest) cost. Best wishes.
Screen freezes signal a problem with that particular setup. Shared memory has the disadvantage of tying up much needed RAM and CPU cycles but it shouldn't cause freezes. You might want to check the driver or OS for problems.
I tried a scalping strategy using ZeroLineTrader as the simulator, I too use IB, I was going for 0.25 to 0.5 points on ES mini, I discovered the strategy didn't work as well as I though it would, but do the comments on this thread imply that if I had somehow a top notch setup, using some different front end, then the above scalping technique might have worked, because then I would have got my limit target at Bid or Ask? In other words some delay meant my software(IB, front-end, whatever) missed my limit price at bid or ask, so in fact having a fast setup gets you prices other users do not receive ? I don't understand that, I thought IB execution speed was very good and once the order is placed it is executed as fast as possible, what possible difference does the hardware/software make once the order is on IB's servers? Sorry if this is a ridiculous question, not trying to imply good hardware/software is not important, obviously you need a stable connetion, and a robust and consistent software, just trying to understand if I should look into scalping again ? Thanks
Probably not. A strategy will probably not work because of the strategy itself or your execution of it. The difference between IB and others will only show up if you trade frequently enough and you are in a fast market. There are many variables here. My comments aren't an attempt to discredit IB. Your hardware is probably sufficient unless your CPU is running at 90+% and your Available Physical Memory is below 80MB or so. In that case, your software might be overwhelming your system enough that it causes a delay in executing your mouse click orders. For this, you really need to be running a lot on your machine. Go to the Task Manager and click on the Performance tab and look at the info there.
FT71, seems like everyone else in this thread is confusing "latency" ie: the ping from their pc to broker server, with the update speed of orders as discussed in your comments. Perhaps you could elaborate on the difference to clarify if you get a moment
Latency is simply the roundtrip time in milliseconds from your computer to the specific server on the internet where the broker handles the orders. This is measured through though a ping test or a tracert command from Windows. The update interval is something entirely different. This is the time between price updates provided by your broker. My comparison across IB, TT and CQG was to measure the difference between one system's update for the next tick up/down versus another provider. It also measured the difference in the depth of market for all 10 price levels on the CME and 20 price levels on the CBOT. I hope this clarifies what I was referring to. Again, none of this matters unless you trade an awful lot on a product that is quick and not as liquid as the ES, ZN, etc.
The issue with Internet routing, is that geographical proximity doesn't necessarily equate network proximity. E.g. in my case, although I'm in Europe, my ISP routes all international traffic via Alternet (a big ISP in US), so my ping times to IB gw in Europe (CH) are between 80ms and 250ms (depending on utilisation), avg about 125ms.