Edit Your Comment
Chart of Latency Times to Most Forex Brokerages
会员从Aug 22, 2012开始
11帖子
Aug 24, 2012 at 21:55
会员从Aug 22, 2012开始
11帖子
It is no wonder it is hard to get fills at some brokers when markets are moving fast.
I stumbled onto this chart which reveals in quantifiable time how long the latency is from San Diego, New York City or the UK to nearly all major Forex brokerages. It ranges from <1ms to over 200ms which is huge when you realize it is double that because the data must move both directions. If it takes over 200ms for you to get the quote and 200ms to send the order the price has likely moved dramatically in a fast market.
http://helpdesk.commercialnetworkservices.net/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=174
I stumbled onto this chart which reveals in quantifiable time how long the latency is from San Diego, New York City or the UK to nearly all major Forex brokerages. It ranges from <1ms to over 200ms which is huge when you realize it is double that because the data must move both directions. If it takes over 200ms for you to get the quote and 200ms to send the order the price has likely moved dramatically in a fast market.
http://helpdesk.commercialnetworkservices.net/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=174
Where there's odds . . . there's hope.
会员从Jan 14, 2010开始
541帖子
Aug 25, 2012 at 08:04
会员从Jan 14, 2010开始
541帖子
nevercaughtup posted:
It is no wonder it is hard to get fills at some brokers when markets are moving fast.
I stumbled onto this chart which reveals in quantifiable time how long the latency is from San Diego, New York City or the UK to nearly all major Forex brokerages. It ranges from <1ms to over 200ms which is huge when you realize it is double that because the data must move both directions. If it takes over 200ms for you to get the quote and 200ms to send the order the price has likely moved dramatically in a fast market.
http://helpdesk.commercialnetworkservices.net/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=174
Marvin, this chart is valid only if you set you vps with cns, however it has no value if your vps is elsewhere. Moreover, the chart doesn't include execution time - the latency could be super low, but execution time super high, so the total of both would be probably the best metric for comparison (not to mention the execution time must be measured for a real account, not demo).

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 25, 2012 at 08:38
会员从Jun 06, 2012开始
1434帖子
James
there must be ref point. He cant measure from our machines. So instead of getting values from Marvin s house it is beter to get values from a VPS with a know IP ( Geo Location ).
walker
there must be ref point. He cant measure from our machines. So instead of getting values from Marvin s house it is beter to get values from a VPS with a know IP ( Geo Location ).
walker
James_Bond posted:nevercaughtup posted:
It is no wonder it is hard to get fills at some brokers when markets are moving fast.
I stumbled onto this chart which reveals in quantifiable time how long the latency is from San Diego, New York City or the UK to nearly all major Forex brokerages. It ranges from <1ms to over 200ms which is huge when you realize it is double that because the data must move both directions. If it takes over 200ms for you to get the quote and 200ms to send the order the price has likely moved dramatically in a fast market.
http://helpdesk.commercialnetworkservices.net/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=174
Marvin, this chart is valid only if you set you vps with cns, however it has no value if your vps is elsewhere. Moreover, the chart doesn't include execution time - the latency could be super low, but execution time super high, so the total of both would be probably the best metric for comparison (not to mention the execution time must be measured for a real account, not demo).

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 25, 2012 at 09:58
(已编辑 Aug 25, 2012 at 10:15)
会员从Jun 06, 2012开始
1434帖子
"If it takes over 200ms for you to get the quote and 200ms to send the order "
when latency is 200ms it is round trip.
so it takes 100 ms for you to get quote and anoyther 100ms for sending the order to server.
means your delay will be latency/2 + latency/2
delay = 1/2 latency + 1/2 latency
means
delay = latency
walker
PS:
NOT CORRECT
delay = 2 x latency
when latency is 200ms it is round trip.
so it takes 100 ms for you to get quote and anoyther 100ms for sending the order to server.
means your delay will be latency/2 + latency/2
delay = 1/2 latency + 1/2 latency
means
delay = latency
walker
PS:
NOT CORRECT
delay = 2 x latency

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 25, 2012 at 10:03
会员从Jun 06, 2012开始
1434帖子
also MT4 has an autentication problem for every 30 sec.
when there is no trade or modification activity with broker server after 30 sec MT4 switch to sleep mode. when any order send to broker after 30 sec (when in sleep mode) MT4 first ask for autentication which it takes around 500ms.
walker
when there is no trade or modification activity with broker server after 30 sec MT4 switch to sleep mode. when any order send to broker after 30 sec (when in sleep mode) MT4 first ask for autentication which it takes around 500ms.
walker

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 25, 2012 at 18:17
会员从Jan 14, 2010开始
2279帖子
Try Oanda. execution is a way faster than MT4 and practically no slippage. Position sizing is also head and shoulders above any MT4 broker not to mention that Oanda has got its own MT4 platform now. It is much easier to manage trades opened with Oanda.

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 25, 2012 at 18:35
会员从Jun 06, 2012开始
1434帖子
is it java based or desktop platform?
Chikot posted:
Try Oanda. execution is a way faster than MT4 and practically no slippage. Position sizing is also head and shoulders above any MT4 broker not to mention that Oanda has got its own MT4 platform now. It is much easier to manage trades opened with Oanda.
会员从Aug 22, 2012开始
11帖子
Aug 26, 2012 at 07:57
会员从Aug 22, 2012开始
11帖子
Steve, Your points are very valid and I think you see it the same way I do except for the part about the ticks. No doubt a tick is about half a second, but ticks do not fill orders. Liquidity has to be there to fill the order with liquidity providers offering in advance X amount of liquidity at current price, x amount at the next level, x amount at the next level and on down. So as soon as one level is gone it moves to the next and then the next as so on.
I think of it as 10 shelves of the same product in a store and when the top shelf is empty the price is higher on the next shelf down and each shelf downward raises the price again. So in order to get the product on shelf one you have to get to that shelf before the next person trying to get to it.
Speed matters in fast moving markets so if a strategy isn't dependent of getting product off of shelf one it is not a big concern. For a strategy that needs top shelf product it is a big concern.
I think of it like this. If I need to run to the store and get some of the product on shelf one, my chances of getting it are much better if the store is only a few blocks away rather than a several miles away. I hail the taxicab (wake up MT4), tell the driver where to go (latency) and run into the store and buy it (execution).
Execution can also be improved by dealing with brokers with redundant gateways (doors to get into the store), but that is a whole different topic.
I think of it as 10 shelves of the same product in a store and when the top shelf is empty the price is higher on the next shelf down and each shelf downward raises the price again. So in order to get the product on shelf one you have to get to that shelf before the next person trying to get to it.
Speed matters in fast moving markets so if a strategy isn't dependent of getting product off of shelf one it is not a big concern. For a strategy that needs top shelf product it is a big concern.
I think of it like this. If I need to run to the store and get some of the product on shelf one, my chances of getting it are much better if the store is only a few blocks away rather than a several miles away. I hail the taxicab (wake up MT4), tell the driver where to go (latency) and run into the store and buy it (execution).
Execution can also be improved by dealing with brokers with redundant gateways (doors to get into the store), but that is a whole different topic.
Where there's odds . . . there's hope.

forex_trader_79941
会员从Jun 06, 2012开始
1434帖子
Aug 26, 2012 at 08:27
会员从Jun 06, 2012开始
1434帖子
Hi Marvin
I did not say VPS is useless.
- no power off
- no connection problem ( %99 of the time )
- auto Backup
- 7/24 up and running
- low latency
- no maintanance
agree;"Speed matters in fast moving markets so if a strategy isn't dependent of getting product off of shelf one it is not a big concern. For a strategy that needs top shelf product it is a big concern."
I am also VPS user.
walker
PS: trader can reach better liquidity by being fast or slow. it depends the market condition😉
I did not say VPS is useless.
- no power off
- no connection problem ( %99 of the time )
- auto Backup
- 7/24 up and running
- low latency
- no maintanance
agree;"Speed matters in fast moving markets so if a strategy isn't dependent of getting product off of shelf one it is not a big concern. For a strategy that needs top shelf product it is a big concern."
I am also VPS user.
walker
PS: trader can reach better liquidity by being fast or slow. it depends the market condition😉
Aug 26, 2012 at 10:35
会员从Jan 14, 2010开始
2279帖子
At least for live account it is desktop platform but java based browser also works ok. my live account has desktop p[latform. no issues.
stevewalker posted:
is it java based or desktop platform?Chikot posted:
Try Oanda. execution is a way faster than MT4 and practically no slippage. Position sizing is also head and shoulders above any MT4 broker not to mention that Oanda has got its own MT4 platform now. It is much easier to manage trades opened with Oanda.

*商业用途和垃圾邮件将不被容忍,并可能导致账户终止。
提示:发布图片/YouTube网址会自动嵌入到您的帖子中!
提示:键入@符号,自动完成参与此讨论的用户名。