Hey Dom i had the same connectivity problem sunday night and monday morning and had to restart - it happened the past two sunday nights. Maybe you should post an issue on ninjatrader forum. Did you get it resolved?
CL always-in
- Thread starter dom993
- Start date
We (that is, my VPS provider, really) are working this with Ninja support, and the network providers identified so far has having issues.
Let me know what they tell you.
Results for the week ending February, 21th:
- 2 losses ; 1 BE ; net -985
Here is the status email just received from my VPS provider re. the connectivity issues:
"During the last 8 to 10 days, we've been alerted about data feed or brokers disconnection on our servers, in Chicago and in NJ. The problem was mainly visible on CQG, but also for Interactive Brokers and DTN IQFeed.
After analysis by our network providers, who are not the same in NJ and in Chicago, they found that the root cause was a congestion on Verizon and Qwest networks. These companies are providing the big network pipes which connects Internet providers in the US and all over the World.
Our network providers are using FCP (Flow Control Platform) equipments, which smartly prioritize the best routes, depending on latency, error rates, bandwidth, ..., and these equipments are now avoiding these busy network links and these two providers.
I'll keep monitoring these aspects until further notice, and will add other data feed providers, but it seems the problem is solved now."
Even though this congestion stuff looks a little bit iffy at first, given the Olympic Games & related amount of streaming, it might actually be the real issue.
- 2 losses ; 1 BE ; net -985
Here is the status email just received from my VPS provider re. the connectivity issues:
"During the last 8 to 10 days, we've been alerted about data feed or brokers disconnection on our servers, in Chicago and in NJ. The problem was mainly visible on CQG, but also for Interactive Brokers and DTN IQFeed.
After analysis by our network providers, who are not the same in NJ and in Chicago, they found that the root cause was a congestion on Verizon and Qwest networks. These companies are providing the big network pipes which connects Internet providers in the US and all over the World.
Our network providers are using FCP (Flow Control Platform) equipments, which smartly prioritize the best routes, depending on latency, error rates, bandwidth, ..., and these equipments are now avoiding these busy network links and these two providers.
I'll keep monitoring these aspects until further notice, and will add other data feed providers, but it seems the problem is solved now."
Even though this congestion stuff looks a little bit iffy at first, given the Olympic Games & related amount of streaming, it might actually be the real issue.
Hey DOM, are you concerned that the VPS administrators are snooping on your strategy code? I have my strategies uploaded as a compiled assembly but don't have any sort of encryption and I've noticed the admins sign on every so often, usually on a Friday afternoon. I have steadfast VPS.
I only load compiled+obfuscated (CLiSecure) NinjaScripts on my VPS, and I have a proprietary UsageControl mechanism that prevents my strategies to be used without authorization (I had to do that as I lease my strats to a few clients).
That said, my VPS provider has no access to my VPS - unless I give him my password.
Results for the week ending February, 28th:
- 2 wins ; 3 losses ; net -2515
A few more data-feed disconnects this week, in one instance the historical data reload got interrupted, resulting in missing data after the strategy restart ... there was no harm & I did another data reload manually, but I am also working on a fix for this scenario which should be finished testing tomorrow.
I am still working 100% on the new system, CL Selective, which should come to a "finished" state next week. I got a couple ideas from this work that I will test on CL AlwaysIn asap.
- 2 wins ; 3 losses ; net -2515
A few more data-feed disconnects this week, in one instance the historical data reload got interrupted, resulting in missing data after the strategy restart ... there was no harm & I did another data reload manually, but I am also working on a fix for this scenario which should be finished testing tomorrow.
I am still working 100% on the new system, CL Selective, which should come to a "finished" state next week. I got a couple ideas from this work that I will test on CL AlwaysIn asap.
Results for the week ending March, 7th:
- 6 wins ; 4 losses ; 3 BE ; net +135
Very disappointed by the week, which started well (Long from Friday), but Wednesday/Thursday took 2 max losses in those spikes down then up.
I had time this week to work again on this system, still using 2007..2012 for stat analysis, and 2013 as forward testing. Made 1 change + 1 addition to the set of patterns, the result seems correct (P&L +5.5% overall, all years better except 2010), #trades up 7% (from 5700+ trades to 6100+ trades)(avg / trade down from $99 to $97).
I'll post performance pictures later.
- 6 wins ; 4 losses ; 3 BE ; net +135
Very disappointed by the week, which started well (Long from Friday), but Wednesday/Thursday took 2 max losses in those spikes down then up.
I had time this week to work again on this system, still using 2007..2012 for stat analysis, and 2013 as forward testing. Made 1 change + 1 addition to the set of patterns, the result seems correct (P&L +5.5% overall, all years better except 2010), #trades up 7% (from 5700+ trades to 6100+ trades)(avg / trade down from $99 to $97).
I'll post performance pictures later.
Hey Dom, how many parameters did you have for the system to optimize?