Hi FrostEngine, Does your 1.7 gig ram allowance on EC2 include the server software space? In other words, when you check in resources, do you have 1.7 gig of ram left once your running the server instance? Also, does EC2 offer any kind of auto ping functionality? So you can be notified if the server instance goes down? Finally, is it easy to use EC2 via remote login? Or is it slow and clunky? Sorry for all the questions. I'm really keen to give EC2 a try as a dedicated trading server and I just want to make sure it will work for my needs? Love the thread by the way. thanks, Runningbear
The 1.7gig of ram is the total ram given to the machine. Roughly 600mb is used by the OS. However, even with 7 strategies running currently, I am only using about 950mb of ram. I am unsure about auto ping functionality at EC2, however I am about to implement my own system to alert on NT conditions as well as up time. Remote desktop is working great with EC2, very responsive.
My biggest hurdle right now is not the strategy, its actually performing well. I keep getting burned by NinjaTrader it appears. I am treating this as a "turn on and walk away" system. For instance, I have been traveling since Monday for business. During that time several of my strategies "turned off". Randomly a few order rejected issues happen and NT just turns the strategy off. I can live with that, will put a system in place to monitor and alert.The big issue is "phantom" positions that seemingly do not belong to ANY strategy. For instance this afternoon, I log on and realize I have a position open yet all strategies were in a "quite" period after market close where they are not allowed to initiate positions. The entry name was Strategy1_Exit. Which is a bit peculiar, why is a EXIT signal being listed as the ENTRY to a strategy?? Best of all, it was initiated 3 minutes after the market opened back up. At a time when I place no orders. To top it off,I have a second unexplained position opened. Seemingly at some point a strategy opened TWO positions separated by about 10 seconds. Yet,I have NinjaTrader configured to allow only 1 trade per direction. The second position apparently the strategy was NOT aware of, as it thought it only has 1 position open. So at end of day it closes its one position, while the other remains there. It seems a couple times each week I end up with phantom positions. If I take those trades out, I am up good. But with those, I am barely treading water. Not sure how to correct this situation at the moment, except for checking on the system fairly frequently (which is not what I was intending to do)
Frostengine, Thanks for your feedback. The phantom trades you are talking about have been a long standing problem with Ninja. I have heard a number of traders report the same type of problem. It's a real shame Ninja doesn't commit 100% to fixing these issues.They let another wise decent product down.
Move over to multicharts frosty. I don't think I have seen many reports of phantom trades on multicharts forums. It is definitely a better software. Now a C# version also.
3/4 pip slippage on mkt orders is ok once in a blue moon when mkt gaps. Generally it is not acceptable unless you are trading 30 million size per trade. Your broker is playing with you man. Even if your average slippage is 1.5 pips, it will completely skew your results. At IB, I get -1 to +1 pip slippage. So, sometimes price improves and sometimes price moves against me. All these other brokers are bucketshops. Move away from them if you are serious about trading and want to make money. If you just want to keep programming, well everyone gets what they want from the markets. No offense, just the hard truth. Cheers.
<b>Results for the week of 8/12-8/17</b> <b>Starting Balance:</b> $1003.48 <b>Ending Balance:</b> $998.98 <b>P/L:</b> -$4.5 <b>Pips:</b> -45 <b>Trades:</b> 29 <b>Avg Trade (pips): </b> -1.5 <b>Skipped Trades:</b> 7 <b>Slippage:</b> -20.9 pips <b>Commission:</b> $.53 <b>Total Cost Per Trade (pips):</b> -.9 pips <b>Challenges</b> This week was challenging,I lost nearly $16 due to "phantom" trades. To combat this, I am doing away with the "exit on close" feature of NinjaTrader. Instead I will utilize a timer that will fire 10 minutes before market close where I will cancel all pending orders and go flat. Hopefully, this will reduce the likelihood of phantom trades. I also need to explore better realtime alerting.
Are you sure that's why you lost $16? Could it be that your edge might not be materializing with real money? Has the edge disappeared? Whenever your real money results are deviating from what you expect, this could be a very strong indication that what you thought was an edge was only fools gold...just something to think about. Personally, I would get more defensive with the strategies and drop my position size considerably with the goal being not to make any substantial amount of money but test whether the edge still exists.
I am certain. It is very easy to see. I did a post 2 days ago about the issue. There was one trade that had my exit signal as the entry,and another that seemingly wasn't tied to a strategy at all because the exit on close functionality created it. I lost over $7 on each of those two trades.
Hey Frost: I believe the problem is a known compatability issue between MB and Ninja. I used MB and Ninja for several months and had the same problems you discribe (drove me nuts). Phantom trades, odd errors that kicked me out of my scripts, dropped connections, etc. I switched to IB 3 weeks ago and the problems went away. Suggest you go with a different broker that will allow micro trading and plays well with NT. Re: Different results between live trading and backtesting. I had similar results using MB and I am now testing with IB to see if the problem carries over. Ninja has an article concerning this at http://www.ninjatrader.com/support/helpGuides/nt7/index.html?discrepancies_real_time_vs_bac.htm which may shed some light. The way your broker sends data can affect results in live/historical/market replay (ie. if broker time stamps its data or if Ninja has to use local computer time when storing data leading to time mismatch on replay) Appreciate the journal.