Missing Candles...

[Pipskateer]

Hello,

While going over daily trade reviews recently, I began noticing that my charts did not have the same candle formations as my trading partners. In fact, my charts are actually missing several candles and my moving average lines and other indicators have completely different signals.

One of my partners had discovered from talking with NT tech support that it is advisable to rebuild the database periodically to bring the charts back into condition. I find that I have to do this almost constantly (every 5min at LEAST) to keep accurate candle patterns on my screen. I have missed good trade entries and have taken bad trade entries directly because of this issue. This is a completely unacceptable and untradeable situation which must be resolved. You can't trade profitably when your charts are 8-10 candles delayed behind the real price action.

I have read some of the posts that might be related to this issue in my forum search for answers to this problem. I have read the user guide which talks about how the chart candles are created and the time stamping factor. I understand that not every machine will build the charts identically and that some candles will be different, but that is not what's happening here. I should still have the same number of candles that everyone else has, even if they are not identically formed via OHLC data. My NT terminal is not set to filter tick data, btw...

I use a Linksys wireless 802.11g LAN router in my home with a 5mbit broadband cable ISP. My PC is a Toshiba laptop with a dual core AMD processor with 2gb RAM running fully updated drivers and WinXP SP3, all fully updated. I have disabled my Commodo firewall software and my McAfee AV software with no improvement. There is plenty of available RAM memory (almost a full GB) and I keep other programs and network connections to a minimum during trading. There is only one other PC (my wife's) running on the LAN with fairly low traffic.

I am now running another instance of NT on my wife's PC (new HP w/3gb RAM on Vista Home) using my AMP Futures live data feed through Trading Technologies. I am running my own PC concurrently with the AMP demo TT server connection. I will have the results of that testing tomorrow. In the meantime, I am posting copies of my charts here to show the issue I'm facing and hopefully someone will have some idea of what might be wrong that I haven't considered yet.

I have done some other LAN testing, but I want to see what suggestions you guys might have before I get into that here. I need to know where this problem is coming from before I go to my ISP and complain or approach AMP with questions about their datafeed.

Here are a couple of chart pics I took earlier today before the market closed. These are 600 tick charts of the S&P 500 emini. I did a database rebuild at 2:00pm, but the first pic is at 2:38pm edt just before I did a FRESH database rebuild. The second pic is at 2:40pm edt just after the db rebuild. Note the missing candles after the red MA bounce down and subsequent run back up. The 2nd pic represents the degradation from the db rebuild at 2:00pm to the one at 2:38pm.

Thanks for your help in advance!

[Pipskateer]

Followup...

I started up my wife's PC shortly after the market closed today to run along with mine to see if they are both having the same data loss. I did a database rebuild on both computers as soon as I started the test.

I am still running data on both computers on the 600 tick charts, but just did a comparison of a 55 tick chart on both machines WITHOUT any database rebuild yet. Both charts have very similar candles but they're not identical...exactly what you might expect with normal data variances between any two machines. There are no missing candles between the two machines, only different OHLC data.

I then compared our two charts with a chart from one of my trading partners' 55 tick chart around the same time. There are at least 10 candles missing from our charts as compared with my friend's chart between identical price point swing highs. So, it would appear that my data problem is not specific to my computer. I still need to know if it's a NinjaTrader problem or if it is a local network versus ISP connectivity issue.

[ctrlbrk]

TT uses filtered tick data.

Zen fire, also available thru Amp (but you'll have to open a new account and have them transfer the funds), does not filter data.

If you are using tick based strategies and charts, Zen Fire is preferred. This is not just my opinion, but also the opinion of Amp (Dan).

If the person you are comparing to is not using TT this could be the culprit.

Mike

[Pipskateer]

TT uses filtered tick data.

Zen fire, also available thru Amp (but you'll have to open a new account and have them transfer the funds), does not filter data.

If you are using tick based strategies and charts, Zen Fire is preferred. This is not just my opinion, but also the opinion of Amp (Dan).

If the person you are comparing to is not using TT this could be the culprit.

Mike

Thanks Mike,

I was not aware of that fact. It certainly could be the problem. I'll do some comparisons to be sure.

[Pipskateer]

Final Results....

After further comparison of my charts with a friend's Zenfire charts, I can say 100% this is clearly a case of data filtering from TT. I have now run a Zenfire data feed alongside a TT data feed on both of my machines and there are between 30%-50% fewer candles generated for the same price swing high and low points on the TT feed. So...a TT data feed is completely unusable for my trading style.

Thanks for your information Mike, it saved me a lot of additional testing and wrangling with my ISP company. I did have an outdated cable router and some defective cable fittings that have now been replaced, but the main issue is the TT data. I'll either be moving to the Dorman/Zenfire feed at AMP or I'll be changing brokers entirely. Thanks again!

[cesas]

filtered tick data

TT uses filtered tick data.

Zen fire, also available thru Amp (but you'll have to open a new account and have them transfer the funds), does not filter data.

If you are using tick based strategies and charts, Zen Fire is preferred. This is not just my opinion, but also the opinion of Amp (Dan).

If the person you are comparing to is not using TT this could be the culprit.

Mike

Hi ctrlbrk!!
Do you know if this can happen with 5 min chart too? I noticed that my real time charts (opened in real time work) are different from off line reload data. On real time time the bar closes and indicator signals are different from the reload past data. I use onbarclose strategy and this happen all the time. I think this problem is related with the filtered ticks or the timestamps, or exchange clock. Do you have any suggestion? thanks!