Gaps in INTRADAY bars Dan ? A common occurence ? Maybe for illiquid instruments/symbols. How about ES ? Ok, ok...even then it could occur....like when Treasury Secretary Baker uttered that infamous word "regrettably": ...respected money manager Monroe Trout describes his ordeal: "January 9, 1991 was the day that Secretary of State James Baker met with the Iraqi ambassador in an effort to avert the Gulf War. At the time there was a reasonable degree of optimism going in to the meeting. Addressing the press after the meeting, Baker began his statement with the word 'Regrettably.' A wave of selling hit the stock and bond markets. I lost about $9,500,000, most of it in about ten seconds." Trout was holding 700 S&P futures contracts at the time. It is my understanding that IB's REAL TIME BARS are not from sampled "prints" but from ALL of them.
Well since I think you had some knowledge, let me ask u for ohlc bars, its very applicable to Daily charts, since we have gap ups and down due to market open and closing. Hmm for 5 mins i m not sure, cos im running 30s bar charts, and I create the bars myself, with just O, and thats open prices. Since prices come in continuous flow, so why be bothered with gaps ? And with Google Finance and very very liquid stocks @ 5 mins, I don't see many gaps, do you ?
Of course...that always occurs in markets that close. That's why many traders look at TRUE RANGE. Like I said, this is sort of a test of the data. If the prior close and current open are consistently different, then ticks are being missed or it's an illiquid symbol. In the case of intraday data, true range and range should be the same.
IB's Real Time Bars is in 5-sec periods, unfortunately with no Close price data being provided at present for some reason. I am trying to convert these into 5-min periods. So, what you guys are saying is that if say the open price of a 5-sec bar at 00:00 (min:sec) is 5, then I should take the close price at 04:55 (min:sec) [end of that 5-min period] to be the open price of say 5.5 at 05:00 (begining of a new 5-min period). Is this assumption correct?
That's crazy. They've been working on this now for over 2 YEARS....and it's still not right ? Unbelieveable ! Close of the prior period should be the same as the open of the subsequent i.e. current period. For your 5 minute bars, because of IB's omission as stated above, your CLOSE cannot be determined until the OPEN of the 1st 5 second bar after the end of the 5 minute interval.
Err well, dan. if I were you I'll have to prioritize the entire project and strategy, and spending effort to determine proper closing price, is it worth ? I think its not effective at all as there are more things to worry about. What about those with algos which sank deep into level 2 quotes with the orders that flashs at 100,000 times per second bid/ask I am running my algo on 10+ stocks/futures over several exchanges in the world, and heck.. closing price @ 5 seconds ? my 2 cents worth...
Thanks for your reply savagemp5. Iam not running any algos or anything of that nature. All I just wanted to do is to be able to monitor the 5-min OHLC of say about 100 stocks on an excel spreadsheet at quick glance, that's all. Cheers
You are confusing high frequency trading with algorithmic trading....the two are not the same. One moves away from the HFT category when the average time-in-trade goes over 2-3 minutes. That being said, I don't think this is a hard-and-fast "rule", but one must draw the line somewhere. And using that metric makes sense. Heck, who knows, maybe the SEC is trying to come-up with that rule as we speak. After all, they are hell bent on doing SOMETHING to stop those HFT boys. Since they have been making all of that money, they're the new "bad boys" in town.
well.. thats still an algo, just that it requires high speed and more in depth of details, and I am not quote staffing in anyway. And as fast as I wish for my algo, the lag between my excel generating a buy signal to the time my order gets executed is around +- 1-2 sec. So in fact, it sucks i havent break into the millisec level.