I was allowed to remote into my friend's pc and actually saw the mismatch in the high or low of a swing on his 5 minute chart vs my unfiltered feed. It is easy enough to see if you run one setup with IB minute data and another with an unfiltered datastream set for the same minute bars. While it's true that IB would best be used only for order execution, it seems to me that it's a waste for them to even provide the minute data if it creates inaccurate bars. It might "represent" a picture of what the market is doing, but sort of like the way near-beer represents beer...
I'm with kostia .... I don't believe this. And to be honest esignal data can be faulty because, at least in the past it used to suffer lag under fast market conditions. I have compared a lot of charts based on ib 5 sec data with exchange data and yet to see such issues. The 5 second data is correct and the backfeed should also be fine as its from the same data base. I have never seen evidence that the "true" data was missing ticks so I suspect its either your interpretation or the software you are using.
>kiwi trader You should give eSignal another look. The lagging data that you saw in the past does not now occur in their new eSignal 11. You and I are a long way from their servers. But I find that the data lags only 0.0 to 0.3 seconds compared to that of local brokers. -------------------------------------------------------------------------------- Regarging IB's data one thing to consider is if they are getting the data directly from the exchange or through a vendor as they do in some countries where they are not exchange members. Where I am I noticed a significant improvement in stability after they swiched from getting data from GTrade to there own data which is now solid as a rock. However, as others have mentioned there are missed ticks on 1 minute charts. But they are corrected to the letter upon refreshing the chart, as seen when compared to Reuters or eSignal. As for the 1 in 20 bar misticks after refreshing that someone mentioned, I would have to see it to believe it. In fact, after refreshing, I doubt that I could find even 1 mistick in 20 DAYS!! Face it, any software, including Reuters or BB, can miss a tick now and then. But I have never seen such to the extent that has been claimed here. And the chance of a missed tick coming within a bar is , of course, much greater than at either end. I would say that a missed high or low with contracts I trade occurs maybe once or twice a day at the most.
OK I will post on monday. I'll try to find my CSVs and test outputs but if not i'll just re-run them and post results. Note: the IB data i am referring to was downloaded via the TWS API and with the historical request i think i only specified ISLAND (since you cant obviously specify SMART) so I'm not sure if that matters.
I asked for some more info after reading the posts here, and the discrepancy might be because he is using NinjaTrader to view his charts, and NinjaTrader does not yet use TWS 9.2, which is where the added 5 sec option in the API is implemented. If I understand correctly, the 5 sec option is a refresh of OHLCV in case a top or bottom tick of a bar is missed because of the data bundling. So NinjaTrader (or any other charting package) would have to use TWS 9.2, and then code something to use the API call to correct the bars in realtime.
Forget that last post - apparently it is the API version 9.2 that started the 5 sec bars, and has been out for a long time. So that means that you would have to build bars that are sampling at 5 sec rates and adjust the OHLCV values until the time of the bar you actually want to watch is finished, then restart sampling for the next bar to get the true OHLCV of each bar as it is built.
As some people know, I have a love/hate relationship with IB. One thing I do love about IB is its "snapshot" data. In hyperfast markets (eg, right after Fed announcements, etc.), IB is the ordinary trader's best chance of knowing what the market is actually doing. Quite frankly, I can't understand why any professional or even semi-professional trader would rely solely on a brokerage data feed instead of spending $125/month or so to subscribe to a dedicated data service. Frankly, it's incomprehensible to me. Auto mechanics, surgeons, and concert pianists are most demanding when it comes to the tools of their respective trades. What an irony that traders--whose success depends on their ability to discern value--practice such false economies.
+1. and traders should make sure they quality for any cme fee waivers so they don't pay money they don't need to. there's also an arb b/w data providers if you use nt b/c kinetick uses the iqfeed but it's cheaper than buying the iqfeed directly. no i don't work for nt, kinetick or iqfeed.
I would agree with using an alternate feed, but there are a couple of issues that are obvious here. 1 - you have to know that the data feed from your broker is unreliable (or that the charting package used is not displaying pricing correctly using that data), or you are likely going to make inadvertent mistakes and have no idea why. 2 - people starting out are usually looking for the lowest cost solution because it will take some time to become acclimated to what they are watching, so can't really blame them. 3 - the cost to provide charting data that is unreliable means the broker can cut that part of the service and maybe pass the savings on since they don't have to fund that ongoing development.