The data feeds we gather intraday data from ( eSignal, DTN IQFeed and TradeStation for example) all have their own chart displays. Frequently users notice minor differences between the bars displayed by charts from the feed and NeuroShell and feel that differences indicate that there must be a bug or flaw in NeuroShell. However, the nature of the way data is sent from these sources makes it impossible to always match the bars tick for tick.

First of all, we have to “poll” some of these sources (e.g. TradeStation) to see when a new tick is available. If we poll too frequently, the CPU will become saturated with polling especially if the user is gathering data from dozens of ticker symbols. Therefore, we have tried to poll every few seconds, but this polling could miss a tick or so in a fast moving market.

Secondly, there are sometimes disruptions in the transmission of data from the source to NeuroShell. This can occur because the Internet delays the traffic, or because some other program in your computer has grabbed the resources, meaning that either the feed can’t get the tick or NeuroShell can’t get it from the feed. However, NeuroShell has to close the bar at the end of the bar’s time period, because it doesn’t know if there are delayed ticks in the pipeline or not.

Finally, some sources have multiple servers that aren’t always in absolute sync. NeuroShell may be serviced by one and the other charting program may be serviced by another.

The bottom line is that completed bars in NeuroShell may not always look exactly like completed bars in the chart of the transmission source, and until better transmission sources are developed, we cannot improve on that much (although we are always trying). Although the open is usually correct, the high, low, or close could be as much as a tick off. This in turn means that the completed bar will look different to the eye. The volume could have some differences too for the same reason.

If it is any consolation, we can tell you that the data feed vendors themselves are experiencing the same problems as they gather data from THEIR sources, and as THEIR sources gather data from the exchanges.