NT8 bug with stock Renko bars while using multiple series on same chart?

[forrestang]

Hello.

I’ve noticed what seems to be a bug in NT itself with collisions that occur when loading Renko bars of different sizes on the same chart(one primary, one secondary) series.

This works fine w/no data feed connected and the chart is built w/historical data.

I noticed this issue as I built my own renko bar, modeled off the stock NT renko bar, and realized this issue existed in the standard one also.

---

When there is no feed connected, workspace open, and then connecting to a data feed, the bars begin stacking on top of one another.

---

I can temporarily repair this error by opening up the chart settings, switching the bar sizes to a different value, and then the chart will seemingly repair itself. Although any cached data that was built incorrectly will be built into the data. See image below:
[img2]

---

However, after the chart runs a while, and this happens randomly, the error exceptions below will pop up. That and one or both of the series will again begin misprinting or omitting bars.
[img3]

---

I think the issue is that the multi-series use a merged x-axis computed during the workspace’s start, and that layout goes stale during live ticks because each series has a timestamp that gets rewritten on each tick in UpdateBar(). Over time, the cached layout and live timestamps drift apart, and and the render loop that draws the chart style against a collection that isn’t there, causing the IndexOutOfRange errors.

To put it more simply, consecutive bars can be emitted w/identical timestamps during live ticks. W/one series, this is no problem. But w/two series, the merged x-axis tries to index via these dupe timestamps and causes the out-of-range index lookups, which is what I think leads to the error?

I’m wondering if this is a known NT bug w/NTs renko bars, and if there is a way around this?

[forrestang]

I could only attach 1 image in my opening post, so I’ll include the other 2 where those [img] placeholders where.

img2

[forrestang]

img3

[forrestang]

I could be wrong…
But I thought maybe the delayed data might be the problem. I got live data, and my bar series w/a secondary series seems to work fine

[Andrew_Tsujiguchi]

This is a difficult question because when we use Renko charts for indicators, they usually plot like this, which looks very strange. Have you tried another Renko chart from a different code?I also faced this problem when evaluating NinjaTrader, even with charts containing different data series. It’s a programming issue if there are two data series on the same chart.

[forrestang]

Howdy,

I got live data, and my bar series w/a secondary series seems to work fine

Actually, I don’t know what was happening at the time, and maybe I didn’t explain it well in my post above, but everything works fine now.

I’m not sure what was happening? Either I just wasn’t paying attention, or it was due to using the free delayed data feed?

But once I connected to LIVE data, everything worked fine. I explained above, that I suspected the bar series I built just struggled w/the DELAYED timestamps, attempting to plot it on a real time chart or something? I.e., some internal collisions with time.now() and timestamps the bars were receiving in real time? This was happening w/the stock NT renko bars and the custom renko bars I built.

As the historical bars would build fine, but streaming it live would break everything.

But all good now.

[Andrew_Tsujiguchi]

When I encounter problems in my code and need to review and find the specific issue, I sometimes turn to claude.ai, and it always points me to the correct solution.

[Maverick]

You’re using a 3rd party bar type, so it might not have been coded correctly, which could be causing all kinds of issues. I’ve gone through this phase as well and it 9 out of 10 times it was a poorly coded tool. At any rate, as Andrew says you’ve gotta dig into the cause with AI.

Suggestion, screenshot your compilation error and paste it in to Claude or ChatGPT.

[forrestang]

Maybe I’m still not explaining this well?

This was on the STANDARD NT renko bar type.

I noticed this problem on MY bar type initially, then I wanted to see if the problem was MY bar type, or some other issue.

I loaded up the STANDARD NT renko bars, and they too had the same problem.

The images I attached several posts above are showing the standard NT renko bars(the gray bars).

As I mentioned before, I think the problem was a timestamp collision using NT’s FREE, delayed data feed.

Imagine if you stream data that is 10 minutes delayed. When you attempt to push that LIVE, to a time-axis that expects to receive data w/a timestamp of time.now(), and it is delayed by 10 minutes, it causes a collision on the x-axis, and the platform doesn’t know where to put that data.

I think this is the case b/c once I used a LIVE data feed(not delayed), everything worked as expected. Additionally, the historical bars would build fine, it was just the LIVE data.

TLDR: It wasn’t the 3rd party bar type, it was the stock NT bars. The issue was not the bar-type, it was using a delayed data feed that caused a timestamp collision. Everything works as expected when streaming a Non-delayed data feed.

[forrestang]