Support Board
Support Board
Search Board
Control Panel
View My Posts / Threads
Direct MessagesDate/Time: Tue, 16 Apr 2024 12:47:38 +0000
Bar Based Backtest with STudy/Price overlay referencing different symbol
View Count: 568
| [2020-11-21 14:49:53] |
| jwick - Posts: 142 |
|
I'm unsure if this is expected behavior or I have something incorrectly setup. Here is what I'm observing: - I do a bar based backtest on an empty chart. The performance is very fast. - I do the exact same bar based backtest after adding a study / price overlay referencing the main price graph of a different symbol (same timeframe). This is the only study on the chart. Now, the bar based backtest grinds to a halt at around the 40% mark (moving up maybe 1% every 15 or so seconds). Adding a study / price overlay referencing a chart of the same symbol produces no issue. Adding a higher or lower timeframe doesn't make a difference - just that I referenced another symbol. Thoughts? EDIT: To clarify, my actual objective here is to get a study/price overlay of a custom study residing on a chart of a different symbol. The case above is just a test to simplify what I'm observing. Hence, the overlay study / the add additional symbol study wouldn't work in my case. Date Time Of Last Edit: 2020-11-21 14:57:03
|
| [2020-11-21 16:51:41] |
| jwick - Posts: 142 |
|
hmm ... I seem to have found a solution to this use case which might work. So to recap the problem I'm noticing is when chart A references B (each of a different symbol) using Study / Price Overlay and I run a bar based backtest on A. I think because of the bar matching required between the two symbols (despite being on the same timeframe ... maybe chart B has a missing bar or closed at a different time than A on a particular date which is when the slowdown begins?). Anyways, the solution I found was to introduce a chart C (with the same symbol as chart A) and place a study / price overlay on chart C referencing B. Then, add a study / price overlay on A referencing C instead of B. With this config, running a bar based backtest on A runs buttery smooth with it capturing all the needed data from B as expected.
|
| [2020-11-27 21:44:57] |
| jwick - Posts: 142 |
|
Update: I began experiencing the same issue on another symbol ... but this time the source and destination charts shared the same symbol. The only difference was timeframe. However, the workaround above (using a third chart in the middle with the TF of the destination chart against which a bar based backtest is going to be run) works here as well.
|
| [2020-11-29 01:37:23] |
| jwick - Posts: 142 |
|
Sorry to keep posting about this but having another weird issue ... I'm running the exact same chartbook (one a copy of the other) in two different SC instances. In one, the backtest is chugging along at 2-4% every second ... in the other, it's grinding at maybe a percent every 20-30 seconds. I'm confused ... it's the exact same data, same studies, same chartbook ... yet one sub-instance runs it fine and the other goes really slow (Total CPU usage is just 5% so I can't be bottlenecking there). On v2204. EDIT solution: Erpaderp ... there was one critical difference between the two subinstances which I didn't notice ... chart update interval! Once I ratcheted that down, both ran smooth as butter. Date Time Of Last Edit: 2020-11-29 01:53:58
|
To post a message in this thread, you need to log in with your Sierra Chart account: