Misc ACC Results Companion 1.24.0

[Doug Duthie]

Hi,
Found a small bug on session tab. On spa and silverstone (both version 2018, 2019), result for a SP hotstint gives a total time smaller than selected duration .
small example :
"Total Time" column shows 0:13:38 for a 15 minutes hotstint with Camaro at Silverstone 2019
0:14:35 for a 15 min hotstint with BMW GT3 at Spa 2019

This doesn't happen with tracks other than spa and silverstone except one result at barcelona that gives 14:59 for 15 min session but it could be a rounded error. And looking for very old session doesn't show this error. First occurence in my database results start on july 24.

Hi,
Yes...that definitely looks to be a repeatable error. IIRC, I changed the way session durations were calculated in about 1.6.5/1.6.6.

Bit of a long boring explanation follows...

Previously they included the rather random time before you crossed the start line. Now it should get the timestamp when you cross the start line and record that as the starting timestamp. Each time it detects you crossing the line after that, it records the timestamp again. The total session should should be the last lap timestamp minus the initial timestamp.

The timestamp comes from the UDP real time update (SessionTime). This should be grabbed every refresh interval (by default every 100ms). I use this SessionTime, rather than a real timer so if the session is paused, the time should still be accurate.

There is a lap completed event, but there is a lag with this as it is an async event so the timestamp here would be inaccurate unfortunately so I can't use that.

So that's how it should work...however:-

You're right, Spa and Silverstone do appear to have issues. I haven't seen the issue on Barcelona (yet). If you look at the database, you can spot the issue as laps.real_timestamp appears to be repeated on at least one lap, or the first lap timestamp is the lap_zero_timestamp

I'll run some tests to see if I can identify the source of the issue

 

[unpierrot]

Barcelona total time 14:59 for 15 minutes session is from a session in february. So I think it's a rounded error.
in table laps, column timestamp (there is no real_timestamp values) for this session shows no duplicate.

 

[Doug Duthie]

Barcelona total time 14:59 for 15 minutes session is from a session in february. So I think it's a rounded error.
in table laps, column timestamp (there is no real_timestamp column) for this session shows no duplicate.

I think it is a process timing issue. The SessionTime is being reported correctly, but in certain circumstances it isn't updated until after the lap details are recorded, so it picks up the old time. Some circuits must be more prone to it than others.

This is going to be fun to fix

 

[Doug Duthie]

I think it is a process timing issue. The SessionTime is being reported correctly, but in certain circumstances it isn't updated until after the lap details are recorded, so it picks up the old time. Some circuits must be more prone to it than others.

This is going to be fun to fix

I see what it was now. There is a lag in certain events being raised by ACC. I've re-ordered where I take information from now. I've got to test to check this is a more robust mechanism, but it should be in the next release

 

[Doug Duthie]

Doug Duthie updated ACC Results Companion with a new update entry:

ACC Results Companion V1.6.14

• Fix total duration for certain tracks (Spa, Silverstone) - lap completed timestamp not always undated correctly
• Fix scan server results on startup
• Support multiple server result folders
• Ferrari ECU mode stats added

Read the rest of this update entry...

 

[Doug Duthie]

ACC Results Companion V1.6.14 Released

Headlines:-

Server Results
A couple of issues around scanning server results on startup were addressed. An issue was introduced when the architecture was moved to a more multi-threaded model. The server scan attempted to stat before the database connection was initialised. Also results files were flagged as read before they were processed - if an error occurred, they would never be read again.

Both above have been fixed.

Also, there is a slight change to the way server results folders are set up. Most importantly, multiple server results folders are now supported.

Quick extract from the doc...

ACC Server Results Folder
If Server Results files are to be used, this points to the results folders. You can now monitor one or more server folders. In the text box, they must be delimited by semi-colons. Alternatively, you can choose the folders by pressing the

button.

Initially, this brings up a dialog with no entries.

To add folders, right click and select Add Folder.

This creates a blank entry.

To manual edit the field, select the row and press F2

Alternatively, double click the row and a folder selection dialog will be brought up.

You can add as many folders as you want. Note, current ARC only monitors Windows folders by notification events. These events aren’t available on Linux shares so ARC won’t be able to spot new files (except via Scan server results folder on startup option)

The folders will be scanned immediately after you change folders. After that, only new files will be processed. (Files are matched based on file name/update time)

If you want to delete an entry, right click and select Delete Folder.

Session total time
There was an issue with the total session time reported on some tracks. The app got the session time from a broadcast interface event on each lap and recorded it against the lap for each car when the car received an update on the interface. For most tracks this was fine, however for Spa and Silverstone, the lap completed event occurred after the car update, so the times were missed. As a result, session times for these tracks could appear shorter than they were.

The way session timestamps are recorded has now been changed to address this

Ferrari ECU Mode Stats added

The list of cars where metrics are currently available are:-

• Aston Martin Vantage GT3 (V12)
• Aston Martin V8 Vantage GT3
• Audi R8 EVO
• Audi R8 LMS GT3
• Bentley Continental GT3 (2015)
• Bentley Continental GT3 (2018)
• BMW M6 GT3
• Emil Frey Jaguar G3
• Ferrari 488 GT3
• McLaren 720S GT3
• Mercedes AMG GT3
• Nissan GT-R Nismo GT3 (2015)
• Nissan GT-R Nismo GT3 (2018)

Documentation in PDF form has been updated at https://www.assettocorsa.net/forum/index.php?attachments/acc-results-companion-1-6-14-pdf.145453/

 

[MB190E]

Doug, I’m sorry if this was asked before; how does one update? Do I need to remove the existing version somehow?

On another note, I’m excited to see you’re adding additional metrics; I didn’t realize they were not input and could not understand why the Lexus would not show any but the car I was comparing to (GTR 2018) would.

 

[Doug Duthie]

Doug, I’m sorry if this was asked before; how does one update? Do I need to remove the existing version somehow?

On another note, I’m excited to see you’re adding additional metrics; I didn’t realize they were not input and could not understand why the Lexus would not show any but the car I was comparing to (GTR 2018) would.

Hi
Just run the installer and it will upgrade the old version.

I will be slowly adding all the other ECU metrics. It takes on average 40 laps per car to gather the stats - I’ll see if I can do one per week just to close this off

 

[MB190E]

Thanks, Doug for the info!

I am running into problems with the Lexus on Misano when it comes to Fuel calculations on Misano. It is consistently underestimating the fuel required for the time selected. I even tried the conservative setting to see if that would let me finish a stint and I ran out of fuel on the last lap.

Both Session recommendation and Manual calculation show the same figures, which are too low. Could this be because you need to do that data gathering you mentioned? if that's the case, no rush, I imagine its a tedious process. But if not, then I thought it worth bringing up.

 

[Doug Duthie]

Thanks, Doug for the info!

I am running into problems with the Lexus on Misano when it comes to Fuel calculations on Misano. It is consistently underestimating the fuel required for the time selected. I even tried the conservative setting to see if that would let me finish a stint and I ran out of fuel on the last lap.

Both Session recommendation and Manual calculation show the same figures, which are too low. Could this be because you need to do that data gathering you mentioned? if that's the case, no rush, I imagine its a tedious process. But if not, then I thought it worth bringing up.

The fuel figures are based off your history (apart from the original run), If it uses system defaults (which are possibly inaccurate weighted extrapolations), then don't rely on it. It should update to your past history after a session with more than one lap.

If it runs from past history, it will look like this

Both session recommendation and manual use the same underlying data.

If you find the fuel recommendation is low, use Ultra Conservative.

The fuel calc for aggressive is based on your average fuel/min for the car/track multiplied by the session duration, plus fuel for 1 extra lap. Conservative and Ultra Conservative add 1 and 2 extra laps respectively.

I'm currently changing the way that fuel stats are gathered and calculated, and also adding in various options to manipulate the results (eg if you know all your history is based on fuel saving runs, it will allow you to tweak or weight the figures). Watch out for this in an upcoming release.

 

[MB190E]

I look forward to the enhancements you mentioned.

For clarity - the results were as displayed in your screenshot, indicating they were from history and not system defaults. However, they were the same numbers as system defaults (other cars showed differences between the two) and this is the only car and track combo (so far) where I had consistently too little fuel even on Conservative. I was learning the track and wasn’t driving flat out so that surprised me, as your tool is otherwise bang on.

 

[Doug Duthie]

I look forward to the enhancements you mentioned.

For clarity - the results were as displayed in your screenshot, indicating they were from history and not system defaults. However, they were the same numbers as system defaults (other cars showed differences between the two) and this is the only car and track combo (so far) where I had consistently too little fuel even on Conservative. I was learning the track and wasn’t driving flat out so that surprised me, as your tool is otherwise bang on.

There might be an issue with the on the fly updating of fuel stats (one of the tings I'm changing).

If you select Monza/Lexus then click Re-calculate, it should completely rebuild all fuel stats from scratch (based on your full history). Note this takes a minute or so to run in the background...you'll see it updating the log page. Once finished, refresh Monza/Lexus on the fuel page and see if the fuel/lap and /min have changed

 

[MB190E]

I think Monza worked fine, the issue was on Misano, but I did Previously use the recalculate button for that track and nothing would change, the values remained what they were (and matching the default non-historical values on the left panel. I’ll try again and give it a full minute, just in case.

thought: it cant be having this issue because I selected ECU map 3 (full power, progressive) and not 1 (full power linear) in game, can it? In your tool there is no way to make it use a specific engine map.

I appreciate the time you invest in discussing this, please don’t think I’m nit-picking or unappreciative of this great tool.

 

[Doug Duthie]

I think Monza worked fine, the issue was on Misano, but I did Previously use the recalculate button for that track and nothing would change, the values remained what they were (and matching the default non-historical values on the left panel. I’ll try again and give it a full minute, just in case.

thought: it cant be having this issue because I selected ECU map 3 (full power, progressive) and not 1 (full power linear) in game, can it? In your tool there is no way to make it use a specific engine map.

I appreciate the time you invest in discussing this, please don’t think I’m nit-picking or unappreciative of this great tool.

I've had a dig through the code and I've decided to completely re-work this area as it is overly complex and slow to generate. I've got a better source of data now, so I'll leverage that and it should simplify things a bit and make the calculation more robust. It looks like there were a few edge cases to generate bad results from old sessions, which then got amplified as the older results were given exponentially higher weighting than the new ones.

The shiny new version is currently in test. I just need to make sure I've not introduced any new issues.

The calculation doesn't take into account ECU mode - it just works off all your history. I'd thought a lot about whether it is worth (ie easy) to give fuel figures per ECU mode, or just the prime ECU mode. However, that would make the calculation much more complex and prone to issues, so at the moment, I've elected to keep it simple. There will be the issue though that if you run the majority of your events in one ECU mode, it won't give an accurate estimate for other ECU modes.

There will be something in the future to cater for skewed results though...

BTW - don't worry - it is the constructive criticism received on this forum that helps the app improve

 

[omg.imba]

Hi Doug,

I have a question regarding Exports:
Is there a way to export the information shown in the session view? I.e. the list of my laps in a session or the final standings? Would be great to have that as an easy csv or text export to paste that info to discord channels or excel.


**edit: the former is incorrect. when clicking in the right location, everything works. see my next post **
In addition: the currently available exports (leaderboard for example) only work for me, if i start ACC RC as administrator. In a regular start, the "choose file location" window does not open and seemingly no file is exported.

 

[unpierrot]

I have a question regarding Exports:
Is there a way to export the information shown in the session view? I.e. the list of my laps in a session or the final standings? Would be great to have that as an easy csv or text export to paste that info to discord channels or excel.

From session tab, you can right click in lower window display -> export as CSV.
Works fine with 1.6.14 version.

 

[omg.imba]

Thanks, I now found my mistake:
If right clicking into the empty area below the last entry, the context menu opens, but the csv export does not work. If clicking on one of the entries, everything works (also without admin privileges).
My initial testing was therefore flawed as I did not click into the identical positions.

Still, this could be improved. Right clicking a single line to export all entries does not seem straight forward. Maybe rename the Menu point to "csv export all entries"?

 

[Doug Duthie]

Thanks, I now found my mistake:
If right clicking into the empty area below the last entry, the context menu opens, but the csv export does not work. If clicking on one of the entries, everything works (also without admin privileges).
My initial testing was therefore flawed as I did not click into the identical positions.

Still, this could be improved. Right clicking a single line to export all entries does not seem straight forward. Maybe rename the Menu point to "csv export all entries"?

I'd never noticed that before, and yes it doesn't make sense. I'll fix it in the next release

 

[unpierrot]

You can also check "Show Summary" menu because it doesn't seem to work.

[00000001] 2020-09-21 20:49:25 DEBUG ACCSessionSummary.showSessionInformationPopup : Showing dialog info b8883aff-7e64-4064-ac89-587f07950858
[00000001] 2020-09-21 20:49:25 INFO ACCSessionSummary.sessionRecordInformation : Showing session summary for session b8883aff-7e64-4064-ac89-587f07950858
[00000001] 2020-09-21 20:49:25 DEBUG ACCSessionSummary.showSessionInformationPopup : End showing dialog info b8883aff-7e64-4064-ac89-587f07950858

From log, it seems popup window closed itself.

 

[omg.imba]

@Doug:

The "show summary" issue seems to be connected to the "Results summary" option under "Overlays" settings. Once this is selected, the summary window pops up.