What constitutes an 'incident'?

Do you have a reference ?
I'll happily read it and define it, if I'm able to access documentation for it ... ????

 

Interesting take on the shared memory as a source of SRS data. I have about five apps listening to it for various things.

I'm going to say no for one simple reason: it needs to be turned on in the settings to output but SRS never stipulates any settings changes in their instructions.

More data from a new race: clean race, two track limit warnings, no incidents.

 

You should really have half a clue before being so critical.
Have a read - remember your BS comment ????????????????????? ... lol

https://onedrive.live.com/view.aspx?resid=F4E1176A98D98E1!330142&ithint=file,xlsx&authkey=!ABsalKX5fwnWtu0

Do you need me to explain it to you too ?
...I'll give you a hint - it outputs exactly what I was implying.
I love your use of the word "theoretical" by the way.

 

I agree it outputs the right type of data, but in ams2 it's unreliable and pretty sure SRS doesn't use it based on my post above.
Server log is the most likely by far, on page I have a link showing the incident output, just not for an SRS event.

 

It's exactly the same concept as per GPS or inertial navigation.
Once you have a defined initial point, relative movement data (column C in that spreasheet - the first one visible) will let you easily calculate the exact position and movement of the car.
i.e. all the initial items labelled "car dynamics".
This is pretty simple stuff.

I'm kind of shocked that Herr Wolfgang thinks this isn't achieveable.

 

Perhaps - but the obvious question is - where is the alleged log ?
I don't believe there is one for this game engine (under normal circumstances).
It's quite possible the SRS application extracts data and writes it to a log of some kind, but then its their log of API information and not from the game engine itself.

Query them.
Where is the log?
Does it exist?
My money is on the bet that there is no log.
There is no need for it, therefore logic suggests they wouldn't use one.
I would also suggest in AMS 2 it is reliable in my own experience. Crew Chief uses that API and that seems very reliable to me.
You'd need to back you that claim.
I'm not trying to have a go at you - just saying, that's an easy example of an existing application that uses the API and it works perfectly.

 

here is the original api
And it definitelly does not output incidents just damage states.
I'm working with the api since pCARS1 in 2015, wrote some udp readers and a dash board and also was in DS LUA programming.
Whats outputted on DS is not documented. You'll have read the original lua scripts to write your own code in lua or js to reverse engineer whats happening.
API For External Apps - Project CARS 2

 

You don't seem to understand this at all do you?
There is no need for a defined incident state.

Do you not understand how movement and velocity data can be interpreted?
By calculating the relative positions and movement of multiple objects, incidents, impacts and other events become self evident from the data.

Let's go through an example.
Object 1 is player1's car.
Object 2 is player2's car.
If I know all of the positional and movement data for both cars, I can easily determine when an incident happens.

There does not need to be anything coming from the API that actually defines something as an incident.

Do you have another argument Herr Wolfgang ?

 

its on a dedicated server. If you use p2p sessions , there is no log.
If a DS is setup correctly and you rewrite the sms_stats.lua script like i suggested in the DS thread, you get the output to sms_stats_data.json. This is the logfile that is interpreted in simresults and also from other services.
As you seem to not know that, your assumption went in the wrong way

 

i understand that. But then this discussion is obsolte. Because SRS creates incidents for clean races and its definitely not worth using their system. Dependend what OP asked.
But i am very shure, they use the buggy output of the ds and interpret this also wrong

 

Wrong - you have literally proved my exact point.
The logic that the SRS application code is using to determine an "incident" is clearly wonky.

Does that make sense to you ?
(and it's what I said in my very first post in this thread).

I'll clarify that as I suspect you're going to interpret my words incorrectly.
I already know it's on a dedicated server.
The SRS application is a client server style application.
On one end it interrogates the clients application API - AMS 2.
On the other the SRS application will be interacting with either a back end application, perhaps a web server, perhaps scripts - maybe a combination of the previous items and perhaps something else. I'm explaining a concept here and not the specifics as frankly - it is not relevant.
Bottom line is the SRS application is doing _something_ with the data and it's pretty obvious it's then using that data from all the client applications in a given session (a race and remember the client applications are dynamically extracting data from the game - AMS 2), applying logic and using that results to record incidents.

I think that's about as clear as I can make it.

 

btw. I have NOT found any hint at SRS that you have to activate telemetry output like shared memory or UDP.
If telemetry would be used as you suggested, it would be written every second line at their forum, as its deactivated in default.

 

That's not at all correct.
That depends entirely on code at their end.
How and what they choose to do with it is their decision.
I could write an application that extracts data from the AMS 2 API and then updates it once per second or I could write another application (in another language, just to mix things up further) that then stores than data and updates it once per year.

That's entirely down to application logic.
Real time applications are written quite differently to non real time applications.

 

yes. But when telemetry is off, you can't.
And as SRS gives no hint to activate telemetry or insists on activating, i'm 100% sure they don't use it

 

It may not be using the actual API interface.
Regardless, my previous comments still stand.

Have a read:
Automobilista 2 App

1) it's obviously memory resident
2) it's being started as per the information in that thread.
3) it's pretty damn easy to directly read the memory addresses if you know them. and given that the API has been documented - that would be very, very easy to determine.

Are you familiar with the concept of a kernel level debugger ?
My point being - my initial comments were guesses. If there's an API then you would expect an application to use it, but given they are using a memory resident application, if it's not relying on the API itself, the logical conclusion would be it is directly reading the relevant memory addresses.

Effectively it's doing the same thing - just not using the proper API methods.
Look at the code behind memory resident game trainers for an example of that sort of software.

To repeat my earlier comments:
If you are using their application and the incident logic is faulty, that is due to the programming logic that is being used in their application. Not in AMS 2.
If it was inherent to AMS 2 - then it would behave in the same way - regardless of if you were using the SRS application or not.
The original post in this thread - the very first line states that is not the case.

 

I don't know why you're so quick to blame SRS. It could from that end, ams2 or both.
Ams2 has plenty of bugs and suffers at times from some very dodgy netcode. Very possible that what comes from ams2 is unreliable.
Just look at the disparity in the HUD timing between the relative and standing. Even the live lap delta constantly disagrees with what is reported each sector and after the lap is done.

 

Using this thread to keep a log.
Latest race, I was clean and has no incidents recorded.
Three collided at the start, with one pushed off track, however none of these drivers had incidents.
The only one who did out of eight in the race had two incidents recorded. I watched back the replay in fast forward. They spun and hit the wall, then pitted next lap for repairs. I would think this is only one incident, not two. Plus the drivers who had actual contact all had zero?!

 

FWIW, I have another example of this. I was in a race where I got collected in a first lap spin. After that I had no collisions with other cars and I got no warnings from Crew Chief about leaving the track. But at the end of the race I had 13 incidents. I really don't know why.