So far I suck at making good autocross courses, but I figure I'll get better with some practice and studying. I'll update this post and thread as I make new things, with a description of what I was trying and whether or not it worked.
I will primarily be using the XF GTI (most basic demo car) as my test car because I used to autocross an old 2001 Ford Focus ZX3 and this seems like it has similar specs. My racing wheel is a Logitech G920 with the shifter.
Google Drive folder with all the layouts:
https://drive.google.com/open?id=1r10wusjLfxPtNFUhWLZtMVZMLvFhlyBi
SCCA's autocross design guide
http://www.greatfallsracing.com/wp-content/uploads/PDF/CourseDesign4.1.2.pdf (if you're from the SCCA and I'm not supposed to share that, let me know, but it looks like it's publicly available all over the place /shrug)
------------------------------------------
Course 1 - 'Nuf Said
AU1_First_Track_EVER.lyt
What I was trying - My first course! I tried to use some of the principles in that guide to create a course that isn't painful.
Did it work? - No. The course is stupid.
Course 2 - Sea of Cones Experiment 1
AU1_THE_CONE_SEA_1.lyt
What I was trying - This time I tried to make a course that would flow better, without worrying about what the actual course looked like (hence the sea of cones). I noticed that some of the layouts in that guide are effectively cones laid out around a series of circles, so I thought maybe I could come up with a decent layout by starting with the circles. I took the chart of corner radius speeds and made a bunch of circles of various radii, using 30m/100ft as the base size. I then made a path using route markers and traced it with cones. After that I drove it a few times to get skid marks on the approximate driving line, and made some corrections with green cones because there were a few spots where I would get lost, or I'd accidentally keep following a curve on a section that was actually straight and barrel into a wall. In real life this course is very poorly designed (the crossover towards the middle, etc.), but I was trying to liberate myself from the rules and I figure you have to make some bad ones to start making good ones.
Did it work? - I think so. Looking past the raw unrefinedness, I think it's a course you can actually drive, or at least it's a great starting point. Looking at the skidmarks, I can see how I could turn some of the sections into standard course elements like slaloms. This was a pretty good exercise and may have given me a method for coming up with new courses.
Course 2.5 - Cleaned-up/Finalized Sea of Cones Experiment 1
AU1_THE_CONE_SEA_1_CLEANED_UP.lyt
AU1_THE_CONE_SEA_1_FINAL.lyt
What I was trying - I deleted buckets of cones to get at the few key cones required to create the course. This was also a useful exercise to figure out where to put key cones so I can skip the cone-sea step in the future. I think I got the course visuals okay, but if anyone drives this please let me know if you got lost (and better yet, how to fix it). AU1_THE_CONE_SEA_1_CLEANED_UP.lyt is the same as THE_CONE_SEA_1, with some of the extra cones removed.
Did it work? - Yes. I ended up with a legit-looking autocross course with some nice sweepers and varied slaloms. I think it flows nicely, keeps you in second gear (in the XF GTI at least), and has some challenging sections with plenty of line options that you can get a little bit wrong without necessarily plowing through a bunch of cones.
Course 3 - Accessory Drive Belt Experiment 1
AU1_BELT_F150.lyt
What I was trying - I have a 2001 F-150 5.4L and I noticed the serpentine belt diagram under the hood and thought it looked sort of like an autocross course from the SCCA's design guide. So I snagged this diagram, measured the pulleys with an advanced imaging tool (MS Paint), assigned pixel diameters to meters (30m per 80px), and fired up LFS.
Did it work? - No. It's pretty bad. It ended up with a lot of tight sections and lock-to-lock turns that really suck to drive. If I doubled the radii and distances, it would probably still be terrible because it doesn't have good flow. You'd have the same steering experience, just at slightly higher speeds. It might get better if I spaced out the pulleys a bit more, so that the turns are more like 90-135 degrees instead of mostly 180-270 degrees.
Course 3.5 - Finalized Accessory Drive Belt Experiment 1
AU1_BELT_F150_FIXED.lyt
What I was trying - I took the same circles from Course 3/AU1_BELT_F150.lyt and moved them around so the turns aren't as tight.
Did it work? - Yes. The final product is a short, decent course that is actually drivable. I only added one cone, which was to create a slalom out of an otherwise boring section towards the end. So, are serpentine belt diagrams a good source of autocross inspiration? Maybe. It might not be a bad way to start, but expect to move everything so it looks nothing like what you started with.
Course 4 - Tire Rack Autocross 1
AU1_TireRack.lyt (template)
AU1_TireRack_Autocross1.lyt
What I was trying - I wanted to try recreating a course I had driven in real life. I did my second autocross at Tire Rack in South Bend, Indiana with the BMWCCA in my derpy $500 Ford Focus ZX3. It's approximately parking lot sized, so using chalk lines, InSim circles, and Google Earth, I drew the test track and slapped some cones on it.
Did it work? - Yes. I had drawn a sketch of the course after I walked it, and even though it's kind of a chicken-scratch drawing with no distances, I get roughly the same times I saw in real life: ~46.2 +0 (Focus) vs. 44.78 +0 (XF GTI). It was a really hot day (at least 90F -- the Focus started stalling at the line if I didn't give it some throttle in neutral) and the XF time might be with my stiffened autocross setup. Anyway, that Tire Rack template is pretty sweet!
Course 5 - Tire Rack Autocross 2
AU1_TireRack_Autocross2.lyt
What I was trying - I wanted to see how good my TireRack template is for throwing together a decent course quickly.
Did it work? - Yes. It probably took me about 10 minutes to throw down some cones and maybe 15-20 minutes to make adjustments after driving it. It's not bad! TireRack's track provides helpful constraints with excellent versatility. It turns the problem of designing a course from playing God on a blank canvas to arranging the elements of a preexisting course.
Course 6 - Grissom Day 1
AU1_Grissom_Day1.lyt
What I was trying - I drove at an autocross event at the Grissom Air Force base a couple weekends ago, and I wanted to create that course.
Did it work? - Ehh kinda. I was working from memory, and Google Earth says the concrete pad at the base is a bit bigger than any of the parking lots in LFS. I tried to recreate the feel of the course, but the proportions aren't right. It was a pretty tight course and I felt like I couldn't achieve the feel of how tight it was without shrinking it a little. The visuals are also probably confusing. Since it's a bit too short, the XF GTI beats Corvettes!
------------------------------------------
Now go buld a tark!
I will primarily be using the XF GTI (most basic demo car) as my test car because I used to autocross an old 2001 Ford Focus ZX3 and this seems like it has similar specs. My racing wheel is a Logitech G920 with the shifter.
Google Drive folder with all the layouts:
https://drive.google.com/open?id=1r10wusjLfxPtNFUhWLZtMVZMLvFhlyBi
SCCA's autocross design guide
http://www.greatfallsracing.com/wp-content/uploads/PDF/CourseDesign4.1.2.pdf (if you're from the SCCA and I'm not supposed to share that, let me know, but it looks like it's publicly available all over the place /shrug)
------------------------------------------
Course 1 - 'Nuf Said
AU1_First_Track_EVER.lyt
What I was trying - My first course! I tried to use some of the principles in that guide to create a course that isn't painful.
Did it work? - No. The course is stupid.
Course 2 - Sea of Cones Experiment 1
AU1_THE_CONE_SEA_1.lyt
What I was trying - This time I tried to make a course that would flow better, without worrying about what the actual course looked like (hence the sea of cones). I noticed that some of the layouts in that guide are effectively cones laid out around a series of circles, so I thought maybe I could come up with a decent layout by starting with the circles. I took the chart of corner radius speeds and made a bunch of circles of various radii, using 30m/100ft as the base size. I then made a path using route markers and traced it with cones. After that I drove it a few times to get skid marks on the approximate driving line, and made some corrections with green cones because there were a few spots where I would get lost, or I'd accidentally keep following a curve on a section that was actually straight and barrel into a wall. In real life this course is very poorly designed (the crossover towards the middle, etc.), but I was trying to liberate myself from the rules and I figure you have to make some bad ones to start making good ones.
Did it work? - I think so. Looking past the raw unrefinedness, I think it's a course you can actually drive, or at least it's a great starting point. Looking at the skidmarks, I can see how I could turn some of the sections into standard course elements like slaloms. This was a pretty good exercise and may have given me a method for coming up with new courses.
Course 2.5 - Cleaned-up/Finalized Sea of Cones Experiment 1
AU1_THE_CONE_SEA_1_CLEANED_UP.lyt
AU1_THE_CONE_SEA_1_FINAL.lyt
What I was trying - I deleted buckets of cones to get at the few key cones required to create the course. This was also a useful exercise to figure out where to put key cones so I can skip the cone-sea step in the future. I think I got the course visuals okay, but if anyone drives this please let me know if you got lost (and better yet, how to fix it). AU1_THE_CONE_SEA_1_CLEANED_UP.lyt is the same as THE_CONE_SEA_1, with some of the extra cones removed.
Did it work? - Yes. I ended up with a legit-looking autocross course with some nice sweepers and varied slaloms. I think it flows nicely, keeps you in second gear (in the XF GTI at least), and has some challenging sections with plenty of line options that you can get a little bit wrong without necessarily plowing through a bunch of cones.
Course 3 - Accessory Drive Belt Experiment 1
AU1_BELT_F150.lyt
What I was trying - I have a 2001 F-150 5.4L and I noticed the serpentine belt diagram under the hood and thought it looked sort of like an autocross course from the SCCA's design guide. So I snagged this diagram, measured the pulleys with an advanced imaging tool (MS Paint), assigned pixel diameters to meters (30m per 80px), and fired up LFS.
Did it work? - No. It's pretty bad. It ended up with a lot of tight sections and lock-to-lock turns that really suck to drive. If I doubled the radii and distances, it would probably still be terrible because it doesn't have good flow. You'd have the same steering experience, just at slightly higher speeds. It might get better if I spaced out the pulleys a bit more, so that the turns are more like 90-135 degrees instead of mostly 180-270 degrees.
Course 3.5 - Finalized Accessory Drive Belt Experiment 1
AU1_BELT_F150_FIXED.lyt
What I was trying - I took the same circles from Course 3/AU1_BELT_F150.lyt and moved them around so the turns aren't as tight.
Did it work? - Yes. The final product is a short, decent course that is actually drivable. I only added one cone, which was to create a slalom out of an otherwise boring section towards the end. So, are serpentine belt diagrams a good source of autocross inspiration? Maybe. It might not be a bad way to start, but expect to move everything so it looks nothing like what you started with.
Course 4 - Tire Rack Autocross 1
AU1_TireRack.lyt (template)
AU1_TireRack_Autocross1.lyt
What I was trying - I wanted to try recreating a course I had driven in real life. I did my second autocross at Tire Rack in South Bend, Indiana with the BMWCCA in my derpy $500 Ford Focus ZX3. It's approximately parking lot sized, so using chalk lines, InSim circles, and Google Earth, I drew the test track and slapped some cones on it.
Did it work? - Yes. I had drawn a sketch of the course after I walked it, and even though it's kind of a chicken-scratch drawing with no distances, I get roughly the same times I saw in real life: ~46.2 +0 (Focus) vs. 44.78 +0 (XF GTI). It was a really hot day (at least 90F -- the Focus started stalling at the line if I didn't give it some throttle in neutral) and the XF time might be with my stiffened autocross setup. Anyway, that Tire Rack template is pretty sweet!
Course 5 - Tire Rack Autocross 2
AU1_TireRack_Autocross2.lyt
What I was trying - I wanted to see how good my TireRack template is for throwing together a decent course quickly.
Did it work? - Yes. It probably took me about 10 minutes to throw down some cones and maybe 15-20 minutes to make adjustments after driving it. It's not bad! TireRack's track provides helpful constraints with excellent versatility. It turns the problem of designing a course from playing God on a blank canvas to arranging the elements of a preexisting course.
Course 6 - Grissom Day 1
AU1_Grissom_Day1.lyt
What I was trying - I drove at an autocross event at the Grissom Air Force base a couple weekends ago, and I wanted to create that course.
Did it work? - Ehh kinda. I was working from memory, and Google Earth says the concrete pad at the base is a bit bigger than any of the parking lots in LFS. I tried to recreate the feel of the course, but the proportions aren't right. It was a pretty tight course and I felt like I couldn't achieve the feel of how tight it was without shrinking it a little. The visuals are also probably confusing. Since it's a bit too short, the XF GTI beats Corvettes!
------------------------------------------
Now go buld a tark!