Weight Bias - Redux

[Shiny Side Up!]

As far as the SWR thing goes - I doubt you'll find much about it anywhere but here.  I certainly doubt that Turn 10 would own up to it.  I came up with it mostly like how that dude found Neptune... he never really saw it, but everything around it acted like it was there.

Yes, I understand... I have seen it discussed on this forum and in the context of the tuning calculators, but hadn't paid it much mind since I was happy in my own little world.  However, I find that the technical discussions in most of the FM community is severely wanting, so looking into these alternative approaches is a welcome brain workout!

[Silvo 1981]

I've been back through and gathered information on car type and body style. The information regarding engine position and drive type were already shown. I also collected information on the engine type (cylinder and aspiration) as well in case this showed up any patterns.

I've also tested all the data I collected to highlight any errors and to check the accuracy of the formulas and all seems good. I also double checked through the additional data (mentioned above) to some of the cars when I started to look at the groups as some of them were split over more than one category. These were German Touring Car, GT2, GT3 and Super Car.

I tried to look for patterns within these groups but I struggled to find anything obvious. I thought I had spotted a division in drive type in GT3 but its not clear cut. While there are a lot of RR Porsches in X and a couple of AWD cars there are also some AWD GT3 cars in Y. I also double checked through the spring data and addition info for the German Touring Cars and all was correct. I had a quick look to see if the two GT2 cars outside of X were correct and they were also. The data all seems true but I cannot quite put my finger on a definitive pattern.

Before I had gathered all the additional info I had been trying to use this initial data to tune my cars to see how the different setups behaved. I used an A600 Camaro '10 initially on Sebring Short. I used the cars published Wb to set my tune and saved the setups for W, X and Y (I had already assumed Z is for stock cars only). I then started with W and it was better than the tune I had before but it suffered on the bumps and kerbs. I then tried Y and found it took the bumps better but it was a little too soft for me. I then loaded X and consistently equalled or bettered the ghosts of the other two setups.

With the additional data I gathered I would say that the setups W, X and Y represent smooth through to uneven track surfaces and Z represents the stock cars. The published Wb is something that I agree we should probably just ignore and work out ourselves what feels best. I am a little bit unclear as to the effects messing with this but nothing a little tinkering about won't sort out.

I've attached where I am with my spreadsheet for any that are interested. I had to compress it as it was slightly over the limit. I've accidentally missed the additional data for 3 cars when I was gathering the info and haven't got access to my Xbox at present. The file is split over three sheets now cars, calculator and constants. I know the car sheet is looking quite busy but I’ve been hiding the columns I’m not using if it gets a bit much.

[Blooze]

That all looks pretty good Phil.  And a lot of work!!  Thank you for sharing!

I too have been unable to discern any tendencies with the body type/style information.  It's there, I just know it is - we probably are missing a couple key elements that they have hidden.  

I've been working on a solution where it probably doesn't matter.  Attached is my revised Tonka Toy functions, well, the work sheet, at any rate.  As it shows, you can choose any SWR to any degree of accuracy that you care to, and you get the associated Total ARB and Rebound values back.  Getting too squirrelly, like 46.17569%, is all for naught because the game can't use the results.

Putting these to work would go something like this...

You have three data entry points - the SWR, the Sb (Spring Bias), and the Curb Weight CW

Oh BTW, just for legacy's sake the value is Spring:Weight ratio.  You are actually calculating a Weight:Spring ratio, so all your numbers are an inverse of mine - lest there be some confusion.

To get the Total Spring simply multply the CW by the SWR.

TS = CW x SWR

To calculate the Front Spring...

FS = TS x Sb

...and the Rear Spring

RS = TS - FS

To calculate the Total ARB (TA), plug the SWR into a function like this...

TA = IF($C$15=0,0, ROUND(-1*$C$9^2+29*$C$9+6.63,2))

where $C$9 represents the entered SWR.

Now, to get the front ARB multiply the total by anything you want... 55% gets you what they are doing this year - last year it would be the inverse of the Sb...

FA = (1 - Sb) x TA

and the rear, again...

RA = TA - FA

It is another squirrelly function for the Total Rebound (TR)...

TR = =IF($C$15=0,0,ROUND(-1.58*$C$9^2+24.9*$C$9+4.226,1))

And the same treatment from there.

FR = TR * Sb

RR = TR - FR

To get the bump use the flavour that suits you, somewhere between 50% to 75% to be Kosher...

FB = FR * 50%

RB = RR * 50%

Once plugged into a spread sheet, these work pretty effortlessly.  I have a worksheet for every car I build, and therefore a record of its settings.  The Worksheet is a Boiler Plate, so I just copy files now.

The calcs still need a bit of tweekage... there is a slight bow in the middle of each and I am having the deb-bil's own time beating 'em down...

$

[Silvo 1981]

Where did you get your squirrelly functions from? I've got something similar using a scatter graph and a trend line to calculate the TA and TR for any value of SWR. The thing is, I've only got four points of reference (W, X, Y & Z) so I have no idea if the readings in between are accurate or not. Have you got any more values from the production cars for instance? I've not looked at those yet if I'm honest.

[Blooze]

Tonka came up with those last year using just the method you mentioned.  I have no idea about the mechanism for coming up with a trend line function from a graph (Google was not so friendly) so I just messed with Tonka's functions till I got the line close.

I created the reference points - I used 1% increments from 34% to 66%, then entered the values we had, and calculated for the others.  (the difference between the two point values) / (the number of increments between them).  Now add this amount to each previous point.  The line I created from the four points had a very slight bend to it in the back half, partly due, I think, to the smaller number of increments between the two points.  The back half was made up with two segments, the front half just one.  I put both lines on a graph and they were pretty close to straight, so I decided to build lines using just the begin and end.  I think it would have been better had I used .5% increments - I may do that later today.  You can see the work on the spreadsheet for the second line.  The ARB value is at A6 and its line is in column C.  The Rebound is at A9 and its column is G.  Anyway, the second line was very, very close to the first, only without the the little bow in the back third.  So I reasoned that if I came up with a line that split the difference of the arc of the bow I would be pretty close, at least close enough for the level of accuracy used by the game.

Then, on to Tonka's functions.  I reasoned that they were actually 3 expressions added together, each with their own constant.  So I exposed the 3 constants and started messing with the values untill I got a line that laid over the one that I created.  After a while I figured out that the first expression moved the top one way, the second the other, and the third offset the entire line.  So it didn't take long after that.  Columns D aqnd H show this work.

I started out with graphs, but that was too laborious and I found using a diff column was faster.

That's pretty much it.  Not too elegant, I know... but we gotz work with what God gave us  

As for the production cars, I have pretty well shined them on.  I personally think a number of their settings are desinged to make them all semi drift cars, but that is a whole 'nuther conspiracy theory.  What my plan for them (production cars) is to assign an SWR from somewhere in the first half of the line.  I used a 48 (with a 52 bias) on my D class Acura Integra and it is working fine!  So far I have improved every time I set with the stock configuration.  I have the two setups saved.  From the telemetry I see no tendency to bottom out, nor do I see excessive roll.  In fact, the car doesn't bang the ground on the uphill side of that one tunnel on Kaido like it used to.  I see sparks fly all over from my AI competition.

Unfortunately the Season Moderator (starting 5) keep herding me towards R# cars so I haven't had much time to test.  

Oh yeah, I did use the 50% thing for the Bump instead of the current vogue of 70%-ish.  Next is to see what happens if use the Bias to govern the ARB rather than that infernal 55%.

$

[Silvo 1981]

Well in that case I think I've stumbled across a better formula to model the 4 points we have. I'm just tidying it up at the moment but it hits the all the points accurately and looks to model the lines in between as best we can hope for. I'll put it up as soon as I can and hopefully it will save you trying to crack something Excel does for you.

[Blooze]

...it will save you trying to crack something Excel does for you.

Oh heavens! 

Have I missed something obvious?

[Silvo 1981]

It wasn't that obvious or at least I don't think it is. I knew that you could get a trendline on a graph but I didn't know excel could tell you the formula for that curve. It's under one of the options and is a tad handy for this. I matched up the curve type as best I could with the various options by eye and it gives you the formula for the two relationships we need. I've attached my spreadsheet with the graphs and I have worked in the formulas into the calculator I was using on the other sheet.

I've foolishly got Vista at home so I have saved it in 97-2003 format so everyone else can see what I've done. Let me know if anything isn't showing up as it should and I can explain if needed.

(Removed worksheet as references screwed - see next post for fixed version)

[Blooze]

Boy that is slick!!  Under which options?

the 55% line in your calculator has a busted reference so I can't see how your calculator works.

Great Job!!

I'm going to plug these into my deal and see how they look.

$

[Blooze]

Whups, what have I done wrong here...

Here is my function...

=-268.18*$B3^3+466.62*$B3^2+230.08*$B3+51.204

where $B3 is pointing at 34%.  The result is 172.83  I was expecting something along the line of 16.37.

  $

Oh, the #REF error is looking for a spreadsheet that has data out in cell BA3 - nothing you sent goes out there.

$

[Silvo 1981]

I deleted the cars sheet to keep the size down and still had some references kept in by mistake. Shouldn't matter that much for the formulas you want but I have sorted it out and uploaded it again.

If you're looking for the trendline formula option it's simply called "display equation on chart" and is under the trend line options (tick box). I'm using Vista currently but when I was looking at it yesterday on my machine at work it was pretty similar. Works pretty good though don't you think. Now just to find the right figure for the SWR and spring bias. Easy!

[Blooze]

I think it works great!

I found my mistake...

=-268.18*$B3^3+466.62*$B3^2+230.08*$B3+51.204

When it was supposed to be...

=-268.18*$B3^3+466.62*$B3^2-230.08*$B3+51.204

I do believe I'll spend some time updating the car sheets tonight.  Have to bump the Version counter...

Thanks Phil!

$

[Silvo 1981]

Is it likely that any values used for the SWR would be outside the range 31 to 69. The formula for the AT might be a bit off if it were likely. I figured probably not but thought I'd ask.

I hope you don't mind me asking but what are the affects of altering the Sb and the SWR. I know it's probably written in the initial thread (which I did read through once) but it's quite a lot to trawl through, if you don't mind pointing me in the right direction.

I think the effect of increasing the SWR is stiffening pretty much everything for smoother tracks/surfaces and quicker changes in direction but that's a bit of a guess. I'm not quite sure the effect of altering the Sb but I guess it's going to make the car feel more like the drive types they were trying to emulate.

Is it now just a case of playing about and trying to find a combination that I like the feel of and seeing how it feels on multiple cars.

[Blooze]

Is it likely that any values used for the SWR would be outside the range 31 to 69. The formula for the AT might be a bit off if it were likely. I figured probably not but thought I'd ask.

I don't think it matters if the game has SWR set at, but more what we set it at, and I plan to stay within those ranges.  Interestingly, the two cars I've built with this were quite different to my way of thinking, one a light FF, the other a quite heavy FA.  Yet...

• the SWR's were identical, 51.33% - a number associated we know to associate with low end R2 cars.
• the Sb's were identical, both 50% on cars the had a Wb of 62%
• the ARB values were identical, and off the scale for stiffness.  Again, both set to a 55% bias.
• the Total Rebound were identical, 54% (or Sb + 1??), 17 for them both.  The TR was in line with the SWR; however, the bias was 70-ish, rather than the 50% of the Race cars.

Now, the first thing I have deduced is that we've been right all along.  The Sb is the real Weight Bias, and setting it has direct affect on the balance of the car.  It is basically, the battery in the trunk, or lead billets in the frame rails.  The engine placement has something to do with it, but I haven't figgered it out yet.  Mostly, with RR and MR cars a Sb of around 48% seems to work good - FR and FF cars prefer something in the order of 52%.

The ARB however - very close to 2.66 x TA.


Okay, here's my conspiracy theory.  The high SWR, the 70% Rebound to Bump, the almost tripled ARB are all designed to make the cars easier to drift.  They wanted to make it easier for newbies and non-tuners to drift.

I hope you don't mind me asking but what are the affects of altering the Sb and the SWR. I know it's probably written in the initial thread (which I did read through once) but it's quite a lot to trawl through, if you don't mind pointing me in the right direction.

I think the effect of increasing the SWR is stiffening pretty much everything for smoother tracks/surfaces and quicker changes in direction but that's a bit of a guess. I'm not quite sure the effect of altering the Sb but I guess it's going to make the car feel more like the drive types they were trying to emulate.

Is it now just a case of playing about and trying to find a combination that I like the feel of and seeing how it feels on multiple cars.

That is essentially it.  Look over the opening post to this thread - it is a synopsis of the whole deal, I think.  What it gives us is the ability to fine tune the action and balance of our cars, specially with a spreadsheet.  I have mine set to two decimal places and have been know to make adjustments in the hundredths.  Truth be told, a tenth is about as close as we need to get.

I have been starting with a 48/52 build (SWR/Sb) and modifying from there.  

I still haven't heard anything definitive on the ARB and the 55% or the reason for it.  I have gone to an inverse Sb on both of my weight forward builds and it is working good, and as I expected.

$

[Silvo 1981]

The reason I mentioned the range of the SWR was that if you were working on a R1/R2 and wanted to go above 66.32% SWR or a nascar below 33.94% SWR then the formula for the AT does start to go a little off track past 69% and 31%. If you go lower than somewhere around 31% the strength of the AT starts to increase rapidly instead of decreasing gradually like it had been. I would imagine for the average car you would be working on the SWR would fall well within that range but if you were working at the extremities of the SWR it's something to bear in mind. I'll show you a graph I did of it tomorrow to show you what I'm banging on about. A picture paints a thousand words and all that.

I hear you loud and clear on the drift setup conspiracy. It makes sense to me at least. I don't know that they've been setup completely for drifting but it must have been in the forefront on their minds when setting up the average stock tune.

I'm going to take some time and see what tinkering about comes up with for me on this but it's useful to have your thoughts on these settings.