Or, statstical SHALB analysis.

SHALB stands for Speed, Handling, Acceleration, Launch, Braking; the five categories in which every car is valued.

All of this started because I read this thread: Are racing tires overkill for a B class car, more specifically DesigningLeek47’s third post within that thread. In that post DesigningLeek47 listed good examples of how many “hit points” are satisfactory for classes within each category. I call them hit points because they’re a bit like the set of points one would distribute to assorted character attributes in a D & D type game.

The gears started cranking in my head and I recalled downloading ManteoMax’s Forza Motorsport 4 data spreadsheet awhile back. I opened it, pulled out the selection of data I needed and saved a smaller separate spreadsheet. Each piece of information I have is separated into its own column and every car is set into its own row. Before I go further I must thank ManteoMax for the hard work.

The first thing I did was separate the classes and place each on their own tab. I then ran the five ems (min, max, mean, median, and mode) for the performance index and the SHALB columns of every car. This gives us an excellent breakdown about how many hit points are “too hot” “too cold” and “just right” for each category within each class.

While looking at the individual scores I was doing some quick math in my head and noticed cars have varying amounts of total hit points relative to their performance index rating so I totaled the SHALB values for each car. The final step I did was divide each car’s total SHALB points by its performance index.

Exactly like baseball’s batting average, the quotient is a value between zero and one; it’s a representation of hit points per performance index point, though unlike a batting average the better cars have values closer to zero. To phrase it other ways: A car with a lower hit point total relative to its performance index is a car efficiently extracting performance from its ratings. A car with a higher hit point total relative to its performance index is a car inefficiently extracting performance from its ratings.

What can we do with these figures? Not too much but that which we can do is fairly important. This is a good (not excellent, fantastic, or perfect yet) method to compare built cars, especially for people who build cars to real world specifications and don’t follow the in-game performance index rating. Understand that it is possible to make a B 439 car competitive with a B 490 car. Anyway, sort these SHALB quotients in ascending order and they can rank a group of cars from best (top) to worst (bottom) performer.

I am working on a way to include the important power to weight (P:W) ratio into this rating. At this time I don’t know exactly how though I have some ideas. One method includes a P:W rating contained in the quotient, another method keeps P:W separate. Either way, it’s possible to use P:W to balance the field. Example: we have seventeen cars legal for competition. We could cut the list into six top cars, five middle cars, and six bottom cars. Meaning, give the bottom cars a bit better P:W rating, hamper the top cars’ P:W rating, and hopefully the group tightens around the middle.

How do I rate power to weight? (horsepower + torque) / weight. Use those parenthesis to maintain the order of operations. Ideally we could find the area under the power curve(s) though that would be a TON of work. It would mean breaking out the Forza Kitchen Sink spreadsheet, graphing every car’s power curves (driving each car, saving the replay, watching the replay, entering torque figures), and then mathematically finding the area under the curve. We’d need to do this for every car, stock and modified, and for every different way any given car can make power. All of that work would be worth it because we’d have one value each representing every car’s power through all RPM, not just peak numbers. It’s an awesome thing but is man-hour exhaustive.

We could enter a cluster of SHALB quotients into a graphing calculator to perform further statistical analysis. Of course, using statistics means we can observe sigma and [referring to the list of seventeen cars above] might find we have four top cars, six middle cars, and seven bottom cars. Translation: we can more accurately balance the field because we could determine the varying differences between quotients and sigma.

I am interested in seeing where else this can go. This is Forza Motorsport 4 only so far as I have not yet bought Forza Motorsport 5, though I’m sure SHALB / PI would work there too. What other ideas do people have? Also, could this formula be worked to find which cars are “leaderboard cars” without any driving … ?