Correlation between Performance Points and price

[Memesbro]

I want to see how strong the correlation between Performance Points and the price of the car is. It may seem obvious that cheaper cars are slower and faster cars are more expensive, but there are a lot of edge cases where a car is either a lot cheaper or a lot more expensive than the surrounding cars.
For example, the Ford Mark IV, Ferrari 330 P4, and Jaguar XJ13 are three of 5 cars in the game that cost the maximum 20 million credits (The other two are the Red Bull X2010 and X2011 Prototype), but these race cars, while fast, have a lower PP value than the NASCARs, which cost 40 times less at 500 thousand credits.
Another example is the Volkswagen Schwimmwagen, which has the lowest unmodified PP of any car in the game, costs more than the previously mentioned NASCARS at 625 thousand credits.
While there are other factors that balance out the price and PP, such as rarity, or how novel it is, I still think this is interesting to look at. Also I want an excuse to add onto my GT5 Car List by PP and want to find the best valued car for price

 

[Dave A]

To find the best value car for the price simply divide the price by PP. I think it'll be hard to beat the Daihatsu Midget II for that though, but I haven't done an in depth analysis.

 

[eran0004]

I guess you can plot it in Excel and draw a trend line.

 

[Memesbro]

To find the best value car for the price simply divide the price by PP. I think it'll be hard to beat the Daihatsu Midget II for that though, but I haven't done an in depth analysis.

Its not that simple. Otherwise the Red Bull X cars would have a stupidly low score, possibly lower than the Citroen 2CV, which is the worst standard car in the game (The two worse 'cars', the Schwimmwagen and the Sambabus, are premium cars). I think the equation should be more complex to not rule out higher end race cars and not favour cars that can't get you into the later events

I guess you can plot it in Excel and draw a trend line.

That is what I am planning to do, but first, I need to put down the price of every car in the game

 

[Dave A]

Well, it's however simple or complex you want it to be. At its core, the best PP per credits is the car that costs the least Cr per Performance Point. If you want to include how much PP it can reach and equate a scaling value as the cars PP rating increases that's more complex. But it doesn't answer the fundamental question of the most PP for your credits.

The problem with creating a sliding scale where PP's should become more expensive as they increase, is you set the scale therefore you control the results. So the test will only confirm what you choose for it to confirm. I.e. you determine what it will confirm based on your scale. A different scale will produce different results.

And, shouldn't the Red Bull cars have a stupidly low scale with them costing 20m Cr as they do?

 

[eran0004]

Its not that simple. Otherwise the Red Bull X cars would have a stupidly low score, possibly lower than the Citroen 2CV, which is the worst standard car in the game

You need to define value. One way could be to count the number of events that the car is eligible and suitable for. Another could be to sum the prize money of the events the car is eligible and suitable for. A third way could be to give your own subjective bias to the PP range by assigning weights. For example, if you most enjoy driving cars around 500 PP you could assign 500 PP a weight of 2 and then you interpolate linearly between 1 and 2 for ranges within 100 PP of that. The score would then be weight × PP / price.

The more complex your function is, the more trouble you will have explaining and defending your choice of method and your results, so think about what exactly you want to look for and why and then you can try to design a function that will attempt to do that.

 

[Memesbro]

What I am planning on doing is this: At the lowest point (200 PP), I will put the weight (What I will multiply the PP by as we go up) at 1. I will then increase that weight by, say, 1 every 50 PP. I.e. 200 PP weight is 1, 250 PP weight is 2, 300 PP weight is 3, and so on. That way, stock race cars don't have the chance to be shafted by stock kei cars. (Oh, I forgot to mention is that all the cars will have their PP as when they are new and unmodified. So no upgrades)

 

[Dave A]

What I am planning on doing is this: At the lowest point (200 PP), I will put the weight (What I will multiply the PP by as we go up) at 1. I will then increase that weight by, say, 1 every 50 PP. I.e. 200 PP weight is 1, 250 PP weight is 2, 300 PP weight is 3, and so on. That way, stock race cars don't have the chance to be shafted by stock kei cars. (Oh, I forgot to mention is that all the cars will have their PP as when they are new and unmodified. So no upgrades)

But that's really skewing the results, is there a specific reason for that weighting other than you want the higher pp cars to score better?

I think I have GT5's car database extracted as a csv file somewhere. I might check (as long as its not from a modded game) and see what results are possible. A stepped approach as you suggest however crest cliff edges for cars just before each step up. If you are insistent on a scaled value to PP, I would recommend a scale that increases fractionally for ever PP.

What I feel like you are looking for, is a scale that gives the cars you want to be the best value, best value.

 

[eran0004]

What I am planning on doing is this: At the lowest point (200 PP), I will put the weight (What I will multiply the PP by as we go up) at 1. I will then increase that weight by, say, 1 every 50 PP. I.e. 200 PP weight is 1, 250 PP weight is 2, 300 PP weight is 3, and so on. That way, stock race cars don't have the chance to be shafted by stock kei cars. (Oh, I forgot to mention is that all the cars will have their PP as when they are new and unmodified. So no upgrades)

With those weights you will have some discontinuity issues. A car that sits at 400PP will have an unfair advantage over a car that sits at 399 PP due to the sudden change in weights.

A better way would be to have a continuous weight function W = f(pp).

For example:

W = f(pp) = max(1, (pp/50)-3)

This function would give you a weight of 1 at 200 pp (and below), it would increase by 1 for every 50 pp above 200 and cars that are close in PP would also be close in weights.

 

[Memesbro]

But that's really skewing the results, is there a specific reason for that weighting other than you want the higher pp cars to score better?

I want higher PP cars to score better because on an unweighted scale, cheap ****boxes (Such as a Vitz F for 9,800 credits) would be way better than race cars (Such as a Nissan GT-R Concept LM Car for a little under 1 million), even though stock, the cheap ******* is practically useless in the long term, while the race car can compete at top end events.

With those weights you will have some discontinuity issues. A car that sits at 400PP will have an unfair advantage over a car that sits at 399 PP due to the sudden change in weights.

A better way would be to have a continuous weight function W = f(pp).

For example:

W = f(pp) = max(1, (pp/50)-3)

This function would give you a weight of 1 at 200 pp (and below), it would increase by 1 for every 50 pp above 200 and cars that are close in PP would also be close in weights.

Nah, if a car is at 399 instead of 400, thats a skill issue, lol.
But honestly, this is a good idea, and I cannot believe I didn't think of that

 

[Dave A]

I want higher PP cars to score better because on an unweighted scale, cheap boxes (Such as a Vitz F for 9,800 credits) would be way better than race cars (Such as a Nissan GT-R Concept LM Car for a little under 1 million), even though stock, the cheap *** is practically useless in the long term, while the race car can compete at top end events.

That's fine, but that's not going to tell you which car offers the most PP for the Cr. You need to define what useful means and how you'll scale that properly, not just decide you don't want cheap cars to be better value for money simply because you don't like that outcome and scale in favour of high PP cars. That's simply creating a more skewed outcome.

The earlier suggestion of working out how many events/races each car can enter and creating a scale based around that has some merit, as that's a logical and measurable way to determine usefulness. A car that's realy fast but can only enter one event is no more useful than a car that's slow and can only enter one event.

You may want to consider how close to the events rules a car is as well, i.e. A 350pp car may be able to enter a certain event, but it'll be massively outclassed. So then you need two scales, one that will count how many races a car can enter and another that will determine potentially how compete I've that far will be, albeit in stock form only.

An alternative method that won't only favour cheap cars is to find the car with the lowest PP rating and count that as 1 PP and the essentially deduce that cars PP rating from every car as part of the calculation.

As I said at the start, it's an interesting idea, but to do this in a way that will generate any meaningful consensus is going to be your challenge. If it's purely for you and your benefit ok, ignore all of the above 😂.

 

[Memesbro]

As I said at the start, it's an interesting idea, but to do this in a way that will generate any meaningful consensus is going to be your challenge. If it's purely for you and your benefit ok, ignore all of the above 😂.

I mean it is my spreadsheet, using my previous spreadsheet on the PP of every car in the game, because my mind yearns for a useless spreadsheet that only keeps me busy.

I'm just making this post because I thought it would be interesting for others to look at