Skillpoint calculation formula

[rickostidich]

You might be interested in knowing what I discovered by personal experience.
The number of SkillPoints that you achieve for each Official Time you race, looks like it is calculated by:

(number of players that have completed an official time on that track) * 60 / (your world ranking position)

Then, the result is rounded by some amount, but this is the general logic.

I don't agree with this system, because as noted in many forums, most people only work on the first 4 tracks to gain a huge number of points, and people are discouraged in gaining access to the most difficult tracks - which are definitely the most beautiful tracks of the game! (A01-A04 are sooo boring!!)
In the entire Italy, only me and a couple of other players have mastered the Nadeo on the 5 black tracks, and this is really a pity.
What I would have preferred, is to have a fixed quotient for all tracks, for example SP = 1'000'000 / ranking, so that all track have the same importance: if you are one of the few in doing an official race on a black track, you will gain a better ranking due to the fewer number of players, but in that case you *deserve* a lot of points!
With my system, if you happen to be the best in the World on *all* 65 tracks (!), you would gain 65 million points.
This would be very easy to understand by all players, and it would provide a much more competitive SP challenge.

However: with the current formula, if you make an official race and you obtain the very worst time in the World, you gain 60 SP. Instead, if you obtain the very *best* time, you gain 60*Players points, which for the first 4 tracks means around 2 or 3 millions. On the E04 track (the hardest for me - my 260th medal!), it appears that only 477 players in the World have raced an official time, so you can obtain at maximum 28'620 SP.
(Only *141* players in the World have completed race E05 - easier but veeery long, but as noted in my other post, many other players could have my same problem in uploading a 13 minutes replay to the server!)

If you have in mind the graphic of a formula y=a/x, you can imagine that you will start to obtain *a lot* of points when you approach the ranking position = square root of the number of players. Below that ranking, you always obtain few points, close to the result obtained by players with a ranking similar to yours. Above ranking=players^0.5, you get a huge number of points each position you gain. Of course, the player #1 has twice the score of player #2 and 3 times the one of player #3.

For the most played track, A01, it looks like the number of players is currently around 52'038. So, as a reference to obtain a lot of points, the player position to win is 228 (which is the rounded square root of 52'038).
I'll use this technique in order to see how much below (or above!) I'm from the reference position, and to decide which track to concentrate on, in order to gain SP ranking.

If you want to know an approximation of how many players raced on a specific track, simply multiply your current skillpoints number by your world ranking position, and divide by 60.
For example, on track C03 I'm currently 6th in the World (sigh! one month ago I was 3rd!), and I have 174'617 SP.
6*174'617/60 = around 17'462 players raced officially on that track.

Last note: I find that with the current system of scoring, the most fun part of the game is the solo / medals challenge. It is the only one where good driving is really of importance. With the skillpoints you are limited to too few tracks, and in the multiplayer sessions the ladder points are based more on how much you play instead of how good you do. Moreover, though the user tracks are generally wonderful and very fun, it's more important how much you know by memory that specific track, compared to how well you drive. Good LP winners should download and study each track on a server, and always use that server to play.

Now that I obtained all the 260 medals, I'm a bit sad that the best part of the game is gone.
I'll improve my skillpoints on every track, but it's not the same fun due to the poor scoring system.

Luckily, I didn't experience track-building and replay-editing so far, so I still got a lot of things to do with this game!
Definitely the 20 euros best spent in all of my life!!

Rick Ostidich

[riolu]

Well, I think that the current way of skill point calculation is already good. I mean, of course it is a pity that you get 32.300 SP only for the world record on a hard track like E04, but it is pretty logical (at least for me) to get more points the more players you beat on a track, no matter if it's easy stuff like A01 or "hard" tracks like E03/E04.
I guess you have to remind that on "easy" tracks like A01, it is way more difficult to beat the world record time or to get a top10 time, because the track is short and every mistake can kill your run.
However, on tracks like E04 or even on some D tracks, the tracks are longer, therefore its nearly impossible to drive a clean run without big mistakes. So as a result, it is easier to drive the world record or a top10 time, so you "deserve" less SP for your run than on A or B tracks.

[rickostidich]

The goal is to have people play on all of the tracks.
Currently, 99.5% of the drivers have never played on a black track, and probably never will.

In order to unlock the track E01, you have to achieve the gold medal on all previous maps, and in order to make an official race, you have to win the gold medal on that track.
If you make an "average" official time on that track, you have already "beaten" all the people who never tried to!

Moreover: making the best time on an easy track, means that you have to to a perfect race, for example to press the brake button at the exact millisecond, for the exact number of milliseconds - that could also be achieved by programming the joystick with any mapper or hotkey software - not fun and not that "honest".
Instead, on the long and difficult tracks, also the best time is far from perfect, so what really matters is the average quality on the entire race, thus: your driving skill.

If you are Mr. Riolu, the one who is #1 in the SP World ranking, my best congratulations and admiration to you - I always play with your (or Marco's) ghost time as reference!

And I've checked - yes, also with my system of scoring, you would still be #1 in the World!

Bye!!
Rick

[metoxys]

If you are Mr. Riolu, the one who is #1 in the SP World ranking, my best congratulations and admiration to you - I always play with your (or Marco's) ghost time as reference!

And I've checked - yes, also with my system of scoring, you would still be #1 in the World!

Yo, he really is riolu

[Jadis]

Let x be the amount of SP for rank 1.

Generally: (number of players per map)*100-100=x

That means for rank n:

(x-100)/n=y

(y being the SP of n)
which leads us to the conclusion:

x=yn+100 and n=(x-100)/y

The skillpoints calculation system is the same in TM2 as well as in TM1, the only difference is that in TM2 it is 10 times less likely to equal someone's record thanks to the more accurate timing.

[riolu]

the only difference is that in TM2 it is 10 times less likely to equal someone's record thanks to the more accurate timing.

Yep, that makes the ladder even more interesting
Nevertheless, I noticed that a .xx9 time is impossible, maybe there should be a fix for that soon..I've checked more than 200 replays on different tracks and I never had a .xx9 millisecond ;o

[phil13hebert]

Yep, that makes the ladder even more interesting
Nevertheless, I noticed that a .xx9 time is impossible, maybe there should be a fix for that soon..I've checked more than 200 replays on different tracks and I never had a .xx9 millisecond ;o

Interesting, I never noticed that while playing in multiplayer servers

[Weeeeee]

Not impossible at all,but it does indeed seem to happen less often,here is 2 .xx9 times from a track on mx; http://tm.mania-exchange.com/tracks/9581/twl0-catrace

[riolu]

It's still weird that .xx9 times are so rare..I guess Nadeo should check that

[rickostidich]

Jadis: thanks for your information!
I agree that the multiplier for the number of players is 100 and not 60 as I was guessing: I was considering that most of the people who had ever played an official time on a whatever track (TM2 says 54'140 people as of today), would have at least played on track A01, but this is not true. With my new calculations, only 31'985 raced on A01 as of today.

However, your formula is not the exact one as used in the game.

In the game, for each track I can see the name and ranking (some could be equal ranking) for the best 9 players in the World (plus me), and from the World SP ranking list: I can see, for each of the best 9 players, the SP they achieved for their 10 best-score tracks.
Moreover, of course I know my own rankings and SP for each one of the 65 tracks.

I combined all of these informations in a spreadsheet, and verified your formulas on all data that I have.

You initially said that the number of skillpoints available for each track is x=players*100-100.
That is equal to x=(players-1)*100, and it makes sense since it could be read: each player (except the best one) gives 100 SP to the best player.

Then, you said that the skillpoints for each player is sp=(x-100)/rank.
I was wondering why to remove other 100 points from the available total, and what score should be given to the best player, be it =x, or =x-100.

From experimentation, I concluded that the actual formula used in the game is not (x-100)/rank, but x/rank-100.
So, finally:

skillpoints=(players-1)*100/rank-100.

Still I don't know why they remove other 100 points, but it works always like this, except in 2 cases:

One exception is the very best player, that sometimes takes the correct x-100, but often x-200, and sometimes x-300 or even x-500 (see A03). This could apparently be a matter of server synchronization, but it only happens to player #1.

The other exception is the case in which several players share the "exact" timing.
(The time is shown in the game as rounded to 0.01 s, but we all know that the measured precision is something near the millisecond - probably not exactly 0.001, and this could be the reason why rounding/truncation very seldom results in x.xx9).

Well, if the player #6 and #7 share the same timing (see for example 'devil' and 'galaxym' on A02), correctly they are both shown as ranking #6, and the next one is ranked #8. Instead, for the SP calculation, they are both considered to be ranking #7, thus the worst.
(This also applies to A08, where two players currently share #4 and other two are #7.)
This is the reason why the formulas didn't apply correctly to most of my rankings: on track A03 I'm shown as number 174, but I actually share the same exact time with 26 other players, so our common score is calculated as all of us were number #190. :[

This is a pity since I cannot calculate the exact number of players for a given track, unless I'm very sure that no one else in the World did my same time. The formula

players=ranking*(sp+100)/100+1

is exact only if I am in the 9 best places in the World and I can see how many people share my time.

Last note: out of curiosity, if two cars arrive at the finish line at 300 km/h, one millisecond of difference means 83 millimeters of distance between the respective front bumpers.

Bye!
Rick Ostidich