GT Sport on K' | Profiles, Events, Leaderboards, Stats, Cars | Feedback

  • Thread starter Milouse
  • 755 comments
  • 104,045 views
Thanks for the incredible work! I really like to browse cars on your hp rather than in game.
I have one suggestion though: With kg/hp being maybe one of the more important criteria to get an idea of a car's performance, is it possible to add that ratio like AC does?
 
A Statistics section: GT Sport number of players over the weeks, Players by Country, Final game players Vs Demo, Sport Mode players ratio, Campaign progress, Players by XP level, distance driven, Active players, Daily Activity in Sport Mode.
I've just published an update of those Stats, based on a new and bigger sample.
Interesting to see Japan, which i guess doesn't celebrate Christmas, is way down the share of players compare to previous data.
As a side note, around 9% of existing profiles on the official GT server are somewhat broken or blank. Those are ignored here.

@Spurgy 777 Profile data that are displayed on a given player page are not on K', they are requested by K' server as you ask them (so they are always fresh ;) ). I intended to include Wins, Clean races, pole position and fastest laps in the last poll, but this parts failed so it will be for the next poll.
Then i'll see what we can do with that.
 
This post is for those who try to get a better understanding of the DR system, like @PJTierney , @Famine , @Outspacer , @250Cal , @gapper59 etc.
Using youtube and some on-the-fly data parsing on K' server, i've just made a FIA race-result manual analysis of A to S Rank players.
Basically, i read every grid drivers profile before and after the race.
Sorted by their finishing position:
1. TX3_Devil | 62,339/99 -> 63,149/99
2. jfrod52 | 61,218/99 -> 61,916/99 @Jfrod
3. L1QU1D_G3N1U5 | 43,258/98 -> 44,514/99
4. AdriiRacing | 36,043/99 -> 37,430/99
5. GT_BOCA_ESCAPE | 37,791/99 -> 38,948/99
6. Cherep_GT | 38,383/99 -> 39,357/99
7. Chrisko_44 | 42,705/92 -> 43,350/88
8. GhostlyRazgriz | 37,297/99 -> 37,996/99
9. montero4610 | 53,884/92 -> 53,760/99
10. pK703 | 40,006/97 -> 40,280/99
11. FRoDoPoYoTe | 37,099/96 -> 37,332/99
12. FDP90 | 45,185/99 -> 44,936/99
13. dimgto | 46,727/99 -> 46,255/99
14. mthomas_95 | 58,162/99 -> 57,076/99
15. Palma___94 | 40,826/99 -> 40,271/99
16. Iceman_Berlin21 | 47,623/99 -> 46,640/99
17. adstomko | 46,457/99 -> 45,359/99
18. vinnycris | 49,235/91 -> 47,759/79
19. CSgt_Sheridan | 35,915/99 -> 34,891/99
20. Kyd_Stig | 37,210/92 -> 36,144/92

upload_2018-1-11_21-41-46.png


TX3_Devil video: (race start @ 1h31m3s) - not very interesting, as he did Qual P1 to finish P1.
 
Last edited:
Thanks very much for doing that @Milouse, it's exactly what we need! The FIA race gives a really good mix up as well, with people qualifying out of order much more than in daily races. I'll have to stare at this in a spreadsheet for a while, but the first thing that's popped out is that the total DR (sum of all players) is unchanged by the race.
 
I'll have to stare at this in a spreadsheet for a while, but the first thing that's popped out is that the total DR (sum of all players) is unchanged by the race.
It really looks like an ELO system, which would transfert points from every player to or from every opponent based on how the actual result challenged the expected result.
It would also match the fact that the better ranked P9 lost points as P10 won points.
 
It really looks like an ELO system, which would transfert points from every player to or from every opponent based on how the actual result challenged the expected result.
It would also match the fact that the better ranked P9 lost points as P10 won points.

It very much does, yes. I've used the formula from wikipedia and my calcs are pretty close to the actual DR changes, mostly within a few points. (At the moment I've got 41210 instead of 400 in the Q calc, and 157.75 for K, but these are not accurate and indeed might change depending on the DR strength in the race). Next thing to try out is round() vs floor() vs ceil() when turning each partial score into an integer... this step is what makes working out the correct constants to use difficult.

I can say that the 3 quitters did not gain/lose points between themselves, although of course they lost points to everyone who finished the race.

What's troubling me a bit is where do all these points come from? If a race just transfers points from player to player, the only input to the system is the little boosts that are given when moving up a DR level, including the one after 5 races to D. Are there enough points created in that way to match the number of players at each DR level?
 

Attachments

  • gts_fia_race_results.xls
    37.5 KB · Views: 20
@Outspacer, i'm feeding you with an [edit: two] other FIA races :P

Alphabetical order:
Alexony94 | 58054 / 99 (95 races)
bravi78 | 51710 / 97 (627 races)
donmaynardo | 50267 / 99 (312 races)
ferrari1974 | 46273 / 99 (88 races)
Francescoccx | 52710 / 99 (231 races)
Giorgio_57 | 65433 / 99 (154 races)
Leandrin34 | 45079 / 94 (238 races)
Melmen1187 | 45990 / 99 (422 races)
montero4610 | 51359 / 90 (230 races)
m_verstappen_33 | 55911 / 98 (225 races)
RGT_DAYRONE_08 | 75000 / 99 (283 races)
RGT_MIRAGE_69 | 59920 / 99 (179 races)
Ringkenner | 45708 / 99 (253 races)
Sch456mit789_-qw | 46130 / 99 (134 races)
serfonseca | 46319 / 99 (161 races)
TC-Touni95 | 51174 / 99 (922 races)
TSRacing_Zoky | 64008 / 99 (141 races)
UGO_SPOLDING | 57909 / 99 (76 races)
VP_Dofrer27 | 49036 / 99 (366 races)
WRT-SkyPikmin | 58070 / 99 (59 races)
->
Alexony94 | 58593 / 99 (96 races)
bravi78 | 50912 / 99 (628 races)
donmaynardo | 50004 / 99 (313 races)
ferrari1974 | 45533 / 99 (89 races)
Francescoccx | 53145 / 99 (232 races)
Giorgio_57 | 65995 / 99 (155 races)
Leandrin34 | 44864 / 99 (239 races)
Melmen1187 | 45582 / 99 (423 races)
montero4610 | 46932 / 83 (230 races)
m_verstappen_33 | 55742 / 99 (226 races)
RGT_DAYRONE_08 | 75000 / 99 (284 races) <= We see Dayrone capping at 75 000 pts
RGT_MIRAGE_69 | 60226 / 99 (180 races)
Ringkenner | 44675 / 99 (254 races)
Sch456mit789_-qw | 45238 / 99 (135 races)
serfonseca | 46691 / 99 (162 races)
TC-Touni95 | 51512 / 99 (923 races)
TSRacing_Zoky | 64945 / 99 (142 races)
UGO_SPOLDING | 58615 / 99 (77 races)
VP_Dofrer27 | 48982 / 99 (367 races)
WRT-SkyPikmin | 59085 / 99 (60 races)

Final positions:
upload_2018-1-12_18-50-51.png





Other FIA race with 12 racers:
Annihlator14 | 18413 / 99 (140 races)
camaras666 | 18477 / 91 (356 races)
Cityblueshr | 20244 / 99 (243 races)
frans333 | 18577 / 99 (177 races)
juanudo26 | 19339 / 99 (26 races)
Kevfitz1 | 20206 / 97 (128 races)
koyokdu70 | 19665 / 99 (626 races)
makiaveli69 | 19375 / 97 (162 races)
mambru_46 | 19356 / 99 (234 races)
RacerRay99 | 20041 / 99 (336 races)
robersinte | 19511 / 99 (126 races)
scarfy100 | 18972 / 99 (214 races)
->
Annihlator14 | 17879 / 99 (141 races)
camaras666 | 18100 / 99 (357 races)
Cityblueshr | 19348 / 99 (244 races)
frans333 | 18674 / 99 (178 races)
juanudo26 | 20215 / 99 (27 races)
Kevfitz1 | 20583 / 99 (129 races)
koyokdu70 | 20373 / 99 (627 races)
makiaveli69 | 19613 / 99 (163 races)
mambru_46 | 19118 / 99 (235 races)
RacerRay99 | 19944 / 99 (337 races)
robersinte | 20063 / 99 (127 races)
scarfy100 | 18266 / 99 (215 races)

Final result:
upload_2018-1-12_18-56-45.png


and video in case of need:
 
Last edited:
@Outspacer, i'm feeding you with an [edit: two] other FIA races :P

Thanks! I should be able to have a look a bit later tonight. The second one first - it might show that different constants are needed at B level.

montero4610 | 51359 / 90 (230 races)
->
montero4610 | 46932 / 83 (230 races)

Ouch. One does not want to quit/disconnect from S level race :crazy:

That's what i get for 5 or more races (770profiles out of 6108 profiles):
E 0.6%
D 74.4%
C 15.5%
B 9%
A 0.5%
S ?

Thinking about the problem of driver point creation... how many players have >5 races, that this breakdown corresponds to?
 
Thinking about the problem of driver point creation... how many players have >5 races, that this breakdown corresponds to?
I don't answer your question but shouldn't the average value of all drivers be 2000 if the input was just the level 5 initial points. I check that value from 1/100th of all existing players and found an average DR of 4366.

An other note: i found out that an unusual number of very low SR (14 pts or less) seems to have their DR points capped at 20,000 pts (which is middle-B). A few players are above, but i can imagine that the later tanked their SR after having gained a higher DR score. Maybe extremely low SR can block DR points ELO-like exchange.

I don't try to overwhelm you, but a last race for today:
Arno23 | 46577 / 99 (272 races)
AzoreanGordo | 62788 / 99 (288 races)
bravi78 | 50912 / 99 (628 races)
GTR-Ninja-20 | 44362 / 99 (73 races)
GT_BOCA_ESCAPE | 38948 / 99 (44 races)
GT_BOCA_GIAMP | 48266 / 99 (321 races)
icemannick37 | 51712 / 99 (829 races)
jamie-995 | 72367 / 99 (354 races)
johnny-thilmont | 55317 / 99 (246 races)
Leandrin34 | 44864 / 99 (239 races)
Palma___94 | 40271 / 99 (382 races)
PGMotorsport | 41640 / 94 (265 races)
RGT_DAYRONE_08 | 75000 / 99 (284 races)
Ringkenner | 44675 / 99 (254 races)
schlafhund | 39584 / 99 (312 races)
Thonyboy1 | 59686 / 99 (140 races)
turnupdaheat | 40391 / 99 (446 races)
TX3_Devil | 63149 / 99 (130 races)
VP_Dofrer27 | 48982 / 99 (367 races)
XxIl-NichXx | 61370 / 99 (239 races)
->
Arno23 | 46538 / 99 (273 races)
AzoreanGordo | 63372 / 99 (289 races)
bravi78 | 51174 / 99 (629 races)
GTR-Ninja-20 | 44730 / 99 (74 races)
GT_BOCA_ESCAPE | 40007 / 99 (45 races)
GT_BOCA_GIAMP | 49591 / 99 (322 races)
icemannick37 | 51308 / 99 (830 races)
jamie-995 | 72249 / 99 (355 races)
johnny-thilmont | 53727 / 86 (246 races)
Leandrin34 | 44257 / 94 (240 races)
Palma___94 | 41120 / 99 (383 races)
PGMotorsport | 41322 / 69 (266 races)
RGT_DAYRONE_08 | 75000 / 99 (285 races)
Ringkenner | 43758 / 99 (255 races)
schlafhund | 39029 / 81 (313 races)
Thonyboy1 | 57938 / 97 (140 races)
turnupdaheat | 40279 / 98 (447 races)
TX3_Devil | 64037 / 91 (131 races)
VP_Dofrer27 | 49004 / 95 (368 races)
XxIl-NichXx | 61850 / 78 (240 races)

The race was buggy for TX3_Devil who saw all others stuck on the grid during the whole race.
Anyway, the final positions:
upload_2018-1-12_20-16-0.png
 
I don't answer your question but shouldn't the average value of all drivers be 2000 if the input was just the level 5 initial points. I check that value from 1/100th of all existing players and found an average DR of 4366.

No idea :) The 2000 is the only 'pure' input, but I'm not sure how much to allow for the other boosts/drops that happen when crossing levels. We can assume that there are more boosts than drops, since everyone starts at E. Considering just the D->C transition brings us a bit closer to 4366... 25% of players with more than 5 races have managed that, with its gain of 1200 (?) points. But it still doesn't seem like enough, and only ~10% have managed the next level.

I might spot something when I analyze that B level race, but I suspect that any other input to the DR point pool would happen (or be most obvious) at D or C levels.

4366 is the average of players with >5 sport races, right?

edit: Points effectively get taken out of the system when a player stops playing completely ("done with GTS"), so that will be an interesting stat as time goes on - total DR points of players who haven't played for some weeks.

An other note: i found out that an unusual number of very low SR (14 pts or less) seems to have their DR points capped at 20,000 pts (which is middle-B). A few players are above, but i can imagine that the later tanked their SR after having gained a higher DR score. Maybe extremely low SR can block DR points ELO-like exchange.

Hmm. It's OK if the exchange is blocked completely, but if someone else loses points while those capped don't gain then those points disappear from the system forever.

I don't try to overwhelm you, but a last race for today:

Cool :)
 
Last edited:
No idea :) The 2000 is the only 'pure' input, but I'm not sure how much to allow for the other boosts/drops that happen when crossing levels. We can assume that there are more boosts than drops, since everyone starts at E. Considering just the D->C transition brings us a bit closer to 4366... 25% of players with more than 5 races have managed that, with its gain of 1200 (?) points. But it still doesn't seem like enough, and only ~10% have managed the next level.

I might spot something when I analyze that B level race, but I suspect that any other input to the DR point pool would happen (or be most obvious) at D or C levels.

4366 is the average of players with >5 sport races, right?
Yes, well, > 4 races, so 5 races or more.

I just analysed 15,000 entries of various Sport ranking lines from day 1 in order to spot specific DR where points would be poured into the system. I counted entries for each DR score, and only 3 values pop out:
2000 pts (1685 lines) as the real start point after 5 races.
20,000 pts (281 lines) as an apparent limit in case of Manner points <15
7,000 pts (73 lines) as an apparent probable limit in case of Manner points < 5


edit: Points effectively get taken out of the system when a player stops playing completely ("done with GTS"), so that will be an interesting stat as time goes on - total DR points of players who haven't played for some weeks.
How did you see that?
I'll investigate this in my data.
 
Yes, well, > 4 races, so 5 races or more.

I just analysed 15,000 entries of various Sport ranking lines from day 1 in order to spot specific DR where points would be poured into the system. I counted entries for each DR score, and only 3 values pop out:
2000 pts (1685 lines) as the real start point after 5 races.
20,000 pts (281 lines) as an apparent limit in case of Manner points <15
7,000 pts (73 lines) as an apparent probable limit in case of Manner points < 5

Unless we are very very lucky and catch the stats for a race including one of those limited, we won't know if it adds or removes points to the system, or is neutral (which would be my guess).

How did you see that?
I'll investigate this in my data.

I didn't see it, but it must happen. If a player gives up while they still have some points, and don't play again, nobody can win those points off them. Even those who stop with only a few hundred points could have a fairly big impact, if there's a sizable number of them. A mid-B player leaving would need to be replaced by 10 newbies... and we're coming out of the time of plenty of newbies!
 
If a player gives up while they still have some points, and don't play again, nobody can win those points off them
I don't see this as a problem at all. Even on a given finite number of players, the system would have started without all the points it contains today. The system id designed to produce graciously the points it needs. We don't know how yet, but my best bet would be an introduction of fresh points on very low ranked players that would spread towards higher DR players by a sort of capillarity process. This is where those points would be the more crucially needed - imagine a race full of players with one point, they wouldn't have anything to exchange, so a little boost to wining driver would be welcome.

I tried to find a low DR races to see if they are still zero-sum races, but didn't find one on YT so far.
But since i was there...
BillFlorence68 | 35913 / 90 (675 races)
DEPOSITOdeCARROS | 39066 / 99 (53 races)
DJSchwartz | 56553 / 99 (122 races)
Gui_Josa | 39689 / 99 (37 races)
GumballCGT | 57089 / 97 (102 races)
hellZfirEJP | 70995 / 99 (224 races)
imEbrahimX7 | 55607 / 99 (146 races)
IOF_RACING17 | 61717 / 99 (74 races)
luisfeitosa | 37671 / 99 (324 races)
MarKo_25 | 67250 / 99 (215 races)
maxking39 | 61959 / 99 (109 races)
Monster-Sport71 | 53246 / 99 (123 races)
ORMA_Snow | 67008 / 99 (231 races)
PTslopoke | 64544 / 98 (172 races)
RacingKing89674 | 56337 / 99 (370 races)
The359 | 38196 / 98 (271 races)
TheLadiesMan | 59976 / 99 (132 races)
todally_lit_ | 41452 / 99 (130 races)
TRL_doodle | 69099 / 94 (180 races) @Doodle
tunernaut | 56216 / 99 (139 races)
->
BillFlorence68 | 35147 / 84 (675 races) <= DNF
DEPOSITOdeCARROS | 39762 / 79 (54 races)
DJSchwartz | 56868 / 91 (123 races)
Gui_Josa | 38929 / 99 (38 races)
GumballCGT | 58018 / 99 (103 races)
hellZfirEJP | 71210 / 99 (225 races)
imEbrahimX7 | 54686 / 98 (147 races)
IOF_RACING17 | 62777 / 93 (75 races)
luisfeitosa | 37151 / 99 (325 races)
MarKo_25 | 66339 / 99 (216 races)
maxking39 | 61423 / 83 (110 races)
Monster-Sport71 | 54804 / 99 (124 races)
ORMA_Snow | 67700 / 98 (232 races)
PTslopoke | 64858 / 99 (173 races)
RacingKing89674 | 55229 / 97 (371 races)
The359 | 39085 / 99 (272 races)
TheLadiesMan | 60314 / 99 (133 races)
todally_lit_ | 41419 / 99 (131 races)
TRL_doodle | 68434 / 99 (181 races)
tunernaut | 55430 / 99 (140 races)

upload_2018-1-13_2-2-58.png

@Doodle

EDIT:
Unless we are very very lucky and catch the stats for a race including one of those limited, we won't know if it adds or removes points to the system, or is neutral (which would be my guess).
Guess who got very very lucky a few minutes ago?
I caught a DR->2000 and TWO DR->7000 in one race :lol:
CYGAN-PL-GTA | 5411 / 15 (11 races)
DANY-1984GM | 8612 / 18 (226 races)
elcereceda | 4877 / 19 (236 races)
exl_gambrilus | 7558 / 12 (71 races)
ilpizzu | 8910 / 18 (143 races)
ju130i | 5675 / 12 (300 races)
Mad-man-den | 1 / 29 (4 races)
Matraquilho556 | 7637 / 16 (62 races)
paconet_plus | 9440 / 11 (264 races)
simobima | 5432 / 16 (77 races)
TheSCRAFINATOR | 18625 / 11 (100 races)
->
CYGAN-PL-GTA | 5224 / 16 (12 races)
DANY-1984GM | 8996 / 25 (227 races)
exl_gambrilus | 8281 / 15 (72 races)
ilpizzu | 9449 / 14 (144 races)
ju130i | 5321 / 9 (301 races)
Mad-man-den | 2000 / 30 (5 races)
Matraquilho556 | 7562 / 7 (63 races)
paconet_plus | 7000 / 4 (265 races)
simobima | 4767 / 12 (77 races)
TheSCRAFINATOR | 7000 / 2 (101 races)

Note: A driver was present during the warmup but then quit/got disconnected, but their stats remained unchanged after the race.
upload_2018-1-13_3-5-52.png
 
Last edited:
I don't see this as a problem at all. Even on a given finite number of players, the system would have started without all the points it contains today. The system id designed to produce graciously the points it needs. We don't know how yet, but my best bet would be an introduction of fresh points on very low ranked players that would spread towards higher DR players by a sort of capillarity process. This is where those points would be the more crucially needed - imagine a race full of players with one point, they wouldn't have anything to exchange, so a little boost to wining driver would be welcome.

It's a problem when we don't know how :lol:

Well, actually, when someone has 1 pt they can still lose a considerable number of points... but then their debt is erased. So that is point creation, on a fairly large scale.

It could still be a problem though, because there are cases where quite large amounts of points are lost from the system...

We see Dayrone capping at 75 000 pts

Overall 3789 pts are lost from this race: about 700 that Dayrone would've gained, and 3000 from montero dropping down to A.
My calcs are within 19 pts for each player, apart from those two. Even though it's my worst estimate so far, I'm pretty sure that other players lost points to Dayrone even though he didn't gain them.

Guess who got very very lucky a few minutes ago?
I caught a DR->2000 and TWO DR->7000 in one race :lol:

Oh fantastic! And horrible at the same time... about 14000 pts are lost from the two cappings. Overall, the race loses 11701 pts.
My calcs are within 8 pts for each player, apart from those two and mad-man-den. He got a boost to 2000, but the calculated score for him was -392, so others in the race would've gained those points. Also, the two capped players would've otherwise won about 50 pts between them. 14000 - 2000 - 392 + 50 = 11658 which is pretty close to 11701 (considering that the 392 and 50 are just my estimates).

When you said they appeared to be limited, I just thought they couldn't rise above the limit. This cutting them down is throwing a lot of points away.

Other FIA race with 12 racers:

Zero-sum, and my calcs are within 4 pts for each player.

On these few races I don't think there's anything to suggest any creation of points other than what we know about already, and all of the losses can be explained to within a few pts.
 
It's a problem when we don't know how
The how is a problem of its own, our comprehension of the system doesn't affect it.

Overall, the race loses 11701 pts.
My calcs are within 8 pts for each player, apart from those two and mad-man-den. He got a boost to 2000, but the calculated score for him was -392, so others in the race would've gained those points. Also, the two capped players would've otherwise won about 50 pts between them. 14000 - 2000 - 392 + 50 = 11658 which is pretty close to 11701 (considering that the 392 and 50 are just my estimates).
The capped players exchange points before the cap is applies, right?
Also, wouldn't <5 races players be in a sanctuary zone where no point exchanges occurs at all.
Edit: i put those 7000 pts cap players on a watch-list to see if the game was given them their points back but it seems that they are lost for good:
paconet_plus | 6979 / 13 (267 races) <= now back above SR 5 after two more races

Not sure if there's anything worthwhile here, but still, two low to average Dr races:

RACE 1:
Edit: 963 pts generated!
D_phantom14 | 5668 / 32 (104 races)
ericd427 | 26299 / 31 (779 races)
GEE-33 | 1 / 42 (2 races)
Gnarly__Panda | 10311 / 36 (279 races)
icemanhenry85 | 1 / 46 (2 races)
JeanVogel | 20674 / 37 (62 races)
Koke_TD | 12404 / 34 (76 races)
LEX_FILES81 | 1 / 45 (17 races)
NdN_CHUCKY | 7725 / 32 (25 races)
onix905 | 780 / 48 (158 races)
tamerbrito | 25085 / 38 (234 races)
VeioPizzol | 1 / 47 (3 races)
->
D_phantom14 | 4888 / 30 (105 races)
ericd427 | 26136 / 33 (780 races)
GEE-33 | 1 / 40 (3 races)
Gnarly__Panda | 10849 / 31 (280 races)
icemanhenry85 | 1 / 44 (3 races)
JeanVogel | 21282 / 39 (63 races)
Koke_TD | 12734 / 34 (77 races)
LEX_FILES81 | 150 / 46 (18 races)
NdN_CHUCKY | 7849 / 30 (26 races)
onix905 | 594 / 44 (159 races)
tamerbrito | 25428 / 39 (235 races)
VeioPizzol | 1 / 48 (4 races)

upload_2018-1-13_11-55-46.png


Race 2:
Edit: 2248 pts generated!!!
auluccitis | 19487 / 19 (180 races)
Bladerunner1144 | 10564 / 33 (27 races)
BorbaWayne | 49432 / 29 (228 races)
cinisaru1971 | 6954 / 14 (105 races)
CyberWarGLL | 10615 / 13 (121 races)
Deanw1nchster | 7900 / 35 (28 races)
EdDrJ | 862 / 9 (26 races)
fatmankilla | 2105 / 15 (121 races)
Ken-033 | 5779 / 18 (266 races)
mexicanizer | 20664 / 27 (59 races)
MSBslurpee222 | 2206 / 35 (6 races)
NdN_CHUCKY | 7849 / 30 (26 races)
rolfington | 3162 / 1 (399 races)
sapodolpho | 1 / 23 (116 races)
vinives91 | 8855 / 8 (228 races)
->
auluccitis | 19374 / 18 (181 races)
Bladerunner1144 | 11037 / 30 (28 races)
BorbaWayne | 49488 / 37 (229 races)
cinisaru1971 | 9506 / 24 (106 races)
CyberWarGLL | 9973 / 11 (122 races)
Deanw1nchster | 7176 / 37 (29 races)
EdDrJ | 971 / 7 (27 races)
fatmankilla | 2021 / 22 (122 races)
Ken-033 | 5433 / 22 (267 races)
mexicanizer | 21151 / 24 (60 races)
MSBslurpee222 | 2436 / 38 (7 races)
NdN_CHUCKY | 8241 / 33 (27 races)
rolfington | 2339 / 1 (399 races)
sapodolpho | 1 / 20 (116 races)
vinives91 | 9536 / 15 (229 races)

upload_2018-1-13_12-18-55.png
 
Last edited:
The capped players exchange points before the cap is applies, right?

Yes, I think so. Which in that example means those 50 points were lost as well.

Also, wouldn't <5 races players be in a sanctuary zone where no point exchanges occurs at all.

If my calcs are correct then a race with 10 fresh players, all with 1 point, gives 711 points to the winner through to 79 points for 5th place. 6th through 10th are the inverse of that, but since they're limited to 1 they don't lose points. A total of 1975 points would be created.

Not sure if there's anything worthwhile here, but still, two low to average Dr races:

In the first of those we see 4 players starting at 1 point, and 3 of them 'lost' points to others. Overall, 963 points were created by covering their debt.

So it would seem that there is a lot of point creation at that very early stage, and from anyone who has fallen back down to 1. Enough to allow for that average of 4366 points. I'm still a bit worried that it relies too much on new players, and those who are the most likely to give up, though.

Thanks for posting all these races, and sorry I haven't had time to look at all of them yet - I will! I want to try and make my calculation more accurate, but using spreadsheets is rather limiting - I'll probably end up writing some code instead to try and find the correct constants and roundings. But, it's not really that relevant - we have managed to document and explain most aspects of the system already :D
 
Race 2:
Edit: 2248 pts generated!!!

The point creation seems to be due to two things: last place with 1 point not losing any but giving points to others (approx 744), and 1st place (cinisaru) moving up to DR C (was limited by SR before) and getting a boost (1500 would fit).

So, so far I've seen that S->A is -3000 and D->C is +1500 in the races I've analysed. Are we sure of any other boosts/drops already?


The race was buggy for TX3_Devil who saw all others stuck on the grid during the whole race.
Anyway, the final positions:
upload_2018-1-12_20-16-0-png.705879

I don't know if it's because it was bugged, or just a quirk of the system, or a different bug, but... those that stayed but didn't finish before the timeout (16th to 18th) are listed in reverse order, according to the championship points awarded for the race, and also according to how their DR was changed. That caught me out for a while, looking for a mistake in my spreadsheet!
 
So, so far I've seen that S->A is -3000 and D->C is +1500 in the races I've analysed. Are we sure of any other boosts/drops already?
Would you make use of more race results?

Edit:
andrewsufc91 | 3687 / 93 (14 races)
bofman | 3575 / 99 (23 races)
claudio600rr | 3511 / 99 (91 races)
cyrilus971 | 3423 / 94 (211 races)
FSG_Gerry | 3834 / 99 (90 races)
General-Black-Zo | 3184 / 99 (107 races)
HerrBaer87 | 3619 / 95 (9 races)
kraatz1979 | 3289 / 93 (35 races)
Maniphiko | 3582 / 99 (237 races)
martinpaliwoda | 3419 / 99 (262 races)
ome800 | 3781 / 99 (58 races)
pancyclo | 3557 / 99 (318 races)
Pasadesa | 3245 / 99 (98 races)
perroverde13 | 3178 / 98 (64 races)
samgf | 3522 / 99 (107 races)
Setto10 | 3043 / 99 (121 races)
spike20071983 | 3191 / 99 (25 races)
sprprs | 3200 / 91 (339 races)
vanni40 | 3411 / 99 (168 races)
veng_ | 3789 / 99 (8 races)
->
andrewsufc91 | 6688 / 99 (15 races)
bofman | 6263 / 99 (24 races)
claudio600rr | 2636 / 99 (92 races)
cyrilus971 | 3503 / 99 (212 races)
FSG_Gerry | 3420 / 99 (91 races)
General-Black-Zo | 5570 / 99 (108 races)
HerrBaer87 | 3533 / 92 (10 races)
kraatz1979 | 1942 / 92 (35 races)
Maniphiko | 2861 / 99 (238 races)
martinpaliwoda | 3659 / 99 (263 races)
ome800 | 3529 / 99 (59 races)
pancyclo | 2997 / 99 (319 races)
Pasadesa | 3809 / 99 (99 races)
perroverde13 | 3588 / 99 (65 races)
samgf | 5736 / 99 (108 races)
Setto10 | 5912 / 99 (122 races)
spike20071983 | 1848 / 99 (25 races)
sprprs | 2175 / 99 (340 races)
vanni40 | 2063 / 95 (168 races)
veng_ | 6308 / 99 (9 races)

upload_2018-1-14_19-52-55.png
 
Last edited:
Would you make use of more race results?

I could, but I appreciate that it takes quite some time to find and capture these races (especially when looking for specific events), so it's something we could leave for now if you like. I'm expecting that the changes are symmetrical, so A->S is +3000 - but it would be quite interesting if they weren't! These figures can only be guessed at when watching an individual's DR, since they are added to any other DR change in the race.
 
I could, but I appreciate that it takes quite some time to find and capture these races (especially when looking for specific events), so it's something we could leave for now if you like. I'm expecting that the changes are symmetrical, so A->S is +3000 - but it would be quite interesting if they weren't! These figures can only be guessed at when watching an individual's DR, since they are added to any other DR change in the race.
I try to be selective, watching ranks before the race.

I thought this one could be interesting since there is a good range of DR, but at first glimpse, didn't see anything notable:
AamPT92 | 21907 / 75 (37 races)
BaBiNhO_LH44_MU | 33485 / 71 (41 races)
BladeXtreme-92 | 24022 / 70 (93 races)
boulalasgp | 35813 / 74 (110 races)
bugilointi | 31809 / 79 (52 races)
byrons | 1 / 87 (38 races)
Delosfr | 20651 / 73 (149 races)
FkFortex | 1 / 81 (56 races)
ilias420 | 37762 / 73 (82 races)
kettchup | 23796 / 71 (190 races)
M1tBL31 | 22242 / 75 (354 races)
Meaulnes | 31548 / 73 (351 races)
norminator- | 1 / 89 (84 races)
rafael180 | 19591 / 78 (359 races)
Sashmunel | 19798 / 75 (372 races)
willyglezanduiza | 1 / 83 (151 races)
witchfinder666nl | 1 / 86 (56 races)
Wulffen1981 | 26289 / 74 (246 races)
xhuytonx | 1 / 87 (167 races)
xmbferreira | 1 / 89 (398 races)
->
AamPT92 | 22445 / 89 (38 races)
BaBiNhO_LH44_MU | 34039 / 84 (42 races)
BladeXtreme-92 | 23994 / 70 (94 races)
boulalasgp | 35474 / 68 (111 races)
bugilointi | 32744 / 90 (53 races)
byrons | 1 / 99 (39 races)
Delosfr | 20283 / 67 (150 races)
FkFortex | 145 / 91 (57 races)
ilias420 | 38302 / 82 (83 races)
kettchup | 24576 / 81 (191 races)
M1tBL31 | 22608 / 87 (355 races)
Meaulnes | 31060 / 66 (352 races)
norminator- | 1 / 97 (85 races)
rafael180 | 19105 / 86 (360 races)
Sashmunel | 20578 / 85 (373 races)
willyglezanduiza | 1 / 93 (152 races)
witchfinder666nl | 1 / 93 (57 races)
Wulffen1981 | 25857 / 77 (247 races)
xhuytonx | 1 / 95 (168 races)
xmbferreira | 1 / 94 (399 races)

upload_2018-1-14_21-11-2.png
 
I try to be selective, watching ranks before the race.

I thought this one could be interesting since there is a good range of DR, but at first glimpse, didn't see anything notable:

Just those seven with 1 DR each, even after having 950 races between them! :boggled:

Oh, and it managed to get the finish positions right in that one, for those who didn't cross the line. Probably the bugged race causing it the other time.

Sometimes my calcs work well, but this one wasn't very accurate even after tweaking the constants a bit...
 

Attachments

  • results_177.xls
    41 KB · Views: 12
Edit: i put those 7000 pts cap players on a watch-list to see if the game was given them their points back but it seems that they are lost for good:
paconet_plus | 6979 / 13 (267 races) <= now back above SR 5 after two more races
He's now paconet_plus | 7311 / 3 (269 races)
The SR 5 limit is not really one apparently, but would work as a gate when crossed downward and that gives a penalty (bring score back to 7000 if it was above).

Just those seven with 1 DR each, even after having 950 races between them! :boggled:
I saw one of them, crossing the SR 5 limit, stating that the DR progress is "a long process". It's one of many GT Sport Sisyphus: winning points, destroying them, and repeat.

Sometimes my calcs work well, but this one wasn't very accurate even after tweaking the constants a bit...
Doesn't the number of races have a small role?
 
He's now paconet_plus | 7311 / 3 (269 races)
The SR 5 limit is not really one apparently, but would work as a gate when crossed downward and that gives a penalty (bring score back to 7000 if it was above).

Interesting. Yes, could be triggered like that, hard to think how else. Another possibly is that it happens when DR is > some level and SR < 5, but paconator didn't start that race with much more (and would've only gained about 100 pts from the race).

I saw one of them, crossing the SR 5 limit, stating that the DR progress is "a long process". It's one of many GT Sport Sisyphus: winning points, destroying them, and repeat.

:lol: It's like looking at a completely different game to the one I play!

Doesn't the number of races have a small role?

It's possible, since in chess different values for K are used for experienced vs beginner (in some orgs). So that could vary based on # of races or driver points. It would be very painful try and work that out, so I'm hoping it's something simpler - for example, the Elo formula for Q uses 10 as the base, but maybe that is different (e.g. e).

---

I've put all the data from races so far into uniform tables and saved as .csv, which are then easy to copy-paste into the right part of my calcs spreadsheet. Attached in case anyone else wants to play :)
 

Attachments

  • results20180114.zip
    4.6 KB · Views: 15
Back