Having an early look at the new FIA Points

Discussion in 'Gran Turismo Sport' started by Jomas, Nov 15, 2019.

  1. Jomas

    Jomas

    Messages:
    185
    Location:
    France
    Introduction

    Hello everyone, as you may already know, a new point system has been added to the game. Some of you might say that it is too early to make any kind of conclusion and we fully understand that, so lets say it's an early conclusion. A few months ago, with the help of @khkenni and @GT_Alex74, we've made a mathematical formula which was making the estimation of the FIA Points in a lobby, based from the DR of the drivers, from the previous point system. Its accuracy was decent and helpful for me in the Top 16 races, knowing where I need to be and where my opponents shouldn't finish.

    With that in our bag and the new system in action, we gathered datas from 7 Splits in EMEA, Manufacturer Series, Time Slot 1 during Round 1 (13th November) to try to see how the new point system works and if some issues has been adressed. We assume those Splits are the 7 firsts of this time Slot, based on the DR of the drivers in it. Split 2 has sadly been broken, everyone in it being disconnected except RC_Snake91 who was alone in the lobby. We took the datas from the screenshots/videos he shared in a conversation I am in, and considering how the new system works, based on the "average DR", we figured out it was still some valuable data, using Snake's DR as it is the average.

    Keep in mind that we are players like everyone else, and we don't have access to any special file or documents. We're figuring things out from what we see.

    What was the issues?

    - Disconnections during matchmaking lowering the Split's points
    - Gap between high Splits not big enough, encouraging drivers to tank their DR in a context of competition (for World Tours & World Finals)*

    *: This issue occurs mainly in EMEA due to the high density of players compared to the other regions. It is also the only region which hasn't been split in GTS' lifetime, when Oceania & Asia have their own regions as North and South America. Some may argue that it is not an issue, we will treat it as it is one.

    Is it adressed?

    Yes and... No. The disconnection issue has been adressed which is a very very good point, but the second point is... Slightly better at best for high DR splits and worse for low DR splits.

    How the point system works now?


    Fundamentally, the point system works the same as it previously was but the points displayed nowadays are around 8.5 times smaller than it was before. Moreover, the last place of a split scores 20% of the 1st place's points, instead of 0% before, which is a good thing. We've also noticed that the average DR in a Split isn't linear with the points earned, noticing a trend where the "DR Points/FIA Points" decreases when the average DR in a Split is lower. We're assuming another factor is in the FIA points algorithm but we're yet to find what it clearly is. Although, this unknown factor is not significant, and may alter the points by 1-2 points at most.

    So... What are the consequencies for us players then?

    - For high DR splits, the overlap between them is still way too big, even if it's slightly better than it was before, since last place scores 20% of first place. On our datas, we've noticed that P1 in Split 2 overlaps Split 1's points from P4 and downward. The following N Splits are overlapped from 2 downward by the N-1 Splits. If it looks better for Split 1 compared to the other splits in comparison, the gap between Split 1 and Split 2 was 7000 DR points, when the other only had 3000 DR points gap. In a general way, when we compared splits with an average of 3000 DR Points gap (over 50000 average DR), the overlap of a N-1 Split was P2 of the N Split.

    Gathered Data

    [​IMG]

    Previous System (splits having the same average DR than previous table, according to our formula)

    [​IMG]


    (Sorry, my screenshots crops 19th and 20th position.)

    - For low DR Splits the issue comes from tie points between positions within the same split. We've figured that a Split with an average DR around 5000 Points or lower (or lower than 24 points for the winner) will see tied points between positions. An example I've seen on Twitter was P1 and P2 scoring both 19 points.

    - In a general way, more tied positions will happen in the leaderboard as the scale of points goes from 3125-0 to 364-0. Those tie points will also alter the Manufacturer Rankings.

    - To answer the question of this part, nothing much will change in the end. Sure, the points looks lower which can be weird when we are used to 4 digits numbers but in a competitive aspect of the FIA races, we can't say the global picture has been altered much. As every updates, there is some good and some less good ideas, and I believe that it is an experiment for this exhibition series which can lead to something else for the next season. We also want to remind that it is an early conclusion.

    A possible way to get around the overlapping split issue

    As we don't like to point out (what we think is) an issue, we also thought on how to solve it. We don't claim it as THE solution, but as one of many possible, written in a mathematical way and brought an example of how it would look like: (In the spoiler to make the thread easy to load)

    [​IMG]

    The basics of this system are to make it simple but yet answering to the overlap issue mentionned before as the tied points. Having a way higher cap points, 5000 in this case when the split is maxed out, makes the possibility of having tied points between players less likely while knowing the clear cut of the ideal split. Then, reducing the percentage of points lost per position to lower the overlap, trying to balance the points and skills encountered in the different splits.

    Another way to think about it would be to bring back the 1 shot attempt in FIA Races. Gaps between splits will naturally increase, discouraging DR tanking, as making the race way more intense on the human side.

    Note: I originally thought of making this message into the "2019/20 FIA Exhibition Series Race Discussion" but looking at how long this message was, I felt it was more appropriate to open a new fresh thread. This thread is also an opportunity to share constructive ideas regarding the subjects brought here. Free bashing is not the point here.
     
  2. breeminator

    breeminator

    Messages:
    748
    As I said in the FIA thread, where I showed much the same thing with regards to how the point of overlap can easily be adjusted to wherever it is desired to have it, I think PD are looking at this as a psychology problem, i.e. they want to maximise player engagement at all DRs.

    Setting the overlap where you suggest would be good for people in top split, of course, but what effect would it have on people lower down? Someone racing in a 57k lobby is going to be beaten by 111 people on points, even if they win their lobby, and no matter how much better they drive than people in higher splits.

    As an example, in the first round of last season's manufacturers, the wet RBR event, I was lapping in a 30k lobby significantly quicker than Windfire was lapping in NA top split, because his car was just hopeless for it, whereas the NSX I was using was rather good for it. If he were in the same region, should he automatically get a massively higher score than me, just because he's in top split, regardless of how much slower than me he might be for that race? I think there have been other races where the winner of my <40k lobby has done a total race time that was faster than over 50% of people in top split. For example, I've just checked the Brands Hatch Indy race from last manufacturer season. I think my lobby was around 36k, and was won in 22:21. The last top split race, covered in Kie's "GT Sport FIA Rage" video was won in 22:19, with only 3 drivers faster than 22:21. I assume that PD has access to the times from every race and is assessing this kind of thing in deciding where the overlap should be.
     
    O_Maverick_O, Akwaaba and Granadier like this.
  3. mclarenLB

    mclarenLB

    Messages:
    1,061
    Location:
    United Kingdom
    I mean... a 30k EMEA split vs. a later night slot NA top split is basically the same pace. DR:A's regularly get into NA top splits, especially later at night.

    Also, DR across different regions isn't 1 to 1 exchangeable. Due to the 75k DR ceiling, the faster the region overall, the harder it is to get DR at the high end. I think your cross-regional comparison is totally inaccurate.
     
    Akwaaba and khkenni like this.
  4. breeminator

    breeminator

    Messages:
    748
    The NA comparison was with Windfire, who has 75k DR. He was just slow because he was in the Mustang, which was terrible for that race.

    The EMEA winning time I mentioned was a driver who is currently at 70k, 2nd was a driver who is currently at 74k. The 74k driver was in the same car as the winner of my 36k lobby, and was only 2.4 seconds quicker.
     
  5. JayOTT

    JayOTT

    Messages:
    546
    EMEA points overlapping aside, they've done well to address the main aspect of the previous system that was unfair; getting put in an 18-19 player lobby by the matchmaking. Since the old formula only cared about how many drivers you beat in the race, it was common to see a weaker lobby with 20 players offer more points than a stronger but smaller one. Now your multiplier is constant based on finishing position:

    1st: 24, 2nd: 23, 3rd: 22, 4th: 21
    5th: 20, 6th: 19, 7th: 18, 8th: 17
    9th: 16, 10th: 15, 11th: 14, 12th: 13
    13th: 12, 14th: 11, 15th: 10, 16th: 9
    17th: 8, 18th: 7, 19th: 6, 20th: 5

    I am fairly certain these figures are correct based on every points table I've seen so far. The base points that these get multiplied by are harder to get a precise reading on, but my current estimate is the average DR of the lobby divided by 4800.
     
    darth_tails likes this.
  6. Lazybike

    Lazybike

    Messages:
    509
    I guess it depends what PD want from the FIA races...personally I don't take any notice of the points, I'm A/S, I'm not fast enough for world tour events, so any points are meaningless, They need a fair system to determine finalists for tour events, and they need interesting combo's to keep the rest of us playing.
     
  7. darth_tails

    darth_tails

    Messages:
    1,161
    Location:
    Scotland
    Thanks for the article and the comments above, helps me understand a bit better the system now for the new points which
    will be handy going into tonight's races as tonight's rounds take place at two of my better tracks providing I don't die in
    the infamous Dragon Trail chicane of "death".

    In essence the new points seems to favour trying to get good qualifying time and pole position to avoid carnage or if you have
    good skill and a decent car being able to get up the order but I still don't understand why if the points are as @JayOTT describes
    how higher ranked players are getting more but I am sure that will be explained at some stage.
     
  8. mclarenLB

    mclarenLB

    Messages:
    1,061
    Location:
    United Kingdom
    Race time isn't an accurate measurement either, as any battling that takes place for the lead will slow down the race time considerably.
     
    Akwaaba, Landy1264, GT_Alex74 and 3 others like this.
  9. khkenni

    khkenni

    Messages:
    71
    Location:
    Denmark
    That's the whole point. In my opinion, there's no way someone multiple splits lower should be anywhere near the points for podium in top split.

    If you want to compete in this, and go for higher FIA points, you should be in the highest split you can get into, otherwise (except if you can get a podium in top split) no matter what your DR is, it will be beneficial to lower your DR before FIA races, in order to get more points in a lower split.
     
    Krypton78 likes this.
  10. MPincombe

    MPincombe

    Messages:
    149
    I won EMEA first slot split 4 the other night and I got 275 points for it, P8 top split would've got less than me. Now is this fair? I've come to the conclusion no, and it's not based on the race time, although my race time was pretty slow, race times can be pretty subjective to how fast the guys around you are driving. I was at 56k DR so I'm not slouch but where would I finish in the top split if I was in there, 15th to 20th? Maybe, maybe better but it just doesn't seem right that I got better points for my win than 12 drivers who race in top split, being in top split alone should be enough to get you a decent points haul, I think the crossover to my split should've been P10-P12, P5-P7 for split 2, P8-P10 for split 3

    Also you mention that in lower splits points were tied, well surely anyone with two brain cells can work out that having a lower points system is going to lead to tied points at some point. Surely the best way to have the points would be to have more on offer as a base and bring it down sufficiently, this leaves more wiggle room for the extreme lower splits, I really don't see the logic behind this decision from PD.

    However there are the improvements of last place not getting 0 and a DC not effecting the points, and if one of the lower DR drivers get's disconnected do the points actually increase in the lobby?
     
    Akwaaba likes this.
  11. Lazybike

    Lazybike

    Messages:
    509
    Couldn't they just have a system where lowest points win...so winner of top split gets 0, 2nd gets 1 point etc... winner of 2nd split gets 21 points and so on down through the splits. As I said earlier, I suspect points only really matter to top split and half the 2nd split.
     
    Last edited: Nov 16, 2019
  12. Rangeraus

    Rangeraus

    Messages:
    192
    For some Oceania context:

    The winner of the Manufacturers 2nd split race tonight (1st timeslot) scored 226 points.
    Note that the #1 seed for that race had a DR of 52.6k as of yesterday. Only 3 drivers in the race had a DR > 50k.

    The score of 226 points is more than the score for the following positions in the first two top split races:
    Timeslot 1: P9 (P8 scored 227 points). This race had 9 drivers with DR > 70k
    Timeslot 2: P5 (P4 scored 228 points). This race had 4 drivers with DR > 70k

    There was always some overlap but I don't remember it being this noticeable in previous seasons.
     
  13. mclarenLB

    mclarenLB

    Messages:
    1,061
    Location:
    United Kingdom
    Well no because you could just... not race
     
    Lazybike likes this.
  14. Krypton78

    Krypton78

    Messages:
    614
    Location:
    Germany
    There is no problem with the point calculation which is based on DR in the split. The Problem is... again... the DR-cap at 75k. Perhaps half of top split would have easyly 80 or even 90k DR, and then you would have the point gap between the top and second split. It would automaticly match with the skill difference between these two Lobbys- btw, I think here it is the most. For example, you don't find that skill difference between split 3 and 4, or when you look further, 7 and 8, where is probably no measurable speed/skill difference anymore.
     
    Last edited: Nov 16, 2019
    eAxis Patohm likes this.
  15. pantherjag

    pantherjag

    Messages:
    268
    First off the fact they have fixed the system to stop disconnect effecting the total points and also that points are now scored down to 20th is a positive.

    However I do feel the complaints regarding the overlap in points are made somewhat redundant because of the fact only top 5 results count plus you have multiply times to enter a race.

    If the championship was 1 time entry per race and all 10 results counted I'd be more sympathetic to the top split and understand the sentiment that it's unfair for the winner of split 5 to get more points that 8th in top split but that's not how the championship works

    The way I see it a true top split player probably discards an 8th place finish anyway. When you think about it if a top split player can get 5 top 6 finishes from 30 trys that's probably going to be enough to outscored any regular 2nd 3rd split driver anyway. If a top split driver cant pick up said top 6 finishes then there is no guarantee that they would do much better in 2nd split anyway as they would need regular podiums to procure similar points.

    Also having watched the world tour events since nurburgring i dont recall seeing a representative from EMEA that wasnt a regular top split driver. Could be wrong on that front but at very least the vast majority are true top split players.

    Just my thoughts and opinion on the subject. Sure system might not be ideal but I think the fact it has an element of chance and can throw a curve ball into the works keeps lower splits more engaged in the championship. I certainly think it's better than a system that could see a player finish 20th 5 times in top split and still place 30th in the championship.

    Anyways good work and I enjoyed reading the OP
     
  16. Lazybike

    Lazybike

    Messages:
    509
    I hadn't thought of that....:O:dunce:
     
  17. breeminator

    breeminator

    Messages:
    748
    The proposal in the OP goes way beyond that, though, e.g. the winner of split 6 scores the same as last in split 4. It essentially regards the worst performance of anyone in a race as being that of a driver with 91% of the DR of the standard of the winning drive. In my experience that simply isn't an accurate reflection of the range of performances seen in lobbies. For example, the Brands Hatch Indy manufacturer race I mentioned before - 14th place in my lobby was over a lap behind the winner. That's not a 9% lower DR standard of performance, it's more like a 50% lower DR standard of performance, and that's only 14th place, not last (my recording doesn't show times below 14th). That isn't unusual, I just looked at the gap to 14th place in some other races - 42 secs, 44 secs, 52 secs. I just looked at some of the gaps to 14th in top split from Kie's stream tonight - 30 secs, 38 secs, 31 secs, 23 secs, 44 secs. Thinking about it, I think there might be an issue with the fact that the DR scale itself is not linear, will have to give the implications of that some more thought.
     
  18. breeminator

    breeminator

    Messages:
    748
    Okay, an example from people in the same race - a 65k driver just came 4th in EMEA manufacturers top split. The scheme proposed in the OP would make it impossible for someone at that DR to score the same points as even last place in a 73-75k split, yet he beat 2 x 73k drivers when in the same race.
     
  19. mclarenLB

    mclarenLB

    Messages:
    1,061
    Location:
    United Kingdom
    If he keeps that up he'll be 73k-75k himself in about 5 races, its not an issue.

    You should have prepared your DR in advance if you intend on taking a season seriously, which you can literally do at will in Daily Races
     
    khkenni and Jomas like this.
  20. breeminator

    breeminator

    Messages:
    748
    People aren't the same standard at every combination of track and car, and for manufacturers it's expected that a manufacturer will be better for some races than others. What the proposal in the OP is doing is basing points primarily on someone's overall standard rather than it being about how they do in an individual race. It would be fine if lobbies were split based on QT rather than DR, i.e. if the game got everyone to do their qualifying session solo before grouping people together by QT. That way, if someone's car were particularly good or bad for one race, they'd be in a correspondingly higher or lower split. But with lobbies split by DR, I don't think such tight groupings of points accurately reflect the range of performances that happen in each race.
     
  21. Jomas

    Jomas

    Messages:
    185
    Location:
    France
    The scheme proposed is an example with every splits having a 3k average DR gap, based from the EMEA experience. In reality, it can (and will) be different since the average DR will fluctuate depending of the drivers in it but we wanted to have a standard gap to have a look at what the proposition would give.

    Basing the splits depending of the QT in TT is just a no in my opinion. If in Nations Cup & Supra Cup it would kinda work, it'll be heavily dependent of the time a driver spent in TT mode, and in Manufacturer Series, add the previous bias and add the car bias. Driver Rating, despite its flaws and easiness to manipulate, remains a good factor to decide of the splits composition, and the purpose is to build a context where manipulating your DR is not a viable strategy, motivating the drivers to work on their DR as much as they can.

    On a side note, tanking the DR desn't affect only high DR splits. A friend of mine, who's a the barrier of DR A/A+ (50k DR) also purposely decreased his DR to be in a lower split and having more easiness to score high points, which worked in the end.

    But overall, retries or not (and I wish we go back to 1 attempt, it would solve many issues at the same time), being outscored by drivers you can't see on track and having significantly less DR than you from P3 downward is neither fair, balanced and enjoyable, even more when you fight for a result.
     
  22. breeminator

    breeminator

    Messages:
    748
    Just to be clear, I wasn't suggesting splitting based on free practice times, I was suggesting the exact same structure as at present, but the 10 minute qualifying session would be done solo prior to the splitting into lobbies. If someone has spent a lot of time practising to improve their performance, then it would be fair enough for them to get into a higher split if they are genuinely performing better than higher rated drivers who have spent less time practising. And I'd also see it as fair in manufacturers for someone who has an OP car for a particular track to score better than a higher rated driver whose car is poor for that track, otherwise it wouldn't be revealing the true difference between the different manufacturers.
     
  23. Jomas

    Jomas

    Messages:
    185
    Location:
    France
    I get it better now, thanks for the precisions, but I'm afraid of the points if it stays the same. If the Splits are based on the qualifying times, there is nothing to prevent a faster split to score less points than a slower one. Base points wouldn't work since the number of splits vary too much between time slots, and even between regions. Plus, the DR would lose even more its interest.

    On the other hand, especially in Manufacturer Series, we could end up with 10 times the same car if the combo is heavily car biased regarding it. At this rate, it'd be better to have splits within Manufacturers, but I don't think it is a good idea as some manufacturers will hardly fill one split and the Global Manufacturer Ranking would die instantly.
     
  24. breeminator

    breeminator

    Messages:
    748
    Right, yes, I was saying I'd be happy with your suggestion of last place getting around 90% of 1st place if the splits were based on a 10 min solo qualifying session. I agree it would make no sense at all if last gets 20% of 1st.

    I also agree that 20% is too low for the current DR-based splits, and that there is an incentive to race at lower DR at all levels. It's saying that the worst performer in a e.g. 60k lobby is performing at the level of a 12k driver, and that is too low. So, while I think 90% is too high, as I see people perform much worse than 54k standard at that level, 20% is too low. Maybe somewhere around 50%, which would be saying the worst performer in a e.g. 60k lobby still performs as well as a 30k driver. People do often have pretty shocking races, e.g. they might pretty much stop dead if they hit the chicane of death, so I don't see it as that unlikely that one person in a 60k lobby will perform that much worse.
     
  25. MIE1992

    MIE1992

    Messages:
    1,230
    Location:
    United States
    Or have some sort of reward for being one of the better drivers on a lower scale, such as if you're among the best in the state. Maybe a credit/exp payout based on your ranking in a season, idk. At least some sort of in-game reward, like how if another player from your manufacturer/nation is the best in the region for a season, you get some credits/MiEx points.

    But otherwise, it should be inherently fun to race, and I don't think some aspects of the current setup is as fun as it could be. It all just feels so futile, like unless you dedicate every second of spare time to this game, you're just fodder.

    I mean, one idea would be to use the "sub-class" idea that I and probably others have mentioned, instead of either having each race (either FIAGTC or Daily Races) include all cars in a class, or just be one-makes. On that note, I don't get why, when there's a race that permits the SF19, they don't permit both of them, and it's usually just one, when they're so similar that neither of them would get a benefit from BoP, which is why I think they're in Gr.X instead of a new "Gr.SF" or something. I've said it already that there's a lot of potential with the race setups in general, so I don't think there's any need to repeat it ITT.
     
    Last edited: Nov 18, 2019
  26. Lazybike

    Lazybike

    Messages:
    509
    At my level I don't bother doing practice etc, I did the Manufacturers at Nurburgring because it had no tyre choice, no fuel saving..it was jump in, race, pit stop halfway...nice and simple, I do like the longer races though. Didn't bother with Nations.
     
  27. King_speedy11

    King_speedy11

    Messages:
    13
    i dont undertstand the point from my end i done 1 race my 1st race ever on nations cup and got 11 points,then thier are guys in my country that have 101 point after 1 round how does it work,do you race one of the 3 races or all and are all the point put together
     
  28. watto79

    watto79

    Messages:
    921
    Location:
    United Kingdom
    The points available in a race is dependent on the average DR of the room. So races with lots of A+ rated drivers will get more points than races made up of C and D rated drivers.

    Only the points from a single race (the last race you do) count.
     
    banckzy likes this.
  29. Lost Sheltie

    Lost Sheltie

    Messages:
    582
    Surely the simplest way of ensuring that people strive to get in top split is to simply add 50 points onto everyones score in it. They've given base points to top 16 races before, so it shouldn't be too difficult to code in.

    Adding 50 points would mean that the winner of split 2 would get about the equivalent of top split p7, so that should be enough to encourage people to risk top split rather than the safer/slower split 2 race.

    For us norms in the lower splits the points don't matter that much. It's nice to see we are doing well, but it isn't going to matter at the end of the season like it does for the best drivers. It's not like we are ever going to be racing at Monaco :(