- 3,613
- New England
- jobyone71
◆ SNAIL [Spec] Racing - Currently Recruiting for GT7 - JOIN TODAY!!Open
- Thread starter zer05ive
- 150,018 comments
- 8,687,969 views
JLBowler
Premium
- 7,985
- California
- JLBowler
I'm working on something that looks like it has some serious promise for adjusting based on the size of the grid without giving too much of a reward or penalty to large or small fields. It looks like it's really going to be a good way to measure the performance of drivers in different divisions based on their point total and size of the field.
The maximum number of points a driver can score is always 100, however, the minimum number of points a driver can score varies based on the number of drivers. This minimum is never zero. With 16, the min. is 6 and it goes up 6 for each less driver in the room. By using a multiplier based on this difference in available points, I should be able to come up with a formula that will accurately compare the performance of drivers in different divisions with a different number of racers in each one without having any noticible bias based on room size.
The maximum number of points a driver can score is always 100, however, the minimum number of points a driver can score varies based on the number of drivers. This minimum is never zero. With 16, the min. is 6 and it goes up 6 for each less driver in the room. By using a multiplier based on this difference in available points, I should be able to come up with a formula that will accurately compare the performance of drivers in different divisions with a different number of racers in each one without having any noticible bias based on room size.
, so I wont be able to race on Sundays for a couple of weeks. I let you know when I can make it back.
JLBowler
Premium
- 7,985
- California
- JLBowler
OK, I think I have something here:
Basically, I've applied a multiplier to the winners total points for the night based on how many drivers were in the room. Becasue each driver out of the room makes a six percent difference in the minimum number of points available, the smaller room makes it easier to score more points.
This system takes into consideration the fact that we use a reverse grid and it has a tendancy to turn some races upside down. It is also more difficult to score points in those races if you start at the back.
I have applied a difference of .03 for each racer above or below 10. In other words, if you have 10 in your room there is no adjustment. If you have less than 10 then you get a -.03 adjustment for each driver below 10 and if you have more than 10, you get a +.03 adjustment for each driver above 10. For ease of comparing at this point, all calculations are based on the number of drivers that scored points on the night and it's not adjusted for number of drivers in each race.
I have run the numbers for the last four weeks and the system appears to work quite well from what I can see. The numbers are a little skewed due to dominant drivers in certain divisions but that is not the fault of the formula.
Before applying the adjustment, the lowest three winning totals for the month were 79, 79, and 81. After the formula, the three lowest winning totals are 79, 81.37, and 83.
Before adjustment the three highest totals were 100, 98, and 93. After adjustment the three highest are 100.94, 96.46, and 94.
With adjustments applied the room with the lowest turnout would have finished 3rd, 1st, 4th and 5th (same number in two rooms), and 2nd.
With adjustments applied the room with the highest turnout would have finished 1st and 5th (same number in two rooms), 2nd, 2nd, and 1st.
The original formula I proposed would have made the lowest attended room last and the highest attended room first most of the time. This system doesn't have this extreme of a bias built into it.
The overall distribution of prizes for the month would have been as follows:
D1 2 prizes
D2 1 prize
D3 2 prizes
D4 4 prizes (3 by the same driver)
D5 3 prizes
With the same driver winning three times, some of those prizes would have been shifted to other divisions. Overall the distribution doesn't look to bad especially when you shift prizes away from repeat winners.
Once again, it's just a suggestion and feedback is appreciated. I'm always open to working the numbers if something is found to make it work better.
Basically, I've applied a multiplier to the winners total points for the night based on how many drivers were in the room. Becasue each driver out of the room makes a six percent difference in the minimum number of points available, the smaller room makes it easier to score more points.
This system takes into consideration the fact that we use a reverse grid and it has a tendancy to turn some races upside down. It is also more difficult to score points in those races if you start at the back.
I have applied a difference of .03 for each racer above or below 10. In other words, if you have 10 in your room there is no adjustment. If you have less than 10 then you get a -.03 adjustment for each driver below 10 and if you have more than 10, you get a +.03 adjustment for each driver above 10. For ease of comparing at this point, all calculations are based on the number of drivers that scored points on the night and it's not adjusted for number of drivers in each race.
I have run the numbers for the last four weeks and the system appears to work quite well from what I can see. The numbers are a little skewed due to dominant drivers in certain divisions but that is not the fault of the formula.
Before applying the adjustment, the lowest three winning totals for the month were 79, 79, and 81. After the formula, the three lowest winning totals are 79, 81.37, and 83.
Before adjustment the three highest totals were 100, 98, and 93. After adjustment the three highest are 100.94, 96.46, and 94.
With adjustments applied the room with the lowest turnout would have finished 3rd, 1st, 4th and 5th (same number in two rooms), and 2nd.
With adjustments applied the room with the highest turnout would have finished 1st and 5th (same number in two rooms), 2nd, 2nd, and 1st.
The original formula I proposed would have made the lowest attended room last and the highest attended room first most of the time. This system doesn't have this extreme of a bias built into it.
The overall distribution of prizes for the month would have been as follows:
D1 2 prizes
D2 1 prize
D3 2 prizes
D4 4 prizes (3 by the same driver)
D5 3 prizes
With the same driver winning three times, some of those prizes would have been shifted to other divisions. Overall the distribution doesn't look to bad especially when you shift prizes away from repeat winners.
Once again, it's just a suggestion and feedback is appreciated. I'm always open to working the numbers if something is found to make it work better.
Handlebar
Premium
- 3,969
- Herndon, VA
- HandlebarMustash
How does that compare to the current distribution, Bowler?
What is 'can be' is cool - how it is different from 'currently it would be...' is the data I suppose I'm asking for...
Currently, it is...
This month.... (insert data)
With this plan, it would look like
(data, different)
Supporting idea how these results are better - here
What is 'can be' is cool - how it is different from 'currently it would be...' is the data I suppose I'm asking for...
Currently, it is...
This month.... (insert data)
With this plan, it would look like
(data, different)
Supporting idea how these results are better - here
- 1,545
- California
- Push-The-DZ
Is it possible for two people to move up a division?
kcheeb
Premium
- 4,753
- Alberta
- kcheeb
- kcheeb
Hey SoCalRacer310,
Thanks for joining S.N.A.I.L. 👍 Can you let us know which PSN ID you'll be racing with?
My apologies for not responding sooner, life has not been predictable lately.
Please read the following:
We run a clean league by enforcing the S.N.A.I.L. OLR and we expect all of our drivers to know and follow The Good Racecraft Guide.
Please become versed in both.
Follow These Steps:
1. I add you to the drivers list as a New S.N.A.I.L.
2. You run the Time Trial.
3. You Private Message my GTPlanet account with your time from the time trial (no Sunday racing until this is completed).
4. I assign you to the appropriate Division, to give you the best racing possible.
5. A PSN friend request will be sent from the appropriate SNAIL_Division1, 2, 3, 4 or 5 account. That is where you race on Sunday night.
6. You drive fast and clean on Sunday
The first post has all the rules and regulations for S.N.A.I.L. It is the first place you need to check if you have a question. If it's not answered there, feel free to post your question on the thread.
The second post has the current race lineup for Sunday and is updated weekly.
If you have a preference for car/wheel colour and racing number, post your choices in this thread.
To avoid duplication, the currently claimed combinations can be found here.
During the week we run a number of different events, we encourage all S.N.A.I.L.s to join as many as possible.
Welcome aboard and enjoy
- 4,499
- Tampa, Fl
- Ice_Warden
Yes it happens quite often actually.
Handlebar
Premium
- 3,969
- Herndon, VA
- HandlebarMustash
Interesting point, jobyone.
Let's say that forcedinduckshun gets put into D4.... should he get a pass for all 4 weeks, choose the tracks, and is that in the best interest of all 45+ racers? One guy who's in the wrong division calls the tunes and we all dance?
I'm not in total disagreement, I'm just pointing out a possible flip side.
Let's say that forcedinduckshun gets put into D4.... should he get a pass for all 4 weeks, choose the tracks, and is that in the best interest of all 45+ racers? One guy who's in the wrong division calls the tunes and we all dance?
I'm not in total disagreement, I'm just pointing out a possible flip side.
JLBowler
Premium
- 7,985
- California
- JLBowler
The distribution would have been the same under either system. The difference lies in when they were won, which prize was won, and how close the adjustment makes it. Under the current system, the room with the lowest participation won prize A 3 out of 4 weeks. Under the adjustment system the lowest participation room won prize A 1 out of 4 weeks.
The average differnece from high point division winner to low point division winner under the current system was 18 pts. With the adjustment is would have been 11.70.
The instances of a division winner going from a prize to no prize from system to system are as follows:
Nov. 4 D1 89 pts. with 10 drivers (second low of the night)
Nov. 11 D5 88 pts. with 10 drivers (low of the night)
Nov. 25 D3 86 pts. with 9 drivers (third low of the night)
The instances where a division winner went from no prize to a prize would be as follows:
Nov. 4 D3 87 pts. with 11 drivers (second high of the night)
Nov. 11 D1 84 pts. with 14 drivers (high of the night)
Nov. 25 D5 86 pts. with 10 drivers (high of the night)
From those two lists, you can see how the system works. Each of the division winners that would get a prize under the new system had what looks like a better night than the driver that would have lost a prize. 87 w/ 11 is better than 89 w/ 10. 84 w/ 14 is better than 88 w/ 10. 86 w/ 10 is better than 86 w/ 9.
It's not a huge difference but it does a pretty good job at limiting the advantage a driver would have in a division with lower attendence.
- 760
- DrKronin
And that's a very valid perspective we haven't discussed much (someone did suggest rewarding the lowest scoring division winners earlier today, but I don't see that message right now, so I can't remember who).
It does seem a little weird to give prizes to the highest scores when the goal of the divisions in the first place is to promote close racing. It's almost like rewarding people for finding mistakes in the system.
That said, if no one posts scores proving them too good for their division, kcheeb has nothing to do. Consider it a Canadian employment program
Last edited:
Oshawa-Joe
Premium
- 5,061
- Hillsdale
- Oshawa-Joe
Maybe every second week the prizes go to the drivers with the least number of points.
- 3,613
- New England
- jobyone71
I totally understand that. I would say that if someone scores a perfect 100 in any division other than D1 they should probably be moved up to the next division. If it happened twice in one month in one division they are truly in the wrong Division. In my mind a perfect 100 should be very rare. I think I have only seen 2 since I joined. The most recent was Danger and I think we will all agree he is too fast to be in D4. But I still feel it should be rewarded.
Now if someone in D1 gets a perfect 100 lets just let some air out of their tires or give them some ballast
Handlebar
Premium
- 3,969
- Herndon, VA
- HandlebarMustash
Billy Connolly may have had it right, as a punchline, Joe, but I really like Bowler's idea.
Getting 100 points if it is 1 vs 1 is not like 88 pts in a 14 man grid.
Size matters.
Go big to win big.
So if I have a small grid in my room, and I have no control over it... Scoring is WAI (working as intended)?
That being said, I like Bowler's idea... Lots.
Finesse and nuance.
Getting 100 points if it is 1 vs 1 is not like 88 pts in a 14 man grid.
Size matters.
Go big to win big.
So if I have a small grid in my room, and I have no control over it... Scoring is WAI (working as intended)?
That being said, I like Bowler's idea... Lots.
Finesse and nuance.
Handlebar
Premium
- 3,969
- Herndon, VA
- HandlebarMustash
I totally understand that. I would say that if someone scores a perfect 100 in any division other than D1 they should probably be moved up to the next division. If it happened twice in one month in one division they are truly in the wrong Division. In my mind a perfect 100 should be very rare. I think I have only seen 2 since I joined. The most recent was Danger and I think we will all agree he is too fast to be in D4. But I still feel it should be rewarded.
Now if someone in D1 gets a perfect 100 lets just let some air out of their tires or give them some ballast
Under the current system, that means that racer gets the A and B choice, and the next two weeks, nobody gets squadoosh. I like the current system more than the old D1/overpowered FR only / great for that end of SNAIL model.
The idea of a one month for a division system is good, but recognizing that there needs to be an adjustment if a racer is misplaced rather than having an entire division, or all of SNAIL, becoming a whipping boy with no recourse.
To be clear, I am not dissing Dangerzone, I am thinking of precedents and how I hope things would work to be fair.
- 1,545
- California
- Push-The-DZ
- 330
- Dearborn, MI
- weaver-2009_
Anything interesting going on tonight? I'm as bored as Michael Vick at a PETA meeting...
