Season Points Modeling

[Loren]

We're not, I repeat, NOT making changes to how points are done this season. This is just an thought exercise for maybe a change next season.

I always look at the points for smaller classes (and really large classes) and think "something's not right". But, ANY change we make to the points system has far-reaching implications that are hard to see unless you actually mock it up and go through many possible scenarios to be sure that the system works the way you want it to. (rewards those who are winning the most, but doesn't allow anyone to "run away" with the points unless they've REALLY earned it... fosters good season competition, which is fun)

Even just deciding how many events we should drop in a season gets pretty complicated. You can see where we discussed that here back in 2009.

I'm not going to go through all of the analysis right now, kinda hoping someone else has the motivation to do it. But, here's my thought:

Current points are:
1st - 9
2nd - 6
3rd - 4
4th - 3
5th - 2
6th and below - 1

SMALL CLASSES
To put it simply, if the class only has 3-4 competitors, and they don't all show up regularly, somebody who might not necessarily be a top driver in their class is getting "first place points" merely for showing up. We've accepted that as a part of the game for a long time. You have to show up to win!

Anyway, here's my first thought (METHOD 1): Make the default points for 1st place 6 points, just like 2nd place. And then ADD one point for each car that you finish ahead of, up to a max of 9. That gives the ModeratelyFastGuy who shows up on the day when nobody else is there, or when only one other driver is there LESS of an advantage over the ReallyFastGuy who's going to show up and win the next event.

For classes with 4+ drivers, nothing would change. That could work to help the Small Class issue, if we disregard the Large Class issue.

LARGE CLASSES
Classes like M2 lately... you could be REALLY good, be within a couple tenths of winning, and still only finish 8th out of 20 and get ONE point! Now, I'm sure nobody noticed, because it hasn't actually come into play yet, but we DO award season trophies down to 4th place if the total eligible at the end of the season is 10 or more. (that could happen in M2 this season) So, it's possible that these "lower points" could matter to somebody. Our current system essentially says "if you finished below 5th, you're not a contender", and that's not necessarily the case in a larger class. I think we could make the competition more interesting for a larger number of people in a class like M2 by awarding more significant points deeper into the field.

What to do? This one is tough. (and needs to be weighed against the Small Class phenomenon)

First of all, if you were wondering, our points system is very similar to, and possibly based on old F1 points. (I stole it from Wichita Region SCCA, which is where I started autocrossing in 1995) You can read about F1 World Championship Points for some background. The thing to note here is that F1 is designed around a 22-car field, they will NEVER have a field much smaller than that, and they will NEVER have a field larger than that. So, their points system doesn't need any flexibility built-in. Also note that F1 has gone from a "top 6" points scheme to a "top 10" scheme starting in 2010. Interesting.

How do we make something fit both a large class AND a small class? To my mind, that means "conditional rules". In some way, the points need to reflect how many drivers are in class and/or how many drivers you finished ahead of. But, it also has to be very simple and formulaic.

I'm going to start a new post to rattle off some ideas... they're just ideas.

[Loren]

IDEA: Reverse finishing order (Method 2)

24 cars in class = 24 points for first place, and down from there. I think this would be very sensitive to variations in class size. Would give the guy who won "that one event where the class was really large" an advantage that nobody else could match.

4 cars in class = 4 points for first place, and down from there. It could work, but not enough of a boost for finishing high. Also, doesn't really matter, but it would put the season winner's total points on a completely different scale than a larger class.

Don't like this idea.

IDEA: Incorporate "how many did you finish ahead of" (Method 3)

This one's fuzzy. Might end up being too complicated. But, maybe not.

What if we did one point for each car you finished ahead of up to a max of 10 (so if there are 24 and you finished 20th, you finished ahead of 4, so 4 points... you finished 1st through 14th, you max out at 10 points)... plus...

Another point from top to bottom through half the field with the top car getting points equal to half the field (rounded up to even number, divided by two) (so, 10 cars in class = 5 points for 1st, 1 point for 5th, zero below that)... plus...

One point for just being there. (bottom finisher gets one point) Plus...

4 bonus points for first, 2 bonus points for 2nd.

(I'm free-form rambling here, this might not work at all...)

For the 24-car class, that gives the bottom and middle of the class some additional points, which I think is good. At the top,
1st = 10+12+1+4 = 27
2nd = 10+11+1+2 = 24
3rd = 10+10+1=21
4th = 10+9+1=20
Down to
12th = 10+1+1 = 12
14th = 10+0+1 = 11
20th = 4+0+1 = 5
24th = 0+0+1 = 1

More analysis required, but some variation of this appears to have potential for a large class.

For a 5-car class, does it still work?
1st = 4+3+1+4 = 12
2nd = 3+2+1+2 = 8
3rd = 2+1+1 = 4
4th = 1+0+1 = 2
5th = 0+0+1 = 1

More analysis is required. I think it could work, but it's obviously class-size dependent. Would throw things off if the class size varies widely. Would be better if it scaled so that the top points were not necessarily the same, but at least closer from a smaller class to a larger class. (a class size can vary from 1-5 or from 14-25 from one event to another)

For a 2-car class?
1st = 1+1+1+4 = 7
2nd = 0+0+1+2 = 3

Hmmm... I know I just said more even top points is good, but I'm actually okay with this. It's not a complete "throw away" event for these guys, but it doesn't... and, arguably, shouldn't... count as much as an event where they had more competition.

And for a single-car in class:
1st = 0+1+1+4 = 6

Which is totally fine.

SO, there are some ideas
Not necessarily good ideas. Certainly not the ONLY ideas. Let me know what you think. If you have your own ideas, let's talk about them. If you've got time to do some more in-depth analysis, go for it.

We don't HAVE to change anything. We've been using the same method since 2007 and only changed the number of dropped events in 2009. So, it's not that this doesn't work. I just think (and have always thought) that it could work better given our variations in class size.

[AScoda]

1st thing that came to mind as a potential issue is when you have a class that is only occasionally large. If you have a regular competitor who is in a championship hunt and happens to miss that one event when 20 people showed up, their competitors pick up a huge points bonus that might be hard to make up.

I think your idea in the first post incorporates some bonus for larger classes without causing other issues.

[ImpostorDan]

Maff is hard!

I'm just going to throw out of left field here... What if season standings were accrued time lost by within class as a penalty deducted from 100 [seconds] per season?

I think this eludes the irregular attendance run away wins, avoids winning on participation alone and keeps large classes interesting with multiple competitive drivers.

Edit: I suppose there would have to be a baseline penalty for being a no-go. e.g. lowest driver + an additional something.

[Loren]

1st thing that came to mind as a potential issue is when you have a class that is only occasionally large. If you have a regular competitor who is in a championship hunt and happens to miss that one event when 20 people showed up, their competitors pick up a huge points bonus that might be hard to make up.

I think your idea in the first post incorporates some bonus for larger classes without causing other issues.

Yeah. I'm starting to think we need to universally "start from the top" so that top finishers get about the same points no matter how many people show up. (rather than a bunch more because extra people showed up at one event) With that in mind... I still like the "6+ how many did you beat up to a max number" idea. Maybe couple that with simply extending our points depth from 6 to 10, or maybe even 12-15?

Expanding on Method #1 (Method 1A):
Say we multiply our existing points by 2 as a starting point.

1st place = 12 + 1 point for each car that you beat, max of 18
2nd = 12
3rd = 8
And down one point per position to
10th = 1

Simple. Similar dynamics to what we already have. Gives points deeper for larger classes. Addresses points spread for smaller classes.

[Loren]

FYI, I just went and labeled all of the ideas presented so far with "Method #" so that we can refer to them easier without confusion.

[ImpostorDan]

I've been dwelling on this for hours on and off... I think I'm kind of missing the big picture philosophy / objectives. I can see that you're reading points totals and maybe the relative comparison across small and big classes has got you thinking? I suppose, what's the outcome (from last season) that has your attention?

In thinking about the general levers/dynamics, if I look at the macro range of today's system it's essential 9pts.
- the trophy advantage pattern by position is weighted 3,2,1,1,0...

By narrowing the weighting on the top positions, you'll emphasis consistent position placement to win the season.
By increasing the number of positions paid points, within the same limit, you'll increase the number of drivers in the hunt
By increasing the range of points assigned, you'll penalize those who fall out of the trophy range and irregular attendance; inconsistent drivers will get bumped out of the running faster.

Today, as long as I show up, I can only fall behind -8 pts; if I skip I'm penalized 1 more point.

In thinking about the generalities of the suggestions above and making skewed point scheme changes:
Scheme 1: 9,8,7,6,5,1 - change result, if you are consistently good, you will be in the hunt for longer in the season
Scheme 2: 9,8,7,6,5,4,3,2,1 - change result, more in the hunt, catch some talented, but inconsistent drivers + above
Scheme 3: 24,21,18,16,14,12,10,9,8,7,6,5,4,3,2,1 - The bottom half are now further disadvantaged (-9 increases to a potential of -24), 3 bottom places outside of drops or a short season of attendance is essentially catastrophic. Inconsistency is going to get dinged hard.

As a participant in M2, I'd like a tighter scheme with more positions to keep it interesting for the local level; keeping in the spirit that there are no trophies for showing up, I could see a kicker on 1st, but maybe narrow it to a point - all to keep the heat on for the duration of the season.

9, - 7, 6,5,4,3,2,1.

In a small class, narrowing keeps the pressure on as well. Even if there's just a "duel" between a couple drivers, nobody is given the run away. Now, when there's not enough to fill the trophy list, you could put a penalty kicker in. 9, 7, 6, 1 - e.g. don't be last! I'm not sure if that really addresses the concerns.

Now - the real opportunity exists for Classic events. There should be a point system for those with:
- Lost to the fastiva penalty
- Tied times to the thousand get 2x points
- You beat Drew bonus

[xxAGAVExx]

I like method 3, but maybe you limit it to a certain number of entrants, as an example I'll use 15. So for the 24 car scenario, maybe it only applies to 1-15 and everyone above that only gets 1 point and the 1-15 finishers get their method 3 points but only calculated as if 15 cars were in the class. This way it hedges the advantage someone would get if there happened to be an abnormally large turnout.


Sent from my iPhone using Tapatalk

[twistedwankel]

I still like Rock, Paper, Scissors.
THEN just have a ton of fun for the rest of the season.
You just cannot beat FAST for FUN.

[Loren]

I like the way you're thinking, Chad. That could work, and resolve some of the downside of Method 3. But, the other downside is that it's a little complex. People tend to like simple. I'm still liking Method 1A the most right now. It's got simplicity just like we have now.

Dan, I knew I could count on you for some good analysis! That's a good way to look at things. Difference between Max and Min points in a class.

Currently, 1st is 9 and 6th is 1. 6th place gives you an 8 point deficit.

The more I look at Method 3 (even though I dreamed it up, myself), the more I don't think it's a good choice. Gonna disregard that one for now.

Method 1A is the one worth looking at.

With 1A, 1st is 18 (or as little as 13 for a 2-car class), and 10th is 1. 10th place is a 17 point deficit (but, we've also doubled our points total), and 6th would be 5 points, or a 13 point deficit (equivalent to 6.5 points now... so it puts them closer than the 8 points of the current system.

If I run out of other things to do, I'll fire up Excel and do some mockup scenarios and see how things shake out. But, I don't see method 1A "breaking" anything, merely making a few people lower down in the field a little more competitive.

Dan, as to your question of "what's wrong", it's mostly the small classes that I see. We're starting to see more people in M4 and S5, but they used to be really small, along with R1, and often M0. When you have a 3-4 car class, at the end of the season, maybe half of them will qualify... and it often comes down to who attended more events. Specifically, who attended an event or two where they competed alone, or against one competitor. If they're alone, they get a 9 point advantage. If they have one competitor, they get a 3 point advantage over that guy, and both of those competitors are 6 points ahead of the rest of the class. Method 1A just effectively makes an extremely small turn-out event less of a factor, and reduces the spread between 1st and 2nd.

The stuff about bigger classes isn't really a "problem", it's just a "gee, wouldn't it be nice if". And it's mostly inspired by M2 over the past couple years.

[Loren]

I still like Rock, Paper, Scissors.
THEN just have a ton of fun for the rest of the season.
You just cannot beat FAST for FUN.

We do have a ton of fun, and I don't think anybody takes things TOO seriously. But, a little friendly season competition adds to the fun for a lot of people, and expanding the "possibility of winning" from the top 5 in class to the top 10 in class can't be a bad thing when we've got classes with 20+ cars sometimes.

[RedBRZ80]

Just my quick 2 cents.. Go into JAX Solo group and look at points. It's basically a % off first system. So 1st always gets 100. And if you finished .05 seconds back you'll probally get 99 points.

However. I don't think it rewards winning as much as it should. My advice to my son who won a class champ this year in my old car was simply put. Just win man. Nobody can beat 100 points. And he won a lot of events by very small margins.

[Loren]

David, I actually like the idea of doing it that way. Yeah, it doesn't reward first place finishes as much IF the competition is close. But... if the competition is that close, maybe the points spread between first and second shouldn't be all that great? But, the problem (for us) doing it that way is that we don't have a system in place to "read" the results and calculate that stuff for us. I'm not interested in building such a system right now. And, even if I was, I still know that people prefer to be able to look at the results and EASILY verify their points, or think ahead about their points, or whatever. Too much math muddies the works.

Unrelated anecdote: Back around 1998, I created a local class called the DIC - Driver Indexed Class. It worked a lot like a PAX class, but for cars that weren't necessarily prepared to the full extent of the rules (PAX is really pretty useless unless all the cars being compared are THE car for the class and are prepared to the limit of the rules). What I did was give every driver their own index based on THEIR results over the past 6 events (I think it was 6, may have been 3). Their best time vs FTD for that day was their index, and we averaged the last 6 events. So, the driver's index could change a little bit from one event to the next. It worked REALLY well and the competition was really tight. But, people were frustrated because until I crunched the numbers (which I wasn't able to do at the event at the time, we were still writing times on paper cards), drivers had no idea what their finish position was. That's just one example of how I learned "what people expect" from results. In a nutshell, they want simple answers. "Where did I place?" "How many points did I get?" "Where must I finish next event to maintain my goal for the season.

"Just win" is an admirable goal, and it's what we should all be trying to do... just go as fast as we can. But, sometimes knowing that you're "close" will inspire just a little more effort.

[twistedwankel]

Don't include anyone in the yearly final points until they have completed at least 3 open events. This avoids the occasional super fast T & T people from messing up the yearly points for the amateurs. We have 3-4 events in before anything begins to shape up anyhow. Less work for Loren?

I do agree with Dan closing up the points for each position makes sense going further into the lower group. 9,7,etc.

**Off topic: What would make Jamie actually take a sticker at an awards ceremony? FTD?

[Loren]

I hear what you're saying, Doug. I've struggled with it, myself. I always come back to "if someone shows up and THEY can drive faster than me in a car that's legit for the class, then I didn't DESERVE top points that day". Sure, it messes up our little local points battle... and it does happen from time to time... but, I'm not sure we need to regulate it.

I've also briefly considered making the actual championship "members only", meaning you had to do 5 events in the PRIOR season to qualify. But, there are a lot of reasons to not do that. Not the least of which is that it's not fair to the new guy who starts within the first few events of the season and IS good enough to win something.

Nah, I think if the competitors are there, they should be counted.

We just need to start placing the stickers on Jamie's car for him.

[Jamie]

**Off topic: What would make Jamie actually take a sticker at an awards ceremony? FTD?

I take plenty of stickers...to deliver to other people. And don't mess with my car.

[Evil MS3]

Just my quick 2 cents.. Go into JAX Solo group and look at points. It's basically a % off first system. So 1st always gets 100. And if you finished .05 seconds back you'll probally get 99 points.

However. I don't think it rewards winning as much as it should. My advice to my son who won a class champ this year in my old car was simply put. Just win man. Nobody can beat 100 points. And he won a lot of events by very small margins.

I like this idea and it rewards close competition pretty well from the sounds of it. Winning overall should definitely be rewarded, but if you're behind the car that placed before you by a small fraction of a second, a big points hit hurts.

[RedBRZ80]

Just my quick 2 cents.. Go into JAX Solo group and look at points. It's basically a % off first system. So 1st always gets 100. And if you finished .05 seconds back you'll probally get 99 points.

However. I don't think it rewards winning as much as it should. My advice to my son who won a class champ this year in my old car was simply put. Just win man. Nobody can beat 100 points. And he won a lot of events by very small margins.

I like this idea and it rewards close competition pretty well from the sounds of it. Winning overall should definitely be rewarded, but if you're behind the car that placed before you by a small fraction of a second, a big points hit hurts.

The downside to a % off first system, is the fluke one bad event. - so for example, they allow two drops, so you miss two events(now I know attend more and this doesn't happen). But now every event counts and you are off pace and it's an event where a national champ skus UP first place by a lot. So bit of a perfect storm. But when 1st place is a moving target it happens.

Not saying I don't like it/ dislike it, just a 1 off downside.

[Loren]

I like this idea and it rewards close competition pretty well from the sounds of it. Winning overall should definitely be rewarded, but if you're behind the car that placed before you by a small fraction of a second, a big points hit hurts.

As they say "that's racing". That's how it universally works. You either win or you don't. The margin matters not. You finish ahead of the other guy, you won. To dilute that by trying to say "I only lost by a little bit" is to water down our form of racing, don'cha think?

[Evil MS3]

I like this idea and it rewards close competition pretty well from the sounds of it. Winning overall should definitely be rewarded, but if you're behind the car that placed before you by a small fraction of a second, a big points hit hurts.

As they say "that's racing". That's how it universally works. You either win or you don't. The margin matters not. You finish ahead of the other guy, you won. To dilute that by trying to say "I only lost by a little bit" is to water down our form of racing, don'cha think?

This is true, probably not the best approach then.