Season Points Modeling

[xxAGAVExx]

Sent from my iPhone using Tapatalk

[RedBRZ80]

But the points system is there for basically two reasons. One- the competition. But two, I think a good system keeps people engaged and coming out.. If you reward 1st place too much.. Then others will drop off(could happen)

Again just 2 cents

[Loren]

1st place has always been 3 points ahead of 2nd, and nobody complains about that. (or much of anything else, really, once we got the drops settled) What I don't like is that small classes get overrun easily, and large classes end up with a LARGE number of people who are hopelessly in the weeds, even if they're driving pretty well. Those are the issues that *I* would like to address. Nobody else is instigating this at all.

Small classes aren't an issue in F1 (where our points system originally came from), but even they must have seen the "large class" issue because when they changed from the old 9-6-4-3-2-1 scoring, their new method awards points 10-deep instead of 6-deep.

[RedBRZ80]

I really like the fact that it's just discussed and thought about. Always fun to work through these things.

Martin sports car club rewards points based on how many show up in class, maybe a hybrid of your current system and a modifier for large classes. / I think your on to something there. It's worth more to beat a 20 car field then a 3 car field.

Nothing is ever "fair" cause you can't make everyone happy, but it's still interesting to discuss.

[Amphoteric88]

OK, so i did a (relatively) quick plug and chug of the first few methods. I used a mix of season points results from last year and current standings to come up with a 4 competitor class and a 22 competitor class to run these methods through. Since I set this up last night and couldn't be bothered to look through every individual autocross result, I only got accurate position information for the smaller class; the first five positions I could tell from the season points logs but everyone that received a "1" I just assumed a position for them. Later on I'll refine this and fill in the last page, the results page, to yield more accurate finish information and therefore a more accurate analysis of these methods.
https://docs.google.com/spreadsheets/d/ ... sp=sharing

EDIT: I also have the Excel file; the Google Docs way seemed to be quicker and easier to share

[Loren]

I was hoping someone would do that. Thanks!

At a glance, it looks like 1B does what I expected it to do with the larger class. Doesn't do as much with the smaller class, though it does reduce the spread a bit.

2A actually worked a LOT better than I expected. The points distribution is a lot better, I really like that. But, it still has the downside of penalizing someone who happens to miss a very large event. Their competitor who finishes well at that event will get a points advantage that they may not be able to match. Our 3 dropped scores might be enough to sort of compensate for that. The sample data large event has events with 12 to 23 entries, so it's a good test for that. Just need to play with what happens if one of your top finishers misses the largest event. The low score will drop, and if they have enough other high finishes, it may work out.

Also, we need to consider, what if we have something like 3-4 top contenders (such as M2) who are constantly swapping finishing order at each event. Will there be enough of a points spread to keep them from a tie at the end of the season? May need to add bonus points for top 3 or something like that to compensate. Like extra 4 for first, 2 for 2nd, and 1 for 3rd. Something like that. Just to put a little more spread between the top finishes.

2B is too complicated, I don't want to consider it. But, if we can fix the problems with 2A (or convince ourselves that it's not really a problem at all) without adding complexity, might be worth considering. I really like how it spreads the points out throughout the class.

I think some variant of 2A or 1B are the ones worth looking into the most.

If you're going to have a wide variance in the number of events attended in your sample, you need to be sure to incorporate dropped events into the scoring. Current method is to drop the lowest 3. We should stick with that (we had big discussion about it before, and it makes the most sense for our group)... unless we opt to drop the lowest two and the highest one instead. (which would take some analysis to see if it really works)

Something to include in your spreadsheet would be "points total for this method" and "points total for current method" so that we can easily compare what has changed. And if the new method is doubling points, make it "points for current method x2" for a better comparison. I was flipping back and forth between pages to see that. Easier if you can get that info all on one page.

Also "final finish position for this method" and same for "current method" so that we can see at a glance if finish positions change. Ideally, we'd probably not want the top 3-4 to flip-flop (unless they were REALLY close before, then maybe... if it makes sense), but below that... maybe some of the middle ones would move around a little because they're getting more than 1 point?

Trying not to be overly critical here. I appreciate your effort, and glad you got to it before I did!

We're in no big rush here. If we're going to do anything at all, we need to sort it out by May-June. Plenty of time.

[Amphoteric88]

Glad you appreciate it! Like I said, I just cobbled something together in an evening. Being out of school and having time on my hands lets me fiddle with stuff like this
Filling in 2B was a pain, what with having to sum up the three or four applicable point rules. Glad that one's out of consideration for now; simpler is better!

I completely forgot about dropped events, and adding in the comparisons to the current system will certainly aid those viewing the data. Like I mentioned, tieing each Dr. # to a finish position for each event and then compiling the results will aid in providing some real comparisons.

Once I focus on refining those two methods (1B/2A) we can get to changing other variables, like a close season battle or deciding bonus points for podium positions.

Also, anybody can fiddle with the spreadsheet I've linked. I have the Master Copy on my computer, so that's what I'll be playing around with. You guys get to share the other one

[Loren]

Yeah, I poked around with things a little bit as I was discussing yesterday. (and in doing so, changed quite a bit of what I was saying on-the-fly!) Playing with a mock-up definitely aids in understanding how things will work.

[Loren]

Bringing this discussion out of the closet. If we're going to make a change, we should get it sorted out sometime soon.

Any more thoughts?

[twistedwankel]

Since it took me 4 months to notice that there are now 12 Open events I think I'd better let you all do the math.

It did occur to me that with all the clubs I occasionally run none of them place me in their yearly points unless I am a member of their club. That requires a paid for membership so I am never in their yearly points. However, in their actual events I am scored the place I finish that day with appropriate trophy if desired for that finish.

Since FAST considers anyone who has run 5 Open events(?) a "member" maybe we should do the same thing? Stickers for the event place but no yearly points if they don't do 5 events/year? Just a thought.

Even when I went to the Ft Meyers State Championship it was also their normal "club event" and it stated in rules that non-members would not be tallied in their yearly club points. (The word hotshoe was actually used to describe State Champions.)

[Native]

some variant of 2A or 1B are the ones worth looking into the most.

So I reread this twice. Can't tell which those 2 methods are?

[ImpostorDan]

I'm all for paying more people more points in large classes. Ultimately, the greater the differential (e.g. between 1st and last paid place) the more it eliminates trailing drivers. By narrowing the 1st to second point differential, you incent the driver to be 1st all the time every time (which results in more nerve racking fun and "get in their head games"). Given that FAST has a 3 event drop methodology, the casualness/flexibility is covered.

Thus, I'm for keeping it relative tight... 12,10,8,7,6,5,4,3,2,1,1,1,1,1,1.

[twistedwankel]

Guess I will only enter high enrollment events then to maximize my points? Would love to have gotten 5 points for my miserable finish last time when I netted only 1 pt. One must ask: "Do the same people always show up for these highly attended classes?" If not then the prize t shirt isn't the reason for their showing up.

[Loren]

some variant of 2A or 1B are the ones worth looking into the most.

So I reread this twice. Can't tell which those 2 methods are?

You're a step ahead of me. I'll try to find some time to review and recap what those methods are. And maybe play with Ricardo's mock-up spreadsheet a little, too.

[jev61]

Guess I will only enter high enrollment events then to maximize my points? Would love to have gotten 5 points for my miserable finish last time when I netted only 1 pt. One must ask: "Do the same people always show up for these highly attended classes?" If not then the prize t shirt isn't the reason for their showing up.

12 events minus 3 drops equals t-shirt. Easy peasy. Like David.

[Loren]

some variant of 2A or 1B are the ones worth looking into the most.

So I reread this twice. Can't tell which those 2 methods are?

Finally getting back to this!

The reason you didn't find that info is that it's buried in Ricardo's spreadsheet. Those were his internal references.

2A was simply reversing finish position. In a 10-car class, top finisher gets 10 points, bottom finisher gets 1. Simple. More effective than I expected it to be, but there are still things I don't like about it. The two key things I don't like are that class size at any given event becomes a big factor, and there's no "bonus" for finishing first. If a top competitor misses that one event where there were 8 extra competitors, they're another 8 points down. And, right or wrong, I feel like top dog should get at least a LITTLE points bump.

1B was a little more complicated. 1st was 12 pts plus a point for ever car beaten up to 18 max. 2nd was 8 pts, and down 1 point per position from there. I think that one could give the top and 2nd finisher TOO MUCH advantage. Otherwise, it could work, but it's too complicated.

SO... I pondered this some more and played with numbers a little, and I've come up with something that is very simple and could work.

Personal Goals: (in no particular order)
1. Total final season points in the vicinity of 100, preferably less in most cases.
2. Points compensation for larger classes down to at least 10th place.
3. Good competition (less points spread) in larger classes AND smaller classes. Make it less likely that someone can run away with the season before it's half-way done.
4. Simple math!

All of that should foster better competition through a larger segment of each class, rather than just the top 3-5 competitors.

I like the idea of basing points on class size, at least to some degree. So, what I've done is based points for first place on the class size, but put a cap on it. I like giving first place a boost, so the difference between first and second place is 2 points.

Here's the New Method:
1st Place Points = #Cars x 2 up to a maximum of 12
2nd Place Points = 1st Place minus 2
3rd and Below = one point less for each position, minimum 1 point

And here's a mockup of how it could work.

Points_Demo.pdf

(186.13 KiB) Downloaded 384 times

These points came from actual 2016 season points. R1 and M2, I believe. The names have been obscured to protect the innocent.

It seemed to have some good effects. Points spread is less, especially in the smaller class, which I think is great. Where clearly only the top 5 were in contention before, now competitors as low as 8th for the season are within three 2nd place finishes from a season trophy. Should make the competition more interesting even later in the season. The more "high, but below 5th place" finishes you had, the more it affected your total.

There were some position changes in the larger class, some of the position changes were rather significant and worthy of discussion. Since more points are awarded deeper into the field, it has the effect of rewarding those who attend more events. Where someone might have gotten 1 point for a 6th place finish before, now they get SEVEN points. So, the person who finished 2nd under the old method who had finished 1st or 2nd at exactly FIVE out of eleven events is now down in 5th place. Yeah, he's clearly a good driver. But, I think I'm okay with handing the Season awards to people who attend more events (AND drive well).

For the smaller class, it doesn't make as much difference. But, if a driver is SOLO in a class, they only get 2 points rather than 9. That's significant. Again, goes toward keeping anyone from easily running away with the championship without a challenge.

Have a look at that, play with the numbers, see if it looks feasible. I might be missing something.

[jbrannon7]

Looking at the current season standings there are 2 or 3 people in each class that are in contention. The rest either consistently finish well off the podium or only attend a few events. So if your goal is to make the points system more fair, then you have to take points from the podium finishers and move them to people finishing 4th or lower. I don't see any need for a points change, unless you want to have 4-6 people "in the hunt" by giving the people that attend fewer events more points. Points racing is about attendance as much as being first. If some people only attends 60% of the events they should not be in contention for the championship podium.

[ImpostorDan]

Points racing is about attendance as much as being first. If some people only attends 60% of the events they should not be in contention for the championship podium.

Agree here - if you feel that folks with less attendance should be competitive, this should be a change in the event drop rule. (IMO)

[Z1NONLY]

As someone who doesn't attend enough FAST events to compete for the season, I agree. People who want to win a season need to put in the "work" and show up.

[Loren]

Agree here - if you feel that folks with less attendance should be competitive, this should be a change in the event drop rule. (IMO)

Wait a sec, did you guys misread something that I typed? Or did I type something backwards?

No. Here's what I said:

Since more points are awarded deeper into the field, it has the effect of rewarding those who attend more events. Where someone might have gotten 1 point for a 6th place finish before, now they get SEVEN points. So, the person who finished 2nd under the old method who had finished 1st or 2nd at exactly FIVE out of eleven events is now down in 5th place. Yeah, he's clearly a good driver. But, I think I'm okay with handing the Season awards to people who attend more events (AND drive well).

I'm totally on board with what you guys are saying, and this latest method HELPS that. The current system allowed someone who attended 5 events to finish SECOND in M2 last year. The new system is less likely to allow that.