Yearlong County Contest Discussion Thread

Page 1 of 2 1, 2  Next

View previous topic View next topic Go down

Yearlong County Contest Discussion Thread

Post  dctim on Mon Jan 03, 2011 5:48 pm

Just thought I would separate discussion from hits claimed.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Tue Jan 04, 2011 8:58 am

dctim wrote:Just thought I would separate discussion from hits claimed.

Good idea. Thanks.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Tue Jan 04, 2011 9:53 am

Here is an idea - can you add a tab to your spreadsheet?

Something along the line of date, state, county, person claming.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Tue Jan 04, 2011 11:41 am

dctim wrote:Here is an idea - can you add a tab to your spreadsheet?

Something along the line of date, state, county, person claming.

Not sure what you are requesting...or if it is possible.

Currently, the spreadsheet has tabs for each state, by postal abbreviation.
The counties are listed for each state, when you click the state tab...for example, if you click the IL tab, you will see all 102 of Illinois' counties listed.

When counties are claimed, I put the name of who claimed it - and then highlight it with a color unique to that person, making it easier to count later on. Your color is green, Tim.

Check the NJ tab, and you will see that Burlington County is listed with your name next to it, and is highlighted green.

I could add the date claimed, if you wanted...but, other than that, I believe I already have covered everything you were actually requesting. Do you want me to add the dates? I can do that.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Tue Jan 04, 2011 12:50 pm

Actually, I was thinking that some kind of summary tab(s) might be helpful.

So the first tab could be those counties claimed thus far. Obviously, the most populated will tend to be claimed faster.

I don't object to what you have done so far by any means. I'm just trying to brainstorm.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Tue Jan 04, 2011 2:40 pm

About weighting. I see that mine is 0.3871. That most likely represents the amount of time I have been at TrackDollar since DiamondBack Dave has the same ratio.

I'm wondering if it wouldn't make more sense to weight on factor of My Bills Entered divided by Total Bills for al of TrackDollar.

The downside is that the ratio would be constantly changing. The upside is I have a better chance to have my hits count than the users with more bills in the system.

Back to the "extra tab(s) on your spreadsheet" for a moment. I think it would be cool to have who is leading the contest -- Name, Total Hits, Ratio and Factored Hits
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Tue Jan 04, 2011 4:15 pm

dctim wrote:About weighting. I see that mine is 0.3871. That most likely represents the amount of time I have been at TrackDollar since DiamondBack Dave has the same ratio.

I'm wondering if it wouldn't make more sense to weight on factor of My Bills Entered divided by Total Bills for al of TrackDollar.

The downside is that the ratio would be constantly changing. The upside is I have a better chance to have my hits count than the users with more bills in the system.

Back to the "extra tab(s) on your spreadsheet" for a moment. I think it would be cool to have who is leading the contest -- Name, Total Hits, Ratio and Factored Hits

I will add a tab to the Spreadsheet to show who is where.
As to the weighting...well, that was to be discussed prior to the start of this contest.
I weighted as I did, so as to give new users an equal chance.
As in the rules, IF this weighting system shows one or two users having a runaway, we will open to discuss other weighting methods, and then, if the majority of players agree to change the weighting, then we will.

For now, we should leave it as is and give it a chance to see what it does.

I don't want anyone to have a runaway, I want everyone to have an equal chance.

So let's see what this weighting does first...give it a few months and if it seems skewed, we can discuss alternate weighting methods.

Let's play by the rules I established, eh?

UPDATE: I have updated the map and the Spreadsheet, up thru dctim's claim on DuPage County, IL. I added a Standings tab on the Spreadsheet.
I am open to discussion about the weighting after a couple of months and we see how this works out. If it doesn't seem to be making a competitive balance, we will discuss alternate methods of weighting, and present the option/options to the contestants to weigh in on. Any change in the weighting, as per the rules, must be approved by at least 1/2 of the current contestants.
One possibility would be to add a second level to the weighting...and have a weighting based on 10,000 bills or something.
The idea here is...if a user has 6 months in and 10,000 bills in...his won't have as much circulation as someone who has been in 12 months and also has 10,000 bills.
So we do need the weighting factor for time on TD. We just may add a second level for bills entered using 10,000 bills as a base.
For example...my current weighting is 0.7059. I have approximately 3000 bills entered. (actually, 2,902) We would go by nearest thousand...so that the weighting would not change daily. So, for my purposes it would be 3,000. So I would get an extra weight of 10000/3000 - or 3.3333 to my weighting.
thus, my weighting would be 0.7059 x 3.3333 = 2.353.

Tim, then, with about 4000 bills entered...
He would have 10000/4000 - or an extra 2.5 on his weighting.
Therefore, his weighting would then be - 0.3871 x 2.5 = 0.9678

Again, this is, of course, open to discussion, but that was just an example of how additional weighting factors might work.
There does need to be weighting for time spent on TD, because your bills thus have that much longer to circulate to new places. The number of bills is not as large a factor as this, because the most important factor in spreading out your hits is TIME.
And that is why I initially had time on TD as the main weighting factor.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Tue Jan 04, 2011 5:06 pm

Thanks for your consideration on these hair-brained ideas. And thanks a lot for the new Standings tab.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Tue Jan 04, 2011 5:14 pm

dctim wrote:Thanks for your consideration on these hair-brained ideas. And thanks a lot for the new Standings tab.

No problem. I think it is important to discuss these things, and to make a fair and equitable weighting system that allows all an equal chance to win.

Incidentally, I added one more thing to the Spreadsheet, strictly for my reference...the word MAPPED after a highlighted entry means that county has now been colored in on the most up-to-date map. If it's highlighted, and does not have MAPPED next to it, then it has not yet been colored in on the latest map. that will enable me to not have to pull my hair out each time I do a map update.

Map updates will be done about once per week.
Spreadsheet updates will be done as is convenient - when I notice hit claims coming in.

So that everyone can stay up on it, I will include notation in the map section as to most recent spreadsheet update.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Thu Jan 06, 2011 1:22 am

OK, folks...

I don't like what I'm seeing in the early results on weighting. It seems the newest (me) has a bit of a lopsided advantage.
This may balance out over time, it may not.

I have therefore worked out a second weighting method. I am posting both weighting proposals in the standings, so we can get an idea of how they work out.

Weighting Proposal #1 is as follows....
Time weighting - if you signed up for TD in 2011, that is, will be here 1 year or less then your weighting will be 1.2
If you signed up for TD in 2010 - that is, will be here 1-2 years when contest ends, then your weighting will be 1.0
If you signed up in 2009 or prior, that is to say, you will be here 2 years or more, then your weighting will be 0.9

The reasoning behind this is that...you DO have an advantage being here longer (your bills have had more time to circulate and spread out) BUT...also bears in mind the average bill only circulates 18 months.

Enter weighting - this is based on how many bills you have so far entered...obviously, the more bills you have out there, the better your chances...and again, there is a top amount, seeing as bills eventually DO go out of circulation.
Less than 2,500 entered - weighting 1.2
2501-5000 entered - weighting 1.1
5001-7500 entered - weighting 1.05
7501 - 10000 entered - weighting 1.0
10001 - 12500 entered - weighting 0.95
12501 - 15000 entered - weighting 0.90
15001 - 17500 entered - weighting 0.85
17501- 20000 entered - weighting 0.80
Over 20,000 entered - weighting 0.75

We then multiply the two factors, to get a final weighting factor. that is then multiplied by number claimed.

I like the way the new proposal looks it seems more equitable, but we will give it a chance, say a couple months, to flesh out, and we can then see which weighting proposal we like better.

In the old proposal - the way we are currently doing it...I have as many as Markus, yet have over twice his score...that does not seem right.

To show perspective I added a ghost pacer and a ghost slacker to each proposal.
With the new proposal, the Pacer has twice the number of counties claimed, and thus - as should be...is in the lead....but not by an insurmountable amount.

I think we let these run for a couple of months, say, till the end of February and see which weighting system works better.

We want this to be fair for everyone...and that means the long-timers, too.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Thu Jan 06, 2011 9:10 am

I noticed "Ghost Pacer" and "Ghost Slacker" in your spreadsheet standings. Do they represent comparrisons for your scenarios?
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Thu Jan 06, 2011 9:22 am

dctim wrote:I noticed "Ghost Pacer" and "Ghost Slacker" in your spreadsheet standings. Do they represent comparrisons for your scenarios?

Ghost Pacer represents someone with twice as many claims as the highest actual participant. Slacker represents someone with 1/2 as many as the lowest amount claimed (but no lower than 1) these are there merely for comparison purposes.

I think we would all agree that someone who has twice as many as the next guy OUGHT to be in first place...but not by an insurmountable amount....likewise, someone with only 1/2 the lowest OUGHT to be in last...but not by so much that they cannot hope to climb up a notch or two if they step it up a bit.

So we are seeing, by using them...if this bears true or not. Just a small visual aid to see if the weighting, or handicapping...is doing what it is supposed to do...which is to give everyone an equal chance, allow no-one a runaway, and not take away a win from someone who truly deserves to win...this handicapping is NOT easy!!
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Thu Jan 13, 2011 11:20 am

I thought of two other possible weighting scenarios, I will add them to the Standings List.

The first is based on Track Points.
The second is based on Hit Rate.

Proposal 2, then, Track Points, would work like this...

0 - 2500 Track Points - weighting = 1.25
2501 - 5000 Track Points - weighting = 1.2
5001 - 7500 Track Points - weighting = 1.1
7501 - 10000 TrackPoints - weighting = 1
10001 - 15000 Track Points - weighting = 0.9
15001 - 20000 Track Points - weighting = 0.8
Over 20,000 Track Points - weighting = 0.75

Proposal 3, then, Hit Rate, could look like this...

Less than 1% Hit Rate - weighting = 1.25
1.x% Hit rate - weighting = 1.2
2.x% Hit Rate - weighting = 1.15
3.x% Hit Rate - weighting = 1.1
4.x% Hit Rate - weighting = 1.05
5.x % Hit Rate - weighting = 1.0
6.x% Hit Rate - weighting = 0.95
7.x% Hit Rate - weighting = 0.90
8.x% Hit Rate - weighting = 0.85
9.x% Hit Rate - weighting = 0.80
10% Hit Rate and Higher - weighting = 0.75

I'll post these up, too...just to see how they look.

We could also try a combination of any of these factors...I kinda like the idea of Hit Rate and time weighting as in Proposal 1...combined into a weighting scenario. It seems reasonable. but, again, I just want to see what these numbers look like a couple months from now, and see what everyone thinks is the fairest weighting method...the one that produces the best chance for everyone. and which does not allow a runaway by a long-timer.

This is a bit of a learning curve, this year...but when we come up with the right formula, it can then be used year after year...and maybe in other similar contests.

We also could fiddle with the figures a bit...instead of using increments of 0.05 as I did here, maybe increments of 0.02...just have to see what produces a good weighting factor.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Thu Jan 13, 2011 8:08 pm

I like how you have been crunching the numbers looking for the right fit. It looks like all the scenarios have pretty much the same result.

I'd love to see Darell, Aaron and others jump into this contest to give us a better representation of if you are on the right track.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Thu Jan 13, 2011 8:57 pm

I'd like to see them jump in...to drive up the prize pool. Damned if I got the numbers right or not, we will get there.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Tue Jan 18, 2011 1:07 pm

It's early yet, but I am liking the looks of a weighting proposal based on 2 3 and C.
We could fiddle with the figures a bit as to the levels, but this would give a weighting to Time, Track Points, and Hit Rate.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Wed Jan 19, 2011 1:36 pm

Well, even this early, I think it is clear that the original weighting scenario is a total bomb....we're talking UBER fail...
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Thu Jan 20, 2011 12:53 pm

Kalisiin wrote:Well, even this early, I think it is clear that the original weighting scenario is a total bomb....we're talking UBER fail...
<smile> Don't be too hard on yourself - at least you have several scenarios to try.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Thu Jan 20, 2011 8:20 pm

dctim wrote:
Kalisiin wrote:Well, even this early, I think it is clear that the original weighting scenario is a total bomb....we're talking UBER fail...
<smile> Don't be too hard on yourself - at least you have several scenarios to try.

Yeah, but I'd like some input as to what seems most equitable.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Fri Jan 21, 2011 11:23 am

We may still be in the position of too few data points with 4 players and 42 total hits.

I played around a bit and came up with the following:

Name Entered Hit Entered Ratio Hit Ratio Avg ALL Ratios Hit * Avg
Kalisiin 3,346 16 0.1068 0.3810 0.8180 13.0873
Markus 14,790 14 0.4721 0.3333 0.6779 9.4904
dctim 3,724 10 0.1189 0.2381 0.7105 7.1051
Diamond Dave 9,471 2 0.3023 0.0476 0.6984 1.3968
Totals 31,331 42

The Entered Ratio is Current Bills entered by a participant divided by Total Bills entered by all participants.
The Hit Ratio is Current Hits entered by a participant divided by Total Hits entered by all participants.
The Avg ALL Ratios are these two ratios Averaged with the other propose ratios (Original, 1a, 1b, 1c, 2 and 3).

If we all had the exactly 20 hits, Markus, Dave and I would be separated by just 1.12 and Kaslin would be in the lead by 1.56.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Fri Jan 21, 2011 12:17 pm

dctim wrote:We may still be in the position of too few data points with 4 players and 42 total hits.

I played around a bit and came up with the following:

Name Entered Hit Entered Ratio Hit Ratio Avg ALL Ratios Hit * Avg
Kalisiin 3,346 16 0.1068 0.3810 0.8180 13.0873
Markus 14,790 14 0.4721 0.3333 0.6779 9.4904
dctim 3,724 10 0.1189 0.2381 0.7105 7.1051
Diamond Dave 9,471 2 0.3023 0.0476 0.6984 1.3968
Totals 31,331 42

The Entered Ratio is Current Bills entered by a participant divided by Total Bills entered by all participants.
The Hit Ratio is Current Hits entered by a participant divided by Total Hits entered by all participants.
The Avg ALL Ratios are these two ratios Averaged with the other propose ratios (Original, 1a, 1b, 1c, 2 and 3).

If we all had the exactly 20 hits, Markus, Dave and I would be separated by just 1.12 and Kaslin would be in the lead by 1.56.

Nice work.
Perhaps this would be an ideal weighting scheme, then.
It appears as if it would at least get us all pretty close if we all had the same number of hits claimed...but would give slight advantage to newer players (who would normally expect to get fewer than older players)

Of course, I get a bit of an advantage in PA, as I'm more likely to get those than anyone else...Markus is fairly likely to get a lot of Texas...and split VA with dctim...And Diamondback is likely to get a fair number in Arizona...just because of proximity.

Maybe we take those ratios you came up with...and then come up with a way to balance out the "home field advantage" seeing as I get 67 counties in PA that I'm more likely to get...whereas Markus gets 254 in Texas...and Dave only gets 16 in AZ.

So maybe we use your formula...plus one more tweak to leaven out the homefield advantage thing.

You have to split with Markus on VA, tim...but 1/2 of 134 is 67...the same number as in PA.
Markus, on the other hand...has the 254 advantage, where Dave just gets 16.

(Note: Yes I know Markus lives now in VA, but he's got a fair number of bills floating around in Texas...)

Granted some counties in TX are not likely to ever get hit..like Loving County, for example.

Something to toy with, anyway.

I don't want to be the one making all the "rules" from up on high, as it were, I want active participation in this...and I am glad to see you have played with some figures a bit there, tim.

Of course, the ratios might change for a player mid-year if his track points, number entered, etc, changes.

So...for your additions, we'd need to come up with a range, again...so that ratios are not changing daily.

For example, if you have 10 percent of the bills entered you get x weighting, if you have 6 percent, you get y rating, etc.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Fri Jan 21, 2011 1:33 pm

The Entered and Hit Ratios certainly would not have to be a daily thing, or even weekly. They could follow your Map updates on the 10th, 20th and 30th of the month, for instance.

I like your idea of
For example, if you have 10 percent of the bills entered you get x weighting, if you have 6 percent, you get y rating, etc.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Sat Jan 22, 2011 12:05 am

dctim wrote:The Entered and Hit Ratios certainly would not have to be a daily thing, or even weekly. They could follow your Map updates on the 10th, 20th and 30th of the month, for instance.

I like your idea of
For example, if you have 10 percent of the bills entered you get x weighting, if you have 6 percent, you get y rating, etc.

Could. I tend to try to update the map once a week.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Kalisiin on Fri Jan 28, 2011 9:27 am

In light of my recent health mishap...I think I prefer a weighting adjustment occurring once per calendar month...but doing it at a different time each calendar month so that nobody holds back from entering bills or anything to attempt to manipulate their weighting.

But more than once a month would just not be good. This is a fairly large single-person undertaking, and I did just get taken out for a week...and now I'm on a slow recovery curve back...normalcy shall not be achieved very quickly in my case, I really bit off a big one this time. Healthwise, I mean.
avatar
Kalisiin
New

Posts : 2871
Join date : 2010-07-29
Age : 46
Location : Wendell, NC

View user profile http://www.kalisiin.com

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  dctim on Fri Jan 28, 2011 2:53 pm

I hope you feel better soon.
avatar
dctim
New

Posts : 1760
Join date : 2009-05-15
Age : 106
Location : DC 'burbs

View user profile

Back to top Go down

Re: Yearlong County Contest Discussion Thread

Post  Sponsored content


Sponsored content


Back to top Go down

Page 1 of 2 1, 2  Next

View previous topic View next topic Back to top

- Similar topics

 
Permissions in this forum:
You cannot reply to topics in this forum