Year-end point systems

[j]

I don't have a year-end point system because I tried it in 1997 for my Top 20 (#20 = 1 pt, #1 = 20 pts, multiplied by # of weeks spent on the chart) and it simply wasn't reflective of my tastes for the year. Also, it took me forever to finish, and I have better things to do.

[superbu]

Matt4319 said:

This brings up a good point - even with the system I proposed, four weeks at #50 would roughly equal a week at #1.

The inverse exponential thing is pretty wild though... Even for a top 10, #1 would get 512 points compared to 1 for #10.

Right, which is why I enjoyed doing the 1940s charts much more, since a top ten was all there was from 1940-47! :)

The thing about my system is, it doesn't matter how many positions you're working with -- whether it's only the top ten, the top fifty, or the top 100, whatever. The RATIO between songs stays the same, no matter what. That's not true of straight reverse points.

[insect2]

I think if you guys want the perfect formula to calculate a year-end chart, you should just decide the chart yourselves.

On my chart, two weeks at #21 would be equal to one week at #1. But that doesn't mean that a song that peaks for two weeks at #21 is doing as well as one with a week at #1. That's just an illogical argument. Maybe if it spends 15 weeks at #21, and if it does, then it probably deserves to do as well as a song with only 1 week at #1. There are many factors to consider when you tally points in and then compare two songs' chart performances. It doesn't mean #1 has to be worth 1,048,576 times more than #21. I personally see no fault with the inverse points system, except I add on points for weeks on and peak position, but the system ain't broke, so I'm not fixing it. :)

[mst3k]

I agree with insect...

The whole "2 weeks at #50 equals 1 week at #1" is generally a moot point anyway, since that song that gets to #1 will outscore the #50 peaking song by a huge margin when all of its weekly positions are totalled up (unless song "A" happens to debut at #1 and fall entirely off the very next week, in which case it doesn't deserve to rank that highly on the yearend chart anyway).

I've used inverse points since 1989, and my yearend charts have reflected the weekly lists quite accurately. :)

[Hervard]

geoguy said:

Next year.

AP Stats?? Does that stand for Advanced Probability and Stats? Hell, I couldn't even hack Basic Prob and Stats! The advanced course would probably make me pull all the hair out of my head!

[Matt4319]

ulij20 said:

AP Stats?? Does that stand for Advanced Probability and Stats? Hell, I couldn't even hack Basic Prob and Stats! The advanced course would probably make me pull all the hair out of my head!

Pretty much, although the actual name is Advanced Placement Statistics. There are many types of Advanced Placement classes, so they call it AP for short.

[Hervard]

mst3k said:

I agree with insect...

The whole "2 weeks at #50 equals 1 week at #1" is generally a moot point anyway, since that song that gets to #1 will outscore the #50 peaking song by a huge margin when all of its weekly positions are totalled up (unless song "A" happens to debut at #1 and fall entirely off the very next week, in which case it doesn't deserve to rank that highly on the yearend chart anyway).

I've used inverse points since 1989, and my yearend charts have reflected the weekly lists quite accurately.  :)

Yeah, IMO, the inverse point system works out the best. First of all, it's the easiest one to use (well, the easiest one that's accurate) and it's not miserably frustrating, as many of the ones I've seen look (hey, I already said that I totally failed Basic Prob & Stats).
Many people who use inverse points also give extra points to songs that reach a certain peak position. I do this only for the #1 song. Ten extra points, to be exact, which is what R&R did for most of their year end charts until the PPW era was ushered in.

[Deleted]

I do inverse points... nothing extra.  It works great for me.  It's natural to not agree completely with your year-end chart.  You're going to have burn out on some songs and some newer songs may sound really great and fresh when they debut in November and you put it at #1 year-end and it wears off on you a few months later.

[strong4PMB!]

Matt4319 said:

Pretty much, although the actual name is Advanced Placement Statistics.  There are many types of Advanced Placement classes, so they call it AP for short.

Yes.

My school only offers two classes for people in my math rank: AP Stats or AP Calculus. So I'm taking AP Stats because it'll be more useful for my business major.

[superbu]

mst3k said:

I agree with insect...

The whole "2 weeks at #50 equals 1 week at #1" is generally a moot point anyway, since that song that gets to #1 will outscore the #50 peaking song by a huge margin when all of its weekly positions are totalled up

Well, that's only one example of why I don't use straight reverse points anymore.

Another is that the difference percentagewise between each chart position decreases the farther up the chart you move... and the larger the chart, the smaller the difference between positions as you move up.

Example: Say you're only using each song's run in the top ten to rank your chart. In that case, a #9 position will be worth double the points of a #10 position, a #8 fifty percent of the points of #9, decreasing on up until a #1 only has around 11% more points than a #2.

All fine and dandy, until you start expanding your chart...

Because if you're using the full Top 100 for each song, a position #99 is still worth double the points of a #100... but a #1 is only worth about ONE PERCENT more than a #2. Using the Top 50, it'd be around 2%. See what I'm saying? The point difference between each position decreases as you move up the chart, and the larger your chart, the more it decreases.

I know straight reverse points is a perfectly valid method that still gets very good results, due to bigger hits just naturally spending more time higher up on the charts. But the idea of a #2 being worth 99% of what a #1 is worth just doesn't seem quite right to me.

I like the idea of a consistent percentage increase between chart positions, whether it's doubling points, as I like to use, or, say, 10% added to each position up the chart (which might be more reasonable), or whatever.

But you'll actually get fairly similar results with reverse points.

[chebingeo]

I use the inverse point system with bonus points for weeks spent in the top 10 and weeks spent at #1

[Bob]

chebingeo said:

I use the inverse point system with bonus points for weeks spent in the top 10 and weeks spent at #1

I think that's a good idea

[Devil Marlena Nylund]

Inverse points work but on my chart, a lot of times a song will hang around for a long time even though it didn't peak low and end up having a huge yearend placement, like 'Losing Grip' by Avril Lavigne. It spent 46 weeks on the chart within 2003 and peaked at #16 early in the year but hung around the #50 mark for pretty much the rest of it and managed at #8 yearend placement. I thought that was too high considering a lot of my #1s of the year, which were also huge hits, were lower. And songs that have slow rises but really quick falls place too high too. I usually make the yearend chart with inverse and then add in bonus points.