James, I am all for a level playing field. But this sure doesn't seem to represent that notion. It is so far out of normal for what one might expect given the circumstances that it has left us with no reasonable explanation of its condition. It is completely different from all other tracks AT have uploaded so I wondered if it was an effect of reaching a higher level, but the HBA upload 12 hours ago that posted 1 hour ago with 716 points and Level 23 would seem to indicate that's not the case. (it is level 13 though...) I think the HBA song is entirely deserving of its position, btw.

As I understand normalization (which is not much) it is an attempt to compare data from a wide range of sources and compare it to some normal behavior.

So maybe this song is just too...unusual?

I know you say the algorithm isn't punishing the song, and I'm sure it was not your intent to, however, in actual point of fact...it's hard for me to see how that is not exactly what is happening.

If it's an anomaly that has fallen through the cracks then perhaps the system should be examined to correct for that.

Guess what just posted. :D

haha, finally!

Well, thats good. I still don't understand it.

The exact same problem is now happening with DJ Kaspy - Rakuyou (Fallen Leaves). It is a classical tune uploaded 4 days ago with now 643 bumps and 153 unique bumpers and it is still not posted. James and Sam, if you think this system is working properly, please show a little bit of transparency by clearly explaining whats going on so that both listeners and artists can change their behaviours accordingly. Thanks!

DJ Kaspy might be a good example of the same people bumping the same artist. So while it might seem diverse, if you compare to the bumpers on the other songs, it may just be the same fan base pumping up their bro's song.

I think it's often better for a song to wait to post. It often gets more recognition once it does post. New fans might pay attention to a new post with more bumps.

I wish that were true Akathisia, but it just isn't bearing out to be the case. The AT song in question above posted, but never made it to the homepage. It has over 1000 points so it is very well received by those who hear it, but it did not get anywhere near the exposure that it deserved, and more importantly that others of it's level have been getting. Now a second AT song is sitting at 623 (most of which was gained in the first day or so), unposted after 13 days.

These songs ARE being punished because their listeners are enthusiastic. There is a larger number than usual of maxbumpers. Perhaps that is considered bad? It doesn't seem like it should to me, but...

The formula has decided they are not "normal". Therefore it puts on the brakes.

This seems almost Orwellian to me. Computers trying to squelch anything out of the ordinary to maintain a homogenous society. Is it time to disassemble HAL?

I agree with carpex. If you have no intention of changing this so it doesn't adversely affect song placement, at least tell us what it is that YOU think is wrong with the way it got it's points. Vagueness does not help the situation it just prolongs the confusion and dissatisfaction.

@utterlylinda--What you are talking about happens with tons of other artists--especially the more established acts on t61.

Looking at your bump history it may just be that you aren't as familiar with some of the posting patterns of other acts because you tend to follow a limited number of bands very closely. I'm wondering if the algorithm looks at all the AT songs and notices that the same 200-250 people bump the songs over and over? Diversity means new people bumping too not just the same fans.

That aside, it's hard to figure out what you mean by exposure. I remember that AT song getting to the front page. It happened in a very short period of time. It is also the most-bumped song ever by AT. It did not stay long on the front page because no new people bumped it. Usually a song only stays on the front page if new people are adding bumps to the song. Maybe the only people who bumped it are the only ones who are interested in AT and they bumped early?

Another bump-tactic you might consider is saving some bumps to keep a song on the front page once it posts. I do that with a lot of songs. I feel like it's my contribution to getting a band more exposure. When you MB early it pumps up the points, but late MB's can move a song up the front page. It's fun. Try it some time.

I am a lowly level 14 with only 260 bands I've bumped (and a couple that I am happy to say I've maxbumped many songs for) but I sure haven't seen midlevel 600+ bumped songs go unposted without it raising some attention and questions about why not. Is 260 really a "limited number"?

As for You Hung The Moon, I can't say with absolute certainty that it never hit the homepage as we have lost the ability to see that information, but I never saw it there and I was looking for it actively. The highest I heard about was #18 on the hot list. That was 15 minutes after it posted. How can the formula demand that the song have new listeners to get/stay on the homepage if it gives it 15 minutes or less of exposure there? And it did drop extremely fast despite having comparable numbers to an HBA song that posted about the same time, which miraculously went to #1 on the hot page and stuck there. There is NO visibility on stats so it's really hard to make comparisons, but I scribbled some notes at the time because I thought it was so odd. I noted the HBA song at 931 points an hour after posting. Can't recall exactly where, but I think it was about midway on the homepage. After 3 hours it had gained 26 points and moved up to #1 on the homepage. The AT song (which I thought the highest was #18) one hour after posting had dropped to #38 with 905 points after gaining 24 points. At 3 hours it had gained 62 and was at #40.

That's what I mean by less exposure. Minutes on the homepage translate into listeners and points. You get new bumpers and points by that exposure. I reiterate: Catch 22. You can get new listeners by getting to the homepage but you can't get to the homepage because you don't have new listeners.

Again I can't recall the exact timing on posting but I believe it was about 8 days - do you consider that a very short period of time for an 800+ bumped song to linger in limbo?

As for the diversity amongst AT bumpers...
Mathman did the number crunching on that one(thank you)a day before the song posted, and I quote:
"as of the time i did my comparison the song "half a world away" which posted 30 days ago had 888 bumps from 174 individual bumpers. the song "you hung the moon" had 843 bumps from 213 individual bumpers.

looking at the first 174 individual bumpers the two songs had 83 bumpers in common and 91 bumpers that were different. this means that at that point the two songs had a crossover rate of 47.8%. i'm not sure what the founders need for diversity but to me i think you've achieved diversity if over half of the bumpers of one song are different from those of another. taking a longer view, at the 213 individual bumpers point 99 of the bumpers on "you hung the moon" are the same as the bumpers on "half a world away." this means that 114 of the bumpers on the new song did not bump the last one for a crossover rate of 46.5% at this point."

Over half new bumpers. Does that REALLY defy the diversity unwritten rule?

Oh, and thanks for the lesson in bump restraint. I can practically feel the avuncular pat on the head. I have practiced that method myself for quite some time. I generally save at least half of my bumps for use in keeping a song on the homepage. Unfortunately it hasn't really been effective for these songs that seem to be playing under different rules than most others. I did not maxbump You Hung The Moon early. I got 296 points from a first bump on a 1000+ song (saved some points for the homepage that I never saw and the later points didn't help move it up). So, I went ahead and maxbumped Down somewhere around 5 or 6 hundred in hopes of getting it to post and be heard by more people. My bumps early don't help, saving them for later doesn't help. Apparently it doesn't work the same for these songs as it does for the ones you've been helping out with that strategy. Which is the point of this whole topic.

Hi utterlylinda--really I'm just throwing out some ideas. I don't know shit about it. Sounds like you have done a lot of homework. It seems like you took what I said a bit differently than I intended. That happens with print versus face to face.

P.S. I will be busy looking up avuncular--but I don't expect that it's going to be complimentary.

avuncular |əˈvə ng kyələr|
adjective
1 of or relating to an uncle.
• kind and friendly toward a younger or less experienced person : an avuncular manner.
2 Anthropology of or relating to the relationship between men and their siblings' children.
ORIGIN mid 19th cent.: from Latin avunculus ‘maternal uncle,’ diminutive of avus ‘grandfather.’

haha, Akathisia, don't worry. It was quite complimentary, although it does amuse me a bit to be considered the youngster in the exchange given that I am somewhat ...gifted.. in the accumulation of years.

I appreciate any ideas that might help me to understand why this continues, but the more I check out possibilities, the less it seems like it is business as usual. If I knew just what the non-normal criterion was, and what effect the formula imposes it might help.

OK someone explain this one to me.. Yesterday the Suspects put a new song.. "Sizzle" (i got FB on it), then about 10-15 min later put up "Lies are Free". Lies are free just posted a few minutes ago and here are the current stats on both songs... Lies has 261 bumps from 106 bumpers. Sizzle has 303 Bumps from 108 bumpers. Both songs belong to the Ska genre. Im too lazy to do the research but im sure that most of the bumpers are the same people considering its the same band/genre.

I dont really mind which posted first, im sure Sizzle will get its day in the sun, just thought it was odd. Any ideas??

@Liquidtrailz
Tha answers to the questions you seek are written in a sacred book, hidden high in the himalayas guarded by blind monks whose only purpose in life is to protect the book (and contemplate their navels...but thats a different story).

Many have tried to find it some even dabbling in the black arts ... all have perished and failed doomed to forever live in Level 10 or below.

It is said that one day a person of pure heart will read it and on that day will the fate of the universe be decided.

....or you could email James and ask.

"...level 10 or below." lol.

Level 19>Level 18>Level 17>...

I personally have no problem with the principles James described in an earlier thread.

"...draw from a wide array of behavioral data/patterns to understand the song's universal appeal in the overall community. this is simply a measure to encourage a level playing field between more established acts on thesixtyone (or popular groups coming into it) and newer ones"

The problem I have is with implementation. Take DJKaspy's Rakuyou for example. If we make the ridiculous assumption that half of the 810 bumps the song received are from a small number of groupie listeners who max bumped the song. If the algorithm flags this as suspicious activity, it should reduce the effect of those listeners. If we make the other ridiculous assumption that the algorithm should completely remove the effect of their bumps, that leaves over 400 bumps on a classical tune, which should still be enough for that song to be posted.

What the algorithm seems to do is to flag the entire song based on some metadata analysis, thereby ignoring all the bumpers, including those who have never heard of this artist before. So this has the effect of penalizing both the song and the artist, which IMHO is not right.

"What the algorithm seems to do is to flag the entire song based on some metadata analysis, thereby ignoring all the bumpers, including those who have never heard of this artist before. So this has the effect of penalizing both the song and the artist, which IMHO is not right."

Even if this were NOT the case, and only the 'frequent' bumpers are ignored, the other bumpers are still penalized, because they might make their point investment based on the perceived popularity of the song. And if the impact of even SOME of a song's bumps is diminished, then fair market value is not gotten by any investor.

One could argue that a bump should be about one's opinion of the song, and not about point speculation. But if the game is there, it should be expected that people will play it. And if I can't know which bump is really worth a bump (because of the SYSTEM, not the SONG), then I'm left feeling pretty gun-shy about spending points on anything.

james i know you're not in the habit of giving out detailed information about the inner workings of the system but i would kind of like to be thrown some sort of bone. something slightly more specific than "there is data involved," preferably. all i'm looking for is a little clarification so that discussion of the subject is marginally less speculative.

seriously the conjecture is thick enough to kill a man

@jimt
if some of the rules are a mystery and/or seem unfair or biased it seems to me one has 3 options:
1. opt out of the game.
2. play to please yourself without restraint.
3. allow your intelligence to be guided by your experience while you play to please yourself.

option 3 is the one i go by.

@johnasdf
in some ways the easiest way for the founders to get rid of the complaints about the various mysteries of how the game works would just be to print the formulas that are used to generate point returns to listeners and artists as well as the formulas that determine when and whether a song posts to the home page. i strongly suspect that these formulas would require a knowledge of higher-level statistics and calculus to use to make informed decisions about bumping. of course, then there would probably be complaints about how complicated they all were =0

Sam and James walk a thin line. They are running a business here. They need to communicate with us so that we know how the site works, and we feel good about the changes they make, etc, but on the other hand, they have a product to protect. If they give us all the rules and formulas that their system uses, what's to stop some clever users from modeling a competing site off the 61? I've sometimes felt like we haven't had enough communication, but in this case, I think James is trying to respond to our questions without compromising the site's security.

@ganglyblueheron
i hope you understand that i was being somewhat ironic with the second half of my post above.

btw, i put a reply to your 7/24 post to my wall above your post. it's currently on page 7 of my wall.