Re^4: A Level Playing Field

[demerphq]

I'd like to address a few of your points, and not necessarily in order.

First is this issue of votes bonus. You seem to think that people will lose out because of the vote bonus. This wrong. Under the new system users will continue to recieve their vote bonus until they have 250XP. Under the old system the user would lose their vote bonus at 200XP. So that has actually moved up.

Regarding the new vote allocation scheme. You say that people lose votes. And you are right. That was a deliberate design decision. Under the old scheme all users over 3k XP recieved daily 40 votes, which represented a voting pool of 16.6k votes. Those same 400 odd users now receive about 7k votes. Generally speaking this is perceived as a Good Thing as it makes newer and lower level users votes mean proportionally more, and compensates for the increase in total number of users. Overall the vote count is still almost certainly signifigantly higher than it was when the original level scheme was created.

Also, it seems to me that you (and many others) have become used to the old system without realizing that it in fact was a very different system when users like myself first joined. Not because the system changed of course, but because of the size of the user base. When I started there were less than 20 Saints and only about 100-200 regular users. We currently have a regular user base of about 4000 and we have over 400 saints. This is a drammatic inflation. And the way we have compensated for it is to reduce the number of votes available. This will mean that the $NORM will drop over time, which in turn will mean that lower rep'ed nodes will generate more XP for the users that write them.

Id really appreciate it if you could leave off the doom-and-gloom predictions until the new system has had a chance to "settle down". I think that in two or three weeks you will have forgotten what all the fuss is.

And just to reassure you, the gods involved in this releveling did a fair amount of analysis and debate to come up with the new scheme. We considered a lot of factors, including current user XP distributions. We think that the new scheme will in the long term provide a lot of room for growth, and will return the voting/experience system to something a little closer to what it was when it was first introduced. And frankly if it looks like it needs to be tweaked again to respond to something we have overlooked, then it will be.

So rest easy.

Oh, yeah, about this saint issue. If you are level 13 or higher then you are a saint (or perhaps 'saintly'). If you are level 26 then you are both a Saint and saintly. IOW, we still consider all users over 3k to be 'saints'. Have a look at Saints In Our Book to see.

[Perl Mouse]

When I started there were less than 20 Saints and only about 100-200 regular users. We currently have a regular user base of about 4000 and we have over 400 saints. This is a drammatic inflation.

No doubt there's inflation (I only recently arrived here - I only know the current situation), but the numbers you quote don't show "drammatic inflation". In fact, they show *de*flation. "Less than 20 Saints and about 100-200 regular users", suggest the ratio of users/saints is between "1:10" to "almost 1:5". Currently having 400 saints and about 4000 users gives a ratio of about "1:10". Which means that if the number of regular users in old times is in the high end of your estimate, we have an about same ratio of saints, perhaps slightly more. But if the number of users was at the low end of your estimate, the relative number of saints has almost halved.

Note that I'm not commenting on the real situation several years ago - I wasn't there. I'm just commenting the numbers you state, and the conclusion you draw from those numbers.

[demerphq]

Sorry, i should have been more clear. Its a drammatic inflation of the votes available in the system. The point is that the amount of votes in common circulation is much higher now (even with the new vote allocations), as a relative number, as a ratio of users regularly online, and as a ratio of posts being made.

For instance the saintly ones are so because they come here often, if you took a look at the average days stats a much higher proportion of saints would be online on a given day than any other level. Back when there were much less saints this was much less relevent, in the recent past it was ridiculous.

With something like an average of 20-50 posts a day we had 16k votes in ready circulation. Thats something like 800 votes a node if fully utilized. The only reason $NORM didnt go completely crazy over the recent past was because many of the saints were aware of this and tended towards minimizing their votes.

So the inflation I mean is the amount of currency available to the community at large. We reduced the amount of votes allocated in order to try to counteract this tendency. Hopefully it will work out ok.

[sauoq]

With something like an average of 20-50 posts a day we had 16k votes in ready circulation. Thats something like 800 votes a node if fully utilized.

Are you really saying that PM only gets 20-50 posts a day? I've never tried counting, but that doesn't jive with my experience...

And, uh, we've got over 500,000 nodes. I've been here for about 3 years and my home node is around 180,000. So, that's over 100,000 nodes per year. Without considering the rate of change, that would make the average per day somewhere around 275. With 16k votes, that works out to about 60 votes per node, a far cry from 800.

If I'm right and the above is with 4000 regular users as you said and there used to be only about 200 regular users and if the post ratio remained about the same... then we could extrapolate that there used to be an average of about 13-14 new nodes per day. With about 20 saints, that would work out to about... 60 votes per node. I.e. there has probably been very little net change in votes per node since you first joined.

One thing that probably has changed is the percentage of new nodes that a saint has been able to vote on. Previously, it seems a saint could have voted on every new node in a given day and still have votes left over. Now, if my calculations were correct, a saint couldn't quite vote on 20% of the new nodes in a given day. (This fact supports an abstention option, I think.)

With the sudden cut in votes, I'll be interested to see how the rep of new nodes will compare with the rep of old nodes... I believe we're going to see a drastic drop in node rep. I don't necessarily see that as a good thing because range compression will result in less distinction between great nodes and merely good ones. Or good ones and average ones. Etc.

So, I've got reservations about the changes to available votes. I'm definitely interested in knowing what metrics you'll be collecting and how you'll be analyzing them to determine the success of these changes.

As for the additional levels, I think it's great! I've got something to shoot for again. Once you are a saint 5 times over, what's the point? Well, the truth is that I've always come to try to help others anyway. That's where I get my real satisfaction. But still, I like the "game" too, and this brings new life to it.

[blazar]

Sorry, i should have been more clear. Its a drammatic inflation of the votes available in the system. The point is that the amount of votes in common circulation is much higher now (even with the new vote allocations), as a relative number, as a ratio of users regularly online, and as a ratio of posts being made.

<snip>

So the inflation I mean is the amount of currency available to the community at large. We reduced the amount of votes allocated in order to try to counteract this tendency. Hopefully it will work out ok.

Which makes me wonder, letting aside for the monent any possible inexactness of your claims (which somebody contended), of an alternative possibility. More precisely I read some replies to the OP suggesting an adaptive scheme for levels based on a sort of normalization on the highest ranking(s) at each given time. And an answer was given, to the effect that it may not be such a good idea. I don't think it is, either, although I wouldn't exclude a priori that a less naive and more refined adaptive scheme may be a good one, instead.

However if the actual problem that the reform is aimed at curing is the "drammatic inflation of the votes available in the system", then chances are that any fixed scheme may turn out to be in the long run not perfectly tuned to the status of the Monastry at each given time. So, maybe, an adaptive scheme for the number of daily votes based on the collection of statistical data of the system may be thought of, as an option.(And, if it were, then should also include an "inertia term" to avoid abrupt changes.) Just my two (Euro)cents...