|
You're sorted now. 
|
|
|
|
|
But it still doesn't explain why I was top before; I'm not asking for extra points, rather that the others get due recognition. But thanks anyway.
The best things in life are not things.
|
|
|
|
|
Still isn't right
SA has 4x5 votes now and is 3th in line.
First I thought it was because of total rep points or so but SA still beats you there.
So I'd think something is wrong with the sorting.
|
|
|
|
|
It looks like it has something to do with the final value. Regardless of how many votes we each received, our totals are all 5.00/5.
The best things in life are not things.
|
|
|
|
|
You're just special on this one, I guess... 
I wasn't, now I am, then I won't be anymore.
|
|
|
|
|
 I wish.
The best things in life are not things.
|
|
|
|
|
More than a dozen members, many of whom have names like "Member 8027406", joined the site in the last 3 days and are giving my article a 1/5. They're evidently modeling themselves after "Member 3821620" who joined 4 years ago. So um, can anything be done about that? Or is it impossible to rule out that these are just a bunch of separate individuals that joined the site for the sole purpose of voting down my article?
Head-to-head benchmark: C++ vs .NET[^]
|
|
|
|
|
This is an impressive article. You have obviously put a lot of effort into it, and you present your results in a very clear, easy-to-read manner. My suggestion for improving it: take the negative comments, embrace them, and update your article, either by fixing the code or by explaining the point in question. I have countered the 1-votes on a few of the posts, and the rating went from 1 to 4.5+, meaning that the voter was a newbie who has yet to contribute anything to the site. Or as Christian put it, moron.
That was the "good news" part of my response. Now for the bad news: The voting system here on CodeProject is hopelessly broken. You will get no satisfaction from that statement, and other responses here will probably be along the lines of "over time, the 1-votes won't matter". This is clearly bullcrap, but don't expect anything else. It is what it is, get used to it.
Keep up the good work. Don't let the morons get you down.
|
|
|
|
|
Damn Hans! That was spot on.
-----------------------------
Just along for the ride.
-----------------------------
|
|
|
|
|
Hans Dietrich wrote: This is clearly bullcrap
Word up. Yeah - your rep points don't take much of a hit, but a single hostile 1-vote - REGARDLESS of the weight it carries - can take many many 5 votes to bring the rating back to an acceptable level. "Over time", yeah - it might, but once an article falls off the list of randomly selected new or "highest rated" lists, your vote count RARELY changes, thus prolonging the crap rating you acquired when one retard voted you a 1.
".45 ACP - because shooting twice is just silly" - JSOP, 2010
-----
You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
-----
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass." - Dale Earnhardt, 1997
|
|
|
|
|
Yep, happens to me a lot too... but I think suggestions for improving the voting system haven't been taken seriously. What can you do but move on. Keep contributing though, those of us that aren't a**holes appreciate the good articles and other contributions. I started out just browsing the site for a long time before I decided to start contributing.
Edit: Looks like a great article, I'll read it soon and vote accordingly. This topic is of great interest to me so I'm definitely excited to read! 
|
|
|
|
|
This looks like a gang of schoolchildren, or one person with multiple sock puppet accounts. With luck Chris and the team can find out and fix it.
The best things in life are not things.
|
|
|
|
|
The names take the format "Member nnnnnnn" as these are the CP auto-generated values until the user gives themselves a name. They may or may not be related to "Member 3821620".
As for the "stuffing" you are pretty much correct, these are likely to be sockpuppet accounts for a user that doesn't like your conclusions. I hope Chris can sort this out for you, downvoting has a much bigger effect on articles than elsewhere, the drop out of view as the rating goes down.
I'd like to second the commentthat the voting system isn't perfect, by allowing everyone to vote on articles we get some skewed results in both directions. I've seen pretty poorly written articles heavily bolstered by votes from friends of the author as well as what appears to be tactical downvoting (as opposed to malicious downvoting as in your case).
|
|
|
|
|
Keith Barrow wrote: wing everyone to vote on articles we get some skewed results in both directions
Exactly.
I can turn off downvoting and that solves the problem of members getting a rating of 4.85 / 5 instead of 5/5. It doesn't solve the problem, in the least, of sock-puppets upvoting crap articles with no way for members to balance this out.
In this particular case there are maybe 2-3 votes that look suspicious from the quick look I had. I will look deeper, though, and see if I can get creative.
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
Would it be possible to prevent those who have been members for less than (for example) 6 months, and reputations below (insert sensible number here) from voting at all? In this case (and maybe others) it's quite possible that people are put off reading the article because of all the 1 votes.
The best things in life are not things.
|
|
|
|
|
Richard MacCutchan wrote: t's quite possible that people are put off reading the article because of all the 1 votes.
At the top of the article is a big "4 stars" and a 4.12 (110 votes). To me that says "Good article" and I can't imagine anyone being put off reading it because of that. If you hover to get the histogram then yes, you see the one's, but I would imagine the average person would treat these as noise or spurious.
I'm against [Edit] not [/Edit] letting new members vote - that robs the community of fresh voices. However, we do have in place a requirement that your email is a confirmed email address, thus making it harder to setup casual sock puppets.
I'm working on something today that may help a little.
[Edit] OK, so I have the 'flu and need sleep 
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
modified on Thursday, June 23, 2011 3:14 PM
|
|
|
|
|
Chris Maunder wrote: I'm against letting new members vote
 I do not think that is what you meant.
|
|
|
|
|
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
-----------------------------
Just along for the ride.
-----------------------------
|
|
|
|
|
So most of the votes really are people joining the site so they can give me a 1. Argh. I got suspicious when I saw all those votes with numbers instead of names... but then again it's odd for sock puppet accounts to be so obvious by using a naming pattern.
Anyway, if people are too lazy to give themselves a name, how much weight should their vote carry? The comment that goes with the 1-vote is supposed to be for constructive criticism but most of the 1-vote comments are somewhere between unhelpful and insulting. I might feel better to have a 1-vote without a comment, at least that way the 1-votes wouldn't drown out the more productive discussions that are going on. I took the 1-votes personally at first, but I'm getting over it now.
I'm curious, though, how lots of angry non-members found out about the article so soon after it was posted. I Googled the URL and didn't see any negative reviews linking to the page. Maybe just word-of-mouth then...
|
|
|
|
|
Qwertie wrote: So most of the votes really are people joining the site so they can give me a 1.
No, not most. Just a few.
Qwertie wrote: most of the 1-vote comments are somewhere between unhelpful and insulting
Click the "vote to remove message" link and if enough people vote the message off the island, the vote goes with it.
Qwertie wrote: lots of angry non-members found out about the article so soon after it was posted
You have a jealous rival, maybe? 
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
I know you are against it (from further down the thread), but I think limiting the voting powers (say by a mixture of author and authority points) might be an option for articles, given the much heavier effect that this has than on, say, a forum message. Even by setting the bar pretty low on these categories it might make setting up multiple sockpuppets more trouble than it is worth (i.e. the user would actually have to help in the technical areas before they could use them to vote on an article).
That said, I don't think the voting system is fundamentally broken, it does provide a reasonable filter in practise.
|
|
|
|
|
Take another look at your rating. Hover over the stars to get the histogram and the explanation.
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
I like it, that's a cracking implementation. Kudos to the coders there. When we meet, I will crack open a bottle or three of a decent single malt with you.
|
|
|
|
|
Cool, you added this feature just now? Sweet!
The only thing is, it sounds like an all-or-nothing system. Either the "1" votes are within 2 standard deviations (and always included) or outside 2 standard deviations (and always excluded), which doesn't seem fair as it discounts the haters' opinion entirely (or weights it 100%). What if, instead of completely excluding the "1" votes, you just moved them closer to the average. Taking my article, for instance, since most people voted "5", maybe the "1" votes could be treated as "3" or "4" instead. Alternately, the weight of the "1"s could be reduced.
Another thing that might help to discourage sock puppets angry mobs slightly is to not allow voting from people that joined in the last 24 hours. [Edit: I had written a second idea here but on second thought, it was dumb]
Addendum: Or, since the negative votes are skewed toward new members and nonparticipators, give more weight to longtime members or people with high reputation. If CodeProject reputation worked like StackOverflow, something like the logarithm of reputation would work nicely. But CodeProject reputations can decrease over time, and even be negative, so I'm not sure. Anyway, this kind of approach would also have some effect against sock puppets that upvote articles. It doesn't necessarily distinguish sock puppets from mere lurkers, but I'm not sure if that's possible anyway. You could also mix duration of membership in the calculation in an attempt to discourage sock-puppets (since sock puppets are usually created very close to the time they are used) but then it would discriminate against newbies... well, I see it's a hard problem.
One more silly idea against sock puppets: create a histogram of join dates + votes and only allow one vote per combination of score and date. For example, if 3 people joined Dec. 1, 2010 and all give a score of "5" to an article, only count it as 1 vote. This would mean that sock puppets are less effective if the creator made them all on the same day . Or alternately, have a script email an admin when lots of people join on the same day to vote the same way on the same article, and the admin can investigate whether it's legitimate or not. Whew. Why I am still up at 3AM, I cannot say.
Just some random thoughts. But I really appreciate that I seem to have been the motivation for this feature 
modified on Saturday, June 25, 2011 4:59 AM
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
