Polls

[Rob Monster]

Just wanted to support you in this post @Ategy.com. I voted a number if days ago in the poll, but to me it seemed immediately obvious that the voting was rigged for VettedName. It had 30% of the votes but no mentions (at the time) in the thread as to why anyone was picking it. I also saw it as a relatively poor name and for the vote to be at such an overwhelming majority was highly irregular. I was going to mention at the time that someone should check how many new NP accounts were voting for it, but thought better of it and deleted the comments before I posted them. So I just stopped watching the thread and decided to ignore it since I was pretty sure others would figure it out too. So I've read nothing between page 2 and page 9. I just decided to jump in again as the thread was in the 'popular this week' box. I read this last page and from your post I can pretty well guess what the pages of posts have been about so I'm quite happy I've been ignoring the thread till now. Anyway, your post seems very diplomatic and well thought out so just wanted to say so. I shall now go and hide back under my rock...

I do think @Paul Buonopane will want to look at the problem of "Ballot stuffing" and "Brigading". The technique that I think is interesting is to use a service like Maxmind that measures the integrity of a user. In the case of Maxmind score, the risk goes from 0 to 100, where 0 is riskless and 100 is almost certainly a fraud. We are using this for TrustRatings as a way to adjust ballot-stuffing using bots or proxies. I won't comment on any particular group because last time I did, the moderators considered it off-topic and removed my post. My goodness.

 

[Paul]

We can simply click on some the names and look at post count. I didn't say the "overwhelming majority" I stated a fact, that some joined just to vote. That's the truth. There are a lot of 0 and 1 post voters from Africa/Algiers. I just picked a random 5, the last 5, 3 of them are 0 posts.

select count(*) from xf_poll_vote v inner join xf_user u using (user_id) where v.poll_id = 4157 and u.register_date >= (select post_date from xf_thread where thread_id = 1155658);
+----------+
| count(*) |
+----------+
|        4 |
+----------+
1 row in set (0.01 sec)

A total of 4 voters registered after this thread was created.

select count(*) from xf_poll_vote v inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count = 0;
+----------+
| count(*) |
+----------+
|        9 |
+----------+
1 row in set (0.00 sec)

A total of 9 voters have no posts to their name.

 

[JB Lions]

select count(*) from xf_poll_vote v inner join xf_user u using (user_id) where v.poll_id = 4157 and u.register_date >= (select post_date from xf_thread where thread_id = 1155658);
+----------+
| count(*) |
+----------+
|        4 |
+----------+
1 row in set (0.01 sec)

A total of 4 voters registered after this thread was created.

select count(*) from xf_poll_vote v inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count = 0;
+----------+
| count(*) |
+----------+
|        9 |
+----------+
1 row in set (0.00 sec)

A total of 9 voters have no posts to their name.

You just posted 0 posts, I said 1 posts also. Many have low post counts. And what you posted 50/62 from Africa. The poll was gamed, obviously. They're literally talking about it on their message board. You can put checks in there to help cut that down, I didn't say eliminate but it would cut down on this type of stuff. Again, other forums do this. This isn't life and death stuff, I get that, just a poll on the internet.

 

[Paul]

You just posted 0 posts, I said 1 posts also. Many have low post counts. And what you posted 50/62 from Africa. The poll was gamed, obviously.

select count(*) from xf_poll_vote v inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count = 1;
+----------+
| count(*) |
+----------+
|       13 |
+----------+
1 row in set (0.00 sec)

Edit: To clarify, this is the number of voters with exactly one post. The number of voters with <= 1 post is 9 + 13 = 22.

 

[Paul]

We can simply click on some the names and look at post count. I didn't say the "overwhelming majority" I stated a fact, that some joined just to vote. That's the truth. There are a lot of 0 and 1 post voters from Africa/Algiers. I just picked a random 5, the last 5, 3 of them are 0 posts.

Again, only 4 voters registered after the poll started, so we can safely put an upper limit of 4 on the number of people who registered exclusively to vote. That's a very small number.

Please stop speculating about the integrity of this poll. There is simply no evidence at this time that the poll was manipulated. There's nothing wrong with people encouraging each other to vote, which is all I see happening here.

 

[JB Lions]

Again, only 4 voters registered after the poll started, so we can safely put an upper limit of 4 on the number of people who registered exclusively to vote. That's a very small number.

Please stop speculating about the integrity of this poll. There is simply no evidence at this time that the poll was manipulated. There's nothing wrong with people encouraging each other to vote, which is all I see happening here.

Just a forum thread with the person submitting the name encouraging everybody to go vote for the name, 50/62 voting for that name from Africa. Tell me what other name has that percentage? And again, I didn't say all, but yes, some. Some could have registered before and just had accounts here and just supporting somebody from the same country.

Contrast that with the current front-runner: VettedName.com, which was a late entrant. That was a name I bought after the poll started because a Nigerian domainer suggested it. I think he told his friends to vote for it, so the poll might have a little bias on that one. The competition is still wide open.

As for VettedName.com -- I am pretty sure that this is the flag name of Team Nigeria. They are on NamePros in force, and I am thankful for them. There is so much domaining talent in Nigeria that I think it is going to surprise a lot of people when suddenly they are losing to Nigerians in auctions because they have figured out how to not only buy good domains, but also sell them for compelling returns.

More here:

https://www.nairaland.com/3430037/what-make-money-online/788

 

[oldtimer]

At the bottom of the poll it says:

"Multiple votes are allowed"

Why not expand that and include some more guidelines, a lot of people say that they didn't know that they weren't supposed to vote for a name just to help their friends, it might cut down drastically on false results if rules are clearly indicated at the top and below the poll. Also as I and others had indicated before there should be some minimum requirements for a person to be able to vote such as minimum 3 months membership and minimum of 25 posts. IMO

 

[Paul]

Just a forum thread with the person submitting the name encouraging everybody to go vote for the name, 50/62 voting from that name from Africa. Tell me what other name has that percentage? And again, I didn't say all, but yes, some. Some could have registered before and just had accounts here.

We don't consider that manipulation because we have no way to prevent such things from happening. That is the nature of polls. Someone could just as easily do this in private, and we'd have no way of knowing.

If you're going to hold polls, you have to accept this as a possibility. It's ridiculous to suggest that we can prevent this.

At the bottom of the poll it says:

"Multiple votes are allowed"

Why not expand that and include some more guidelines, a lot of people say that they didn't know that they weren't supposed to vote for a name just to help their friends, it might cut down drastically on false results if rules are clearly indicated at the top and below the poll. Also as I and others had indicated before there should be some minimum requirements for a person to be able to vote such as minimum 3 months membership and minimum of 25 posts. IMO

For the sake of science, here are what the results would be if we were to do that:

Edit: Hold on a sec, updating this; did the query wrong.

Edit 2: Fixed.

Edit 3: Okay, really fixed this time.

select count(*), r.response from xf_poll_vote v inner join xf_poll_response r using (poll_response_id) inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count >= 25 and u.register_date <= (1569387806 - 90 * 24 * 60 * 60) group by v.poll_response_id order by count(*) desc;
+----------+--------------------+
| count(*) | response           |
+----------+--------------------+
|       49 | NameSeal.com       |
|       33 | DNprotect.com      |
|       32 | NameProtector.com  |
|       23 | VettedName.com     |
|       21 | Assurer.com        |
|       11 | NameAssured.com    |
|        9 | NameSafeguard.com  |
|        9 | NameAssurance.com  |
|        7 | NameSquare.com     |
|        6 | NameSurety.com     |
|        5 | NameLegit.com      |
|        5 | NameRoot.com       |
|        5 | NameFax.com        |
|        5 | NameCertify.com    |
|        4 | NomTrust.com       |
|        4 | NameNotary.com     |
|        3 | NameInspection.com |
|        3 | NameSoldier.com    |
|        2 | Trusturo.com       |
|        2 | CrucialNames.com   |
|        2 | NameValidate.com   |
|        1 | NameCertain.com    |
|        1 | NameCleared.com    |
+----------+--------------------+

 

[wwwweb]

I can very clearly see that the overwhelming majority of people did not register just to vote. They are established members.

They are regular established CONTRIBUTING members, or simply come online from time to time to perform a single task to be gone again?

 

[Paul]

They are regular established CONTRIBUTING members, or simply come online from time to time to perform a single task to be gone again?

@oldtimer suggested that we exclude people who have been registered for less than 3 months or who have less than 25 posts. I ran a query with those constraints; the results are above.

 

[JB Lions]

We don't consider that manipulation because we have no way to prevent such things from happening. That is the nature of polls. Someone could just as easily do this in private, and we'd have no way of knowing.

If you're going to hold polls, you have to accept this as a possibility. It's ridiculous to suggest that we can prevent this.

Again, I said there is no way to totally eliminate this and I've posted this many times now, and not once have you addressed this. Why do you think other forums, set join date/post count limits? I've done this myself. If you had a favorite domaining tool poll and the owner of 1 tool, got all his employees, friends and family to come vote for the poll vs. a poll where the voters had to have a min join date, 300 minimum post count. Which poll would be a more accurate picture?

 

[Paul]

Again, I said there is no way to totally eliminate this and I've posted this many times now, and not once have you addressed this. Why do you think other forums, set join date/post count limits? I've done this myself. If you had a favorite domaining tool poll and the owner of 1 tool, got all his employees, friends and family to come vote for the poll vs. a poll where the voters had to have a min join date, 300 minimum post count. Which poll would be a more accurate picture?

As I mentioned before, it depends what you're trying to measure.

Here's what we get if we limit it to voters with >= 300 posts who have been registered for >= 365 days:

select count(*), r.response from xf_poll_vote v inner join xf_poll_response r using (poll_response_id) inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count >= 300 and u.register_date <= (1569387806 - 365 * 24 * 60 * 60) group by v.poll_response_id order by count(*) desc;
+----------+--------------------+
| count(*) | response           |
+----------+--------------------+
|       38 | NameSeal.com       |
|       25 | DNprotect.com      |
|       20 | NameProtector.com  |
|       17 | Assurer.com        |
|        8 | NameAssured.com    |
|        7 | NameSafeguard.com  |
|        6 | NameSquare.com     |
|        6 | NameAssurance.com  |
|        5 | VettedName.com     |
|        5 | NameSurety.com     |
|        4 | NameCertify.com    |
|        3 | NameInspection.com |
|        3 | NameLegit.com      |
|        3 | NameRoot.com       |
|        3 | NameFax.com        |
|        3 | NameNotary.com     |
|        3 | NomTrust.com       |
|        2 | NameSoldier.com    |
|        2 | NameValidate.com   |
|        1 | Trusturo.com       |
|        1 | NameCertain.com    |
|        1 | CrucialNames.com   |
|        1 | NameCleared.com    |
+----------+--------------------+

 

[JB Lions]

As I mentioned before, it depends what you're trying to measure.

Here's what we get if we limit it to voters with >= 300 posts who have been registered for >= 365 days:

select count(*), r.response from xf_poll_vote v inner join xf_poll_response r using (poll_response_id) inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count >= 300 and u.register_date >= (1569387806 - 365 * 24 * 60 * 60) group by v.poll_response_id order by count(*) desc;
+----------+-------------------+
| count(*) | response          |
+----------+-------------------+
|        1 | Assurer.com       |
|        1 | NameSeal.com      |
|        1 | VettedName.com    |
|        1 | NameProtector.com |
|        1 | NameAssured.com   |
|        1 | NameSurety.com    |
+----------+-------------------+

Doesn't seem particularly useful.

That's not right. I just checked NameSeal.com and there are plenty of members that have been here longer than a year with over 300 posts. Not just 1.

 

[oldtimer]

Everyone here is putting their focus on fraud and intentional manipulation of the Poll, I believe that we should give people the benefit of the doubt and focus more on educating them and providing them with clear guidelines on what would be okay and what would be considered unacceptable by the majority of members here when voting. In my opinion the clear guidelines plus the minimum requirements to vote will drastically cut down on false results, why not try it and see.

 

[Paul]

That's not right. I just checked NameSeal.com and there are plenty of members that have been here longer than a year with over 300 posts. Not just 1.

Hmm... maybe I did something wrong. I'll try again.

 

[Paul]

That's not right. I just checked NameSeal.com and there are plenty of members that have been here longer than a year with over 300 posts. Not just 1.

Okay, I think I got the queries right this time.

 

[JB Lions]

As I mentioned before, it depends what you're trying to measure.

Here's what we get if we limit it to voters with >= 300 posts who have been registered for >= 365 days:

select count(*), r.response from xf_poll_vote v inner join xf_poll_response r using (poll_response_id) inner join xf_user u using (user_id) where v.poll_id = 4157 and u.message_count >= 300 and u.register_date <= (1569387806 - 365 * 24 * 60 * 60) group by v.poll_response_id order by count(*) desc;
+----------+--------------------+
| count(*) | response           |
+----------+--------------------+
|       38 | NameSeal.com       |
|       25 | DNprotect.com      |
|       20 | NameProtector.com  |
|       17 | Assurer.com        |
|        8 | NameAssured.com    |
|        7 | NameSafeguard.com  |
|        6 | NameSquare.com     |
|        6 | NameAssurance.com  |
|        5 | VettedName.com     |
|        5 | NameSurety.com     |
|        4 | NameCertify.com    |
|        3 | NameInspection.com |
|        3 | NameLegit.com      |
|        3 | NameRoot.com       |
|        3 | NameFax.com        |
|        3 | NameNotary.com     |
|        3 | NomTrust.com       |
|        2 | NameSoldier.com    |
|        2 | NameValidate.com   |
|        1 | Trusturo.com       |
|        1 | NameCertain.com    |
|        1 | CrucialNames.com   |
|        1 | NameCleared.com    |
+----------+--------------------+

Big difference.

NameSeal.com - 38
DNprotect.com - 25
NameProtector.com - 20
Assurer.com - 17
.......
VettedName.com - 5

This one looks better to me. The current poll, VettedName.com is winning. Put some minimum checks in place and it only has 5 votes.

 

[oldtimer]

Having too strict of requirements will deprive new domainers from voting who maybe just weren't aware of what is okay and what is not when voting, so I still believe that a modest minimum requirements PLUS adequate guidelines is the way to go so we won't leave good people who didn't know any better out (you can always raise the minimum requirements later if necessary). IMO

 

[Paul]

Having too strict of requirements will deprive new domainers from voting who maybe just weren't aware of what is okay and what is not when voting, so I still believe that a modest minimum requirements PLUS adequate guidelines is the way to go so we won't leave good people who didn't know any better out. IMO

Can we discuss this elsewhere? I've provided a lot of aggregate information on the polls; that's all I can do. If you want to discuss making changes to how polls work in general, that's off-topic for this thread.

Update by moderator: These posts have been moved to their own thread.

 

[oldtimer]

Can we discuss this elsewhere? I've provided a lot of aggregate information on the polls; that's all I can do. If you want to discuss making changes to how polls work in general, that's off-topic for this thread.

I believe everything that needs to be said is already said, it's just up to you to see if you want to try them or not, but thanks for your attention and input anyway.

 

[Abdillahi Abdillahi]

Just for the sake of argument , to @JB Lions and anyone else who believe there should be strict requirements on people voting, how can you make sure other members will not be doing exactly this and get support from members who they are good friends with and bypass such strict requirements?

You might have not liked the support the owner of that domain got from his people or anyone else, but there is no way to 100% avoid such thing , and yes, there is nothing wrong with this since it is like elections. People ask for support to get the result they want. Think of petitioning.

Also, @Paul Buonopane said only 4 members signed up after this thread was started. These discussions have taken the thread of topic.