This one is very odd.

The LGS I frequently play FNM at does a lot of random giveaways at the end of FNM. Packs of Modern Masters, $10 prepaid snack cards, etc. They used to do so by having everyone roll D20s, which was kind of fun, but very time consuming, and a logistical headache. There were also issues with people “rolling just for fun” or doing “practice rolls” before the roll-off was announced and doing other sketchy-looking behaviors.

A few months ago, we moved to using the “Random Player” button built in to the event reporting software.

All was well, until one player commented “The same players keep winning. This isn't random at all.” As a mathematician by profession, I can say with confidence that the results I've seen don't fall outside of statistical norms. The problem is that the idea that the “Random Player” button is rigged has caught on like wildfire. I'm sure half the people complaining about it are doing so out of jest, but a large group of players have bought into the hype, and there are tinfoil hats sprouting up all over the place.

One day, we tried pushing the button until it hit each of the 15 registered players at least once. It took about 50 tries until it finally got the last player. That fell in line with statistical expectations, but somehow further convinced a few people that certain names were unlikely to be hit. >.<

So, I have two questions:

1. Does anyone know anything about the Random Player button that I don't? I highly doubt that there's a known issue, but I have to ask.
   
2. Any ideas on the best way to quell the conspiracy theories?
   

Edited Glenn Fisher (Sept. 4, 2014 02:45:12 PM)

Originally posted by Glenn Fisher:

Any ideas on the best way to quell the conspiracy theories?

1. Before the event, use the Random Player feature a number of times. Many software-based random number generators need a “seed” to be truly random, or they'll just repeat the same sequence. Maybe even hit “select random” after each player is registered (if that doesn't slow it too much).
   
2. Announce “OK, folks, we've got an updated version of the software, this should be much better, now!” :)
   

Or, you could just write your own, and seed the random function with the time. A bit more ambitious…

d:^D

Might I suggest using a service such as www.random.org for this sort of thing. At the time of the draw, use the standings as a benchmark of which number corresponds to which player. Also, random.org is considered true random, compared to the formula for the random player button in Wizards Event Reporter. My Local Game Store has been doing this for years, with no complaints whatsoever.

I second the use of Random.org. It's very transparent and does what it says
on the tin. I've used it for everything from pre-registration giveaways
(you could win a pack if you prereg for sunday draft!) to making tables in
Commander League.

WER presumably generates pseudorandom values using the system clock to
seed, just like pretty much everything else, whereas random.org uses
atmospheric noise, which will sound fancy and different. Functionally there
will be little to no difference, but hey, perception is just as important
as reality in customer service sometimes - as long as the reality involves
delivering a good product or service anyway, that's fine!

I would caution you about lying about WER updates or implying that they've
updated the random player function. I would also caution you about giving
in to the rumors that it's somehow rigged. Make sure the players understand
that you have faith in the software's capability to pick random players
(its other functionality notwithstanding, perhaps) but that you're willing
to placate them by using a different system.


On Thu, Sep 4, 2014 at 6:12 PM, Seth Black <
forum-12344-f205@apps.magicjudges.org> wrote:

> Might I suggest using a service such as www.random.org for this sort of
> thing. At the time of the draw, use the standings as a benchmark of which
> number corresponds to which player. Also, random.org is considered true
> random, compared to the formula for the random player button in Wizards
> Event Reporter. My Local Game Store has been doing this for years, with no
> complaints whatsoever.
>
> ——————————————————————————–
> If you want to respond to this thread, simply reply to this email. Or view
> and respond to this message on the web at
> http://apps.magicjudges.org/forum/post/79135/
>
>
> Disable all notifications for this topic:
> http://apps.magicjudges.org/forum/noemail/12344/
> Receive on-site notifications only for this topic:
> http://apps.magicjudges.org/forum/noemail/12344/?onsite=yes
>
> You can change your email notification settings at
> http://apps.magicjudges.org/notifications/settings/
>
>


–
-Eric Levine
Level 3 Judge
Amherst, MA

One LGS here uses tickets with matching numbers, with the one half being randomly drawn from a tumbler or shaken box. Old skool random tech.

Have everyone participate in the same little experiment you did but more hands on for the unbelievers: have them roll their D20 until every number has come up once. That should help them realize what random means and statistical likelihoods and all that.

Originally posted by John Winter:

Have everyone participate in the same little experiment you did but more hands on for the unbelievers: have them roll their D20 until every number has come up once. That should help them realize what random means and statistical likelihoods and all that.

That takes considerable more tries than using dice to determine random door prize winners. Also, it doesn't really provide any understandable argument for someone who is suspicious of WER being ‘totally random’.

With the amount of complaints that WER as a piece of software gets, it's not extraordinary for people to not believe it's random number generator. This is simply solved by using a different method. For < 20 players, rolling X D20s should be an easy and quick method to use. For more than 20, it's slightly more complicated (or needs more expensive dice, then again, most game shops have these in stock anyway)

Originally posted by Jasper Overman:

That takes considerable more tries than using dice to determine random door prize winners. Also, it doesn't really provide any understandable argument for someone who is suspicious of WER being ‘totally random’.

Oh I'm not saying use that as a way to determine prizes, just as a way to quell their conspiracy theories. The point of the experiment is that it will take considerably more tries. Once they see that with their own hand and their own die that it takes just as many attempts as WER took, they might start to come around. Short of bringing a white board and teaching a few weeks worth of Statistics, it's probably the best way to show the crowd what random results look like.

Edited John Winter (Sept. 8, 2014 08:18:25 AM)

I've had to deal with something similar a couple of times. My solution:

“I'd be happy to explain the details and discuss why I believe I am right. But I'm busy with the tournament right now, and it's already late, so I can't do it now and you'll just have to trust me for tonight. But I understand your concern, so feel free to hit me up on Facebook, we'll pick a place, a date and a time, I'll bring a Statistics textbook, and we'll go through it in as much detail as you like.”

I have yet to be taken up on that offer by anyone.

If WER is using the system clock to generate the seed value and presumably the event is ending at roughly the same time every night, you could get around the similar/same seed values by mixing up the giveaway times. I would suggest doing the giveaway after the end of a round decided by the roll of a d6 instead of always at the end of the event; by doing that, you get around the presumably predictable seed value while still using WER to generate the value. Since this keeps the random player selection within WER & doesn't use an outside solution like random.org, it also helps dispel any kind of player impression that WER is rigged.

I would like to think that players were suspicious that WER is flawed not
rigged, but I know not everyone is as charitable as me (and I'm not very
charitable)

That's not how RNG works :)

-Eric Levine

I could be off-base, but one potential issue I would see with moving to something like random.org for this playgroup that has already raised a fuss about WER is the impression that you're admitting to the players that WER is rigged, which could decrease confidence in the randomness of pairings, calculation of tie-breakers, etc.

I agree with John. Switching to another method is a concession that the system is actually rigged, and might cause other problems among the more paranoid players. Just try to explain to them as best you can that “random” doesn't mean “perfectly smooth and even”. It means “random”. You don't know what's going to happen, and very often the result isn't what you were expecting.

Originally posted by Andrew Herber:

If WER is using the system clock to generate the seed value and presumably the event is ending at roughly the same time every night, you could get around the similar/same seed values by mixing up the giveaway times…

This could only be a problem if WER is using a very coarse unit of measurement as the random seed - like the hour of day or something. Generally programs will use the millisecond, as it's easy to retrieve. They typically also mix in some other value that will change each time the program is started - the classic example is the program ID, which is will almost never be the same between runs, unless you write something designed to keep relaunching the app until it arrives at the desired ID.