Double Vision

[tristenc]

It seems whenever I roll 3 or more separate dice (of the same type) in a single command (like in the case of rolling hits or saves for multiple characters in games I run) the 2nd and 3rd rolls are always Identical.

Is there a way to fix this on the roller? I would really rather not make 30 individual rolls when I can make , but it certainly screams into question any randomness of the site's roller. Several of my players are getting frustrated with rolls from this site as well, and while there is nothing quite as blatant as the perpetual double-rolls mentioned above, it is getting very difficult to give the site the benefit of the doubt from a randomness perspective.

From very recent rolls in one of my campaigns: [There are many more examples]
Dex checks Sir Karl, Pytor, Cross, Art [1d20] = 2, [1d20] = 19, [1d20] = 19, [1d20] = 14
Dex checks Mira, Elsa, Pytor [1d20] = 6, [1d20] = 3, [1d20] = 3
Bones fall on: Jerry, Pyros, Pytor, Sir Karl [1d20] = 19,[1d20] = 16,[1d20] = 16,[1d20] = 11
Bones fall on: Adran, Jerry, Pyros, Sir Karl [1d20] = 1,[1d20] = 18,[1d20] = 18,[1d20] = 13
Bones fall on: Pytor, Cross, Art, Merrick [1d20] = 10,[1d20] = 7,[1d20] = 7,[1d20] = 2
Bones Strike: Jerry, Pyros, Pytor, Cross [1d10] = 10,[1d10] = 7,[1d10] = 7,[1d10] = 2
Bones fall on: Adran, Jerry, Mira [1d20] = 11, [1d20] = 8,[1d20] = 8
Bones fall on: Pyros, Sir Karl, Merrick, Pytor [1d20] = 4,[1d20] = 1,[1d20] = 1,[1d20] = 16
Bones fall on: Adrestia, Art, Cross, Elsa, [1d20] = 10,[1d20] = 7,[1d20] = 7,[1d20] = 2
Bones Strike: Adran, Pytor, Adrestia, Cross [1d10] = 2,[1d10] = 9,[1d10] = 9,[1d10] = 4
Dex checks Sir Karl, Pytor, Cross, Art [1d20] = 6, [1d20] = 3, [1d20] = 3, [1d20] = 18



Stumbled upon from another Thread:

Here goes

Drive [1d10] = 5 [1d10] = 2 [1d10] = 2 [1d10] = 7 [1d10] = 1

Drive [1d10] = 9 [1d10] = 6 [1d10] = 6 [1d10] = 1 [1d10] = 5

5 6+ and 2 1's so 3 total. As usual I am a horrible roller on the dice roller. This is how I plan on handling it.

[dmw71]

That's an interesting find. This will definitely be a Greg question.

[ToniXX]

It is actually very hard for computers to generate random numbers. For a while I was using a service that generated random numbers based on atmospheric noise, but that site is no longer available. But, I think I've resolved this issue. Try it again and let us know what you find.

[Rex]

Actually more like near impossible than very hard for computers. Dice are actually not that great either. When I was getting my minor in statistics we did an experiment with 1000 dice each rolled 1000 times (it was a whole class) and recorded every roll. Less 10 percent of the dice were actually truly random based on a 1000 rolls. The best dice by far were casino grade 6 sided, of the 20 sided we tested none were truly random. We theorized that the more sides the more likely wear would affect the results.

[GreyWolfVT]

Actually more like near impossible than very hard for computers. Dice are actually not that great either. When I was getting my minor in statistics we did an experiment with 1000 dice each rolled 1000 times (it was a whole class) and recorded every roll. Less 10 percent of the dice were actually truly random based on a 1000 rolls. The best dice by far were casino grade 6 sided, of the 20 sided we tested none were truly random. We theorized that the more sides the more likely wear would affect the results.

Ah but what of dice made out of metal? I've got a set of those. At any rate yes I would expect a program to "randomly" come up with roll results or numbers is never truly random it can only do what it is coded to do. Closest you can come to is almost random but there generally is a pattern or something along those lines. (Not a coder but I understand how it works in theory.)

[Rex]

Metal are probably better, we didn't have any to test in the early 90's. The best die was actually an old 6 sided made from real ivory that someone brought in from home, had to be 150+ years old and it was nearly perfect for 1000 rolls.

[rredmond]

Of course metal dice are harder on tabletops. And oooo boy, I can only imagine the caltrop damage if they were stepped on. My guess is MUCH more than a d4's worth.

I have a VERY old brown d20 that tends to roll high at home. My oldest calls it the "dice of doom" and she swipes it at the beginning of every game.

[GreyWolfVT]

Of course metal dice are harder on tabletops and oooo boy, I can only imagine the caltrop damage if they fell on the floor. My guess is MUCH more than a d4's worth.

The d4's are the caltrops in my house Ron.

[rredmond]

Mine too Ed, thus the winky face!
I can imagine a metal d4 actually drawing blood. As cool as that sounds, and all four of my daughters would agree it's cool, the missus on the other hand...

[tristenc]

hmm, lot of 1s and 20s but no doubles
Thanks for the quick response!

test for duplicates [1d20] = 3,[1d20] = 4,[1d20] = 10,[1d20] = 9,[1d20] = 11
test for duplicates 2: [1d20] = 13,[1d20] = 3,[1d20] = 20,[1d20] = 13,[1d20] = 11
test for duplicates 3: [1d20] = 5,[1d20] = 20,[1d20] = 18,[1d20] = 16,[1d20] = 1
test for duplicates 4: [1d20] = 2,[1d20] = 17,[1d20] = 1,[1d20] = 14,[1d20] = 1
test for duplicates 5: [1d20] = 13,[1d20] = 18,[1d20] = 16,[1d20] = 11,[1d20] = 9

[Quonundrum]

Random.org has a new API with a free non-profit license that allows 60,000 requests per month.

https://api.random.org/dashboard

[ToniXX]

Random.org has a new API with a free non-profit license that allows 60,000 requests per month.

https://api.random.org/dashboard

Yes, that's the one I was using. In the last hour i've signed up for the new API and I'm trying to get the JSON working...
standby

[tristenc]

as for plastic dice, with a proper suspension liquid (oil or heavily salted water) you can tell if it floats to the same number or side.

also:

Screenshot_20200609-132408_Firefox~2.jpg (128.81 KiB) Viewed 1714 times

edit: i put the gross pic in a spoiler so we don't have to see it

[dmw71]

hmm, lot of 1s and 20s but no doubles

Nice. Thanks for confirming.

[ToniXX]

I've updated the dice roller using the new random.org JSON method. I'm using the free tier of the API so it is limited to 1000 random numbers in 24 hours. I hope that should cover us! If not, I'll upgrade.

Post your results

[dmw71]

i put the gross pic in a spoiler so we don't have to see it

Thanks for that. I mean, that d4 wasn't even that pointed (each time was flattened). Yikes.

[ToniXX]

i put the gross pic in a spoiler so we don't have to see it

Thanks for that. I mean, that d4 wasn't even that pointed (each time was flattened). Yikes.

Yea, that was the easy part. The hard part was learning JSON in 45 minutes! No biggie! LOL

[tristenc]

I think the die was probably photoshopped into a ' stepped on glass' pic

[rredmond]

I think the die was probably photoshopped into a ' stepped on glass' pic

Or someone had to really work their foot onto it! Twas gross

[tristenc]

So, not to revive the issue, but...

Return of the (un)Living Doubles seems to be at hand...

Adran, Karl, Pytor initi [1d10] = 1,[1d10] = 8,[1d10] = 8
skel init [1d10] = 8, [1d10] = 5, [1d10] = 5, [1d10] = 10
zom init [1d10] = 7, [1d10] = 4, [1d10] = 4, [1d10] = 9