Home Rules Tablets

[Rustybarnacle]

I just started a campaign with a new group a couple weeks ago and after 2 sessions I noticed a trend with one player.

Lots of critical rolls, no ones came up.

Instead of dice he was using a tablet with a d20 gaming app. Anyone know if thats a trend I should be looking into? Anyone else ban tablets from dice rolls and go straight old school?

I tried searching on here but for some reason I don't have a search link. Maybe becasue I'm new here?

Anyway I think I'm going to make the first house rule that all rolls are dice rolls and he can save the tablet for looking after his character. Just thought I'd check the crowd to see if there was anything out there.

[lsfreak]

I tried searching on here but for some reason I don't have a search link. Maybe becasue I'm new here?

Searching is down due to... I don't really know. It's been down for the four months I've been sort-of back on GitP.

As for the dice thing, I prefer handrolled dice. With people I already know I'd be okay with computer-rolled, but I'd still want the exact name of the program they're using. A google search for "loaded d20 dice app" brings up exactly that with the first search result, it weights higher numbers heavier and nat 1's are impossible.

[Rustybarnacle]

Don't worry, I palm slapped myself pretty good for not googling it better. I did google tablet rules but didn't think to google tablet cheating or loaded dice.

Thanks!

[SiuiS]

The search function is down because it messes up the server. There are several threads about it in the appropriate branch of the forum.


Instead of just the house rule, ask the bloke if he's cheating. Say it's not cool, and the game is as much fun with dramatic failure as dramatic success. Play from there, and then bring up that, due to the ease of getting loaded programs you will use material dice only.

[molten_dragon]

Don't worry, I palm slapped myself pretty good for not googling it better. I did google tablet rules but didn't think to google tablet cheating or loaded dice.

Thanks!

Whenever these sorts of things come up, It's usually a player issue rather than an issue with specific dice or an app or something. If he's the kind of player that would knowingly use an app that fudges rolls for him, he'd fudge them on his own if using actual dice. Talk to the player about why he's doing it rather than just banning people from using dice rollers on tablets.

[Alaris]

Definitely go for "Old-School" dice rolling. It's much more fun that way anyway. Puts your fate in your hand. Even if you trust him, honestly, everyone should roll the old fashioned way.

[Barsoom]

Anyway I think I'm going to make the first house rule that all rolls are dice rolls and he can save the tablet for looking after his character. Just thought I'd check the crowd to see if there was anything out there.That ... would not be a house rule. The Player's Handbook specifically states, for example, for an attack roll, to "roll a 20-sided die".

Now, if you'd allow a tablet app to generate a random number instead of rolling a die, that would be a house rule. [and one I would gladly see revoked]

[Phelix-Mu]

The only reason to allow die-rolling programs is to make damage rolling at high-levels more streamlined to speed up play. My mighty wallop monk was eating up lots of table time (if it can be called that over skype) when I was counting 12d8 five times by hand. Same thing goes for high-level casters that aren't in the habit of maximizing everything.

But attack rolls, skill checks, saves, anything that just involves a handful of dice should be rolled by hand. Particularly if you meet in person and rolls can be witnessed. Not that you should need to police so heavily with friends, but cultivating a base level of honesty can be difficult with strangers or acquaintances.

Hand rolling also improves math skills. GO MATH!:smallbiggrin:

[Answerer]

1. Computers are not capable of true randomness without specialized hardware (or access via the internet to a computer that does).

2. Pseudo-random number generating algorithms are a major branch of computer science research, and have gotten very, very good. They are not truly random (i.e. do not rely on one for anything cryptographic), but they produce a sequence of numbers that is very close to uniform in its distribution of values.

3. Physical dice are not generally perfect, due to manufacturing variations and wear'n'tear. These produce extremely slight (well, in the majority of cases) biases in any given dice.

4. Weighted dice and apps that skew rolls both exist, and are easily available.

My point is that really, a computer is not appreciably worse (or better) than a physical die for random number generation. A physical die will be random but biased, an algorithm will not be truly random but it should be unbiased, and in any event both the bias and the pseudo-randomness would require careful statistical analysis of thousands of rolls to detect. (It is, for obvious reasons, much easier to run this analysis on an algorithm than it is for a die.)

By the same token, intentionally-skewed dice or algorithms would take similar sorts of analyses to detect, and unless the skewing was very strong, would require similar numbers of rolls.

As with all things probabilistic, it is generally unwise to trust your gut on this sort of thing.

This neither means that you don't have a problem nor that you should do nothing about it, I just want to sort of... temper the discussion a little. Moreover, the problem is not with the use of apps, but with the person who is selecting a skewed app.

If someone wishes to cheat, they can do so with dice as well as with an app. Deal with the problem at its source: the cheater. If you cannot trust this guy, why is he gaming with you? New group, OK, can his friends vouch for him? If so... probably for the best to let it lie. Or find another new group. If not, if they have doubts themselves... then confront him about it. Politely, but firmly.

You could even say something along the lines of "look, I know statistics, I know that string of rolls was entirely possible. I know better than to assume just because something looks lop-sided in one session, that it implies anything. But someone else in the group approached me about it, and what it comes down to is that even if someone suspects, incorrectly, that something's going on, it's breaking immersion and ruining his game. So would you mind just using the dice we've got? There are plenty."

[TuggyNE]

1. Computers are not capable of true randomness without specialized hardware (or access via the internet to a computer that does).

2. Pseudo-random number generating algorithms are a major branch of computer science research, and have gotten very, very good. They are not truly random (i.e. do not rely on one for anything cryptographic), but they produce a sequence of numbers that is very close to uniform in its distribution of values.

I agree quite extensively with the rest of your post, but felt I should note that for nearly all cryptographic purposes CSPRNGs found on modern OSes and frameworks are actually perfectly suitable, due to a great deal of careful algorithm choice and tuning. (Generally they use cryptographic hash functions to condense and mix available entropy.)

But yeah, computer RNGs can be at least as good in the technical sense as dice (i.e., indistinguishable from true random for all practical purposes); the problem is entirely on the user side.

[Answerer]

I agree quite extensively with the rest of your post, but felt I should note that for nearly all cryptographic purposes CSPRNGs found on modern OSes and frameworks are actually perfectly suitable, due to a great deal of careful algorithm choice and tuning. (Generally they use cryptographic hash functions to condense and mix available entropy.)
There's a reason I chose "rely" as my verb in that sentence, yes. Modern cryptographic techniques use pRNGs, but they do not rely on them alone because of their pseudo-randomness. For instance, if an attacker had access to one's database (read: the encrypted versions of passwords), and could watch their new passwords getting created and encrypted as they created new accounts they could (if I understand things correctly) figure out what pRNG is being used (assuming that the time of the account's creation is being used as the seed, which it almost certainly is). That would allow trivial decryption of arbitrary encrypted passwords in the database, which no encryption should ever allow, even in these sorts of worst case scenarios.

But then, I am not anything like a cryptologist by any means. It is an aspect of computing that I am happy to have others take an interest in.

(really, I'm waiting for proper quantum-locked encryption so we don't have to worry about this any more :smallwink: )

[TuggyNE]

OT cryptography:
There's a reason I chose "rely" as my verb in that sentence, yes. Modern cryptographic techniques use pRNGs, but they do not rely on them alone because of their pseudo-randomness. For instance, if an attacker had access to one's database (read: the encrypted versions of passwords), and could watch their new passwords getting created and encrypted as they created new accounts they could (if I understand things correctly) figure out what pRNG is being used (assuming that the time of the account's creation is being used as the seed, which it almost certainly is). That would allow trivial decryption of arbitrary encrypted passwords in the database, which no encryption should ever allow, even in these sorts of worst case scenarios.

I'm only an armchair cryptographer myself, but I do try to keep up on the literature as best I can.

However, that scenario doesn't sound right; any decent implementation will use RNG-derived values for only one column in that table: the salt. The password field will instead contain a combined hash of the password and the salt, and authentication will pull out the salt, hash it along with the provided password, and compare that with the hash stored in the database. The only practical method for reversing hashes is rainbow tables, which are pretty much entirely spoiled by proper use of salts.

Importantly, the salts are stored in plain-text; the method used to derive them isn't especially important, as long as few or no entries reuse salts. So knowing the entropy involved is unimportant.

Of course, that may be what you mean by saying that no implementations rely on CSPRNGs without hardware entropy, except that even if you have true random numbers, storing the actual password, even encrypted, is a terrible idea. (You need to store a secret key in some fashion to decrypt them, which can be stolen in various ways.)

[ArcturusV]

Well I had a similar problem with one of my groups. What I started doing is anytime I called for a series of rolls?

I'd flip a coin, letting them SEE I flipped it, but not what it landed on behind my DM screen. Then they'd roll.

Heads? Results as normal, 1 is low, 20 is high.

Tails? Results were inverted, 1 is high, 20 is low.

Interestingly enough, no one had a problem with this. Except the one I suspected of rigging his rolls. Who one day after getting a couple of Crit Failures on inverted rolls said I should tell everyone is a roll ended up inverted or not before anyone ever rolled, instead of holding the mystery of the coin flip until after the dice/program came up with a result.

I didn't have to defend myself, my players pointed out it defeated the purpose of Inversion chances if I told them. And that if you weren't using rigged dice/programs it shouldn't matter if it was inverted or not.

[Story]

The only reason to allow die-rolling programs is to make damage rolling at high-levels more streamlined to speed up play. My mighty wallop monk was eating up lots of table time (if it can be called that over skype) when I was counting 12d8 five times by hand. Same thing goes for high-level casters that aren't in the habit of maximizing everything.


They're also useful for anything that needs to happen between sessions.

[Rustybarnacle]

The responses here have been awesome, thanks!

This is a new group of random people off the internet. Until our first session none of us had ever met before. So no vouching or previous relationships to go on.

After 2 sessions though, first and second impressions tell me this is whats happening and that he should be rolling dice.

Any way for me to look at his dice to see if they are rigged as well? I remember playing with a DM way back in the way back who looked at my dice the first time we played together but I don't know if he could actually tell if I had a problem or if he was just putting on a show.

[ArcturusV]

I suppose if you were REALLY sensitive you could tell that the weight was off balance? Only time I had a DM do that is one time when I wanted to use one of my Magic: The Gather d20 "spindown" dice because, well, they're neater looking than most D20s. But he noped it on the idea that because numbers were adjacent to one another (20 shares a border with 19, which shares a border with 18, etc) that somehow it wouldn't "roll fair". Even though it seems to based on experience. But I just shrugged and made due with a boring red crystal d20.

But I never was that attuned to minor fluctuations in balance and weight that would make a difference with dice. Nor had some weird belief that how numbers are placed next to one another determines the odds.

Simple truth is I know a lot of RPGers who will spend HOURS in a store just rolling dice to figure out which ones roll high, versus which ones will roll low, and will buy them.

[Pink]

I suppose if you were REALLY sensitive you could tell that the weight was off balance? Only time I had a DM do that is one time when I wanted to use one of my Magic: The Gather d20 "spindown" dice because, well, they're neater looking than most D20s. But he noped it on the idea that because numbers were adjacent to one another (20 shares a border with 19, which shares a border with 18, etc) that somehow it wouldn't "roll fair". Even though it seems to based on experience. But I just shrugged and made due with a boring red crystal d20.

While it is a bit paranoid, I can see the argument of this. Because high numbers are grouped together, a person could practice a way of holding the die and rolling it so that is would regularly fall on a high number area. Of course, I've never seen a player roll their dice so precisely that I could suspect this being the case myself (I think most of us like to hear the rattle of dice in our hand and see them spin accross the table, which pretty much make a precise drop impossible).

[Answerer]

Any way for me to look at his dice to see if they are rigged as well? I remember playing with a DM way back in the way back who looked at my dice the first time we played together but I don't know if he could actually tell if I had a problem or if he was just putting on a show.
Unless the bias is really obvious, there is no way the human hand would be able to tell with certainty that any bias is present. Biased weighting can be fairly subtle and still have a powerful statistical effect. Precision scales might be able to tell, but even that I'm somewhat dubious on. IIRC, Las Vegas casinos test dice by randomly selecting a few from every batch, and using a machine that automatically rolls them a few thousand times, recording each result, to check for biases. You obviously don't have that kind of equipment or the kind of time necessary to do it manually.

If you feel it necessary, either require everyone to use communal dice, or (probably simpler), use that "secret inversion" trick someone mentioned earlier in the thread. That's a pretty solid idea.


I suppose if you were REALLY sensitive you could tell that the weight was off balance? Only time I had a DM do that is one time when I wanted to use one of my Magic: The Gather d20 "spindown" dice because, well, they're neater looking than most D20s. But he noped it on the idea that because numbers were adjacent to one another (20 shares a border with 19, which shares a border with 18, etc) that somehow it wouldn't "roll fair". Even though it seems to based on experience. But I just shrugged and made due with a boring red crystal d20.
Actually, this is pretty reasonable. All dice are going to have subtle biases. By grouping numbers with numbers on the opposite side of the scale, these biases tend to get canceled out over time, since the uneven weighting will favor all of those numbers evenly (and disfavor the adjacent numbers on the opposite side of the scale evenly). Even if your dice wasn't intentionally skewed, its results will, over time, be statistically more biased than normal dice.

For example, a die might be weighted so the 1 is heaviest (most likely to land on the bottom), but in a "fair" die this will be very subtle (i.e. 1s will still roll close to as frequently as they should). Opposite the 1 is the 20, so it will roll slightly more 20s than it should. But it will also roll more 2s and what have you, and fewer 19s, because the 19 is next to the 1 and the 2 is next to the 20. If you use a spindown die, it will favor 20s... and also 19s and 18s, while disfavoring 1s and 2s and 3s and so on.

Of course, on "fair" dice (i.e. ones not intentionally biased), this effect will be very slight, and is just as likely to favor 10 while disfavoring 11 (whoopty-doo) as it is to favor 20 and disfavor 1.

[Big Fau]

I suppose if you were REALLY sensitive you could tell that the weight was off balance? Only time I had a DM do that is one time when I wanted to use one of my Magic: The Gather d20 "spindown" dice because, well, they're neater looking than most D20s. But he noped it on the idea that because numbers were adjacent to one another (20 shares a border with 19, which shares a border with 18, etc) that somehow it wouldn't "roll fair". Even though it seems to based on experience. But I just shrugged and made due with a boring red crystal d20.

There is a good reason to disallow spindowns: WotC weights the 1 side of the spindown d20s (except the From the Vault ones, for some reason) so the die is more resistant to the inevitable table bumps, thus ensuring the player's life total isn't getting tampered with. Using spindowns means you are using weighted dice.

And his point is actually closer to true than you give him credit. On a normal d20 each face of the die is placed opposite of a number that, when the two are added together, adds up to 21. 1+20, 2+19, 3+18, etc. Spindowns don't do this, which makes it easier to use legerdemain to "roll" higher numbers more often (in addition to the spindown's weight).


Spindowns should only be used to count things, not to roll.