Stephen Schubert's Playtest Reports

[JohnSnow]

Correct. Dice misweighting can have an effect, but in a pure plastic die the difference is only statistical. (As in, it might show up across ten thousand rolls, but not noticable at the table.) Vegas cares about it because of volume, but we tend not to.

On a d20, the more likely cause is an imperfect solid, which is worn more on some edges than others. Because a d20 is more "round" than a d6, this comes up in a way it just doesn't in Vegas.

So a player might actually have a LUCKY die which, due to a combination of factors, tends to come up 20 (or 19, or whatever) more frequently. However, since all dice (especially plastic ones that are mostly round) tend to wear as they're used, it's likely to change. So over time, it's usually self-correcting.

Statistics says that extremely unlikely events happen all the time. They just can't be relied on to happen.

 

[Keenath]

On a d20, the more likely cause is an imperfect solid, which is worn more on some edges than others. Because a d20 is more "round" than a d6, this comes up in a way it just doesn't in Vegas.

Eh... I still think any rolling differences from edge-rounding would be pretty minor and not terribly obvious at the table. I mean, even a 5% difference would still only manifest as a single extra 20 out of twenty rolls.

I put "lucky dice" and such things down to reporting bias* rather than any physical effect.

*Reporting bias: Remembering the events that support your idea and ignoring evidence that contradicts it. You remember the night you rolled three crits in a row and forget about the six weeks of totally average or below-average rolling.

 

[JohnSnow]

Eh... I still think any rolling differences from edge-rounding would be pretty minor and not terribly obvious at the table. I mean, even a 5% difference would still only manifest as a single extra 20 out of twenty rolls.

I put "lucky dice" and such things down to reporting bias* rather than any physical effect.

*Reporting bias: Remembering the events that support your idea and ignoring evidence that contradicts it. You remember the night you rolled three crits in a row and forget about the six weeks of totally average or below-average rolling.

Oh, I absolutely agree that reporting bias is a much more likely "cause." But on the off chance that you do see it, consistently and repeatedly, there are explanations that can make sense. Just like not all coins come up heads 50% of the time, not all 20-sided-dice produce any given number exactly 5% of the time.

By way of another example, basic economic theory says it's virtually impossible to "beat the market" with your investments. And yet, there are people like Warren Buffett, who have done this consistently. Does he disprove the theory? No. Because part and parcel of the theory is that even though it's "virtually impossible," if you have enough people in your pool, someone like Buffett probably exists.

Probability is great for predicting beforehand whether a given occurrence is "likely." But it's totally lousy for after-the-fact analysis.

 

[Mustrum_Ridcully]

*Reporting bias: Remembering the events that support your idea and ignoring evidence that contradicts it. You remember the night you rolled three crits in a row and forget about the six weeks of totally average or below-average rolling.

You mean stuff like "Oh my god, a Fortitude Save - My Fighter/Paladin/<insert other class with high fortitude save here> always fails these kind of saves! Remember guys, no more Reincarnates or Ressourections - you can sell anything except the sword and the belt for a true ressourection!" *roll* "Aaargl..." ?

 

[Steely Dan]

his character sheet is such a chicken-scratch that HE can't read it half the time, let alone ME.

I always type up all my player's character sheets in the latest format (the one introduced in DMG II) and hand them out a fresh one at the beginning of each session; and of course a copy for myself.

Really helps.

 

[Plane Sailing]

*Reporting bias: Remembering the events that support your idea and ignoring evidence that contradicts it. You remember the night you rolled three crits in a row and forget about the six weeks of totally average or below-average rolling.

My sorcerer has used truestrike to ensure an enervation ray hits the target exactly twice. On each case I rolled a 1(!). He is considering whether or not truestrike actually does what it says on the tin

 

[RangerWickett]

Heh. I watched one of my players roll five natural 20s in a row. Damndest thing I've ever seen, and had the whole table cheering. This sort of thing does happen.

Clearly the DM was cheating to let it happen.

 

[Thyrwyn]

My sorcerer has used truestrike to ensure an enervation ray hits the target exactly twice. On each case I rolled a 1(!). He is considering whether or not truestrike actually does what it says on the tin

I ran a 2e campaign years ago. The Thief in the party only ever rolled 1's or 20's on his backstab attempts (which given the nature of 2e were accordingly rare): over the course of the campaign character's career he rolled 8 1's and 3 20's

 

[Rechan]

I think this means that WotC is just increasing the frequency with which 20s are rolled at the table.

This is just another example of WotC sapping the "Fun" out of rolling a 20. By making them more common, WotC has tampered with the rarity of rolling 20.

Probability. Is this sacred cow worth slaughtering, WotC?

 

[Keenath]

Oh, I absolutely agree that reporting bias is a much more likely "cause." But on the off chance that you do see it, consistently and repeatedly, there are explanations that can make sense. Just like not all coins come up heads 50% of the time, not all 20-sided-dice produce any given number exactly 5% of the time.

Hehe, well, actually no coins or dice produce any given value exactly -- Interestingly, probability demands that it be also exceedingly rare to come up precisely the predicted value, no matter how many tests you run. A fair coin flipped X times will almost always be a few Heads more or less than X/2, and sometimes quite a bit off; by the same token, a fair d20 rolled a million times will be CLOSE to 5% each -- but will never be entirely evenly distributed. (Technically, it COULD happen, but the chances of precisely correct distribution become vanishingly small as you increase the number of tests.)

You mean stuff like "Oh my god, a Fortitude Save - My Fighter/Paladin/<insert other class with high fortitude save here> always fails these kind of saves! Remember guys, no more Reincarnates or Ressourections - you can sell anything except the sword and the belt for a true ressourection!" *roll* "Aaargl..." ?

I.... I'm not sure what you meant by that. Reporting bias means that an objective analysis would show that your "lucky die" is no luckier than any other die, but psychologically, you remember the times when the "lucky die" came through in a pinch -- times when it met your expectation of good luck -- and you tend to forget that it often doesn't come through.

Or are you saying it's reporting bias when a high fort guy feels like he always fails fort saves (which is, yes, reporting bias)?

Reporting bias (AKA 'recall bias') is a bit of an optimist/pessimist thing -- depending on what you expect, you tend to remember events that agree with you. It happens all the time on the boards here -- you see somebody claim that everyone agrees with them, when actually there are three people agreeing and dozens refuting. They just mentally focus on the agreement and overlook the disagreements.

My sorcerer has used truestrike to ensure an enervation ray hits the target exactly twice. On each case I rolled a 1(!). He is considering whether or not truestrike actually does what it says on the tin

Hehehe.

I think this means that WotC is just increasing the frequency with which 20s are rolled at the table.