May Daggermaster Nerf Rumour

[xXElaDriNDrizZztXx]

No, it isn't a good indicator of anything, other than maybe that's how one fight might go. The sample size isn't statistically significant so it basically proves nothing about the actual math involved. And that's on top of the other issues like not maxing quarry dice on a crit, etc.

That's not how it works. In this case, you don't even need an inadequate number of random die rolls like you've done here. You know the distribution of results over a large number of rolls so you can just calculate the results.

Wrong, because of the little thing called chance* It's something you should look into. Sometimes, you can crit often, when that happens, yes it is significantly different. The odds are just as good you'll roll 1's 2's and 3's instead.

Please tell me you meant this tongue in cheek

"Battles last less than 20 rounds, therefore any 20 rounds will work to prove my point, regardless of whether or not they are statistically sound."

When you're doing any kind of statistics, 20 rounds aren't a good sample, period. You assume an infinite number of trials.

-O

I didn't know we weren't playing dungeons and dragons and we were playing DO IT UNTIL WE'RE RIGHT. The numbers are fine, they work and your only response is that you have to do non-realistic testing. Congrats, I'm done with this thread. You fail to grasp the concept of reality of the game.

 

[Eternal54]

I didn't know we weren't playing dungeons and dragons and we were playing DO IT UNTIL WE'RE RIGHT. The numbers are fine, they work and your only response is that you have to do non-realistic testing. Congrats, I'm done with this thread. You fail to grasp the concept of reality of the game.

I tried to stay out of this, I really did, but this last one made me lol. So just that we are clear, I am yet another person who thinks the "Non-realistic" method of testing is the proper form. As you said chance plays a significant impact on the roll of the die, especially the fewer amount of times you roll. You have a higher chance of getting an average of 20 with one roll of the die than you do with five rolls and so on.

Even if you do not grasp the understanding of the process, do not be close minded and assume everyone else is wrong. We very well may be, but if everyone says something that contradicts your views at what point do you step back and say, "Maybe I'm not looking at it the right way?"

 

[Obryn]

Wrong, because of the little thing called chance* It's something you should look into. Sometimes, you can crit often, when that happens, yes it is significantly different. The odds are just as good you'll roll 1's 2's and 3's instead.
...
I didn't know we weren't playing dungeons and dragons and we were playing DO IT UNTIL WE'RE RIGHT. The numbers are fine, they work and your only response is that you have to do non-realistic testing. Congrats, I'm done with this thread. You fail to grasp the concept of reality of the game.

This is simply wrong in so very many ways. I strongly urge you to pay attention in your math and statistics classes.

On some 20-roll samples, it will come out like your ... "spreadsheet." On some 20-roll samples, it won't. On some, the axe might crit 3 times or 0. On some the Dagger might crit 10 times, or no rolls will be above 9. The key is to work with probabilities in such a way that this individual variation in samples will even out. Your sample of 20 rolls isn't a characteristic one; it's far better to either run a huge number of rolls (aka the Monte Carlo method, which requires a lot more than 20 events), or use relatively simple probability (aka math) to determine the average results of an infinite number of rolls.

Quick analogy....

Your assertion that 20 attack rolls is "just fine" to prove your point would be like me flipping a coin 20 times, noticing that it came up Heads 13 times, and saying that it proves my point that Heads will come up 65% of the time any coin is flipped in the future. Clearly, you'd say that I'm off my rocker, and that it'll come up heads 50% of the time.

Variations like this are extremely likely in small samples. If I were to flip the same coin (for example) 10,000 times, it will come up Heads about 50%. It still might be something like 5,051/10,000 or 4,963/10,000, but the large sample size brings it closer to the mathematical realities of the situation. On 1,000,000 flips, it will be even closer to 50/50. On a theoretically infinite number of flips, it will be 50/50.*

-O


* Quick aside, sblocked, that has nothing to do with the discussion above.
[sblock]Actually, with a coin flip, it's not quite 50%. You're always better off calling the side that's facing up during the flip, if it's going to land on the ground. Or, calling the side that's facing down before the flip, if it will be caught and flipped over one more time.

Basically, when a coin that was heads-up is flipped, it might go HTHTHTHTHTH or HTHTHTHTHTHT or some variation thereof. At no point during the flip will there be more tails than heads in the sequence, assuming Heads was up at first.

But this is neither here nor there for the mathematical ideal of a coin flip, which is really what we're talking about.
[/sblock]

 

[Dannager]

Well our table sits 7 so to speed up rounds we also play one roll to attack for AoE attacks (Not mutle i.e One or Two Target attacks).

Ah, that explains it.

It also made more sense to us RP wise. Why would a fire ball be hotter here, then colder further down, than hottr again?

The fireball isn't hotter in one place, and colder further down (or, at least, it isn't automatically assumed to be the case). Do note that, by the rules, you roll damage once for all targets of an AoE, so that fireball would be "equally hot" everywhere. The attack rolls are to see if the target's defensive abilities were enough to get out of the way (or resist the attack) in time.

 

[Fifth Element]

Wrong, because of the little thing called chance* It's something you should look into. Sometimes, you can crit often, when that happens, yes it is significantly different. The odds are just as good you'll roll 1's 2's and 3's instead.

...which is why you use these things called averages (it's something you should look into). It's been a while since my mind has been boggled. I appreciate that, I suppose, in some strange way.

 

[xXElaDriNDrizZztXx]

I tried to stay out of this, I really did, but this last one made me lol. So just that we are clear, I am yet another person who thinks the "Non-realistic" method of testing is the proper form. As you said chance plays a significant impact on the roll of the die, especially the fewer amount of times you roll. You have a higher chance of getting an average of 20 with one roll of the die than you do with five rolls and so on.

Even if you do not grasp the understanding of the process, do not be close minded and assume everyone else is wrong. We very well may be, but if everyone says something that contradicts your views at what point do you step back and say, "Maybe I'm not looking at it the right way?"

This is simply wrong in so very many ways. I strongly urge you to pay attention in your math and statistics classes.

On some 20-roll samples, it will come out like your ... "spreadsheet." On some 20-roll samples, it won't. On some, the axe might crit 3 times or 0. On some the Dagger might crit 10 times, or no rolls will be above 9. The key is to work with probabilities in such a way that this individual variation in samples will even out. Your sample of 20 rolls isn't a characteristic one; it's far better to either run a huge number of rolls (aka the Monte Carlo method, which requires a lot more than 20 events), or use relatively simple probability (aka math) to determine the average results of an infinite number of rolls.

Quick analogy....

Your assertion that 20 attack rolls is "just fine" to prove your point would be like me flipping a coin 20 times, noticing that it came up Heads 13 times, and saying that it proves my point that Heads will come up 65% of the time any coin is flipped in the future. Clearly, you'd say that I'm off my rocker, and that it'll come up heads 50% of the time.

Variations like this are extremely likely in small samples. If I were to flip the same coin (for example) 10,000 times, it will come up Heads about 50%. It still might be something like 5,051/10,000 or 4,963/10,000, but the large sample size brings it closer to the mathematical realities of the situation. On 1,000,000 flips, it will be even closer to 50/50. On a theoretically infinite number of flips, it will be 50/50.*

-O


* Quick aside, sblocked, that has nothing to do with the discussion above.
[sblock]Actually, with a coin flip, it's not quite 50%. You're always better off calling the side that's facing up during the flip, if it's going to land on the ground. Or, calling the side that's facing down before the flip, if it will be caught and flipped over one more time.

Basically, when a coin that was heads-up is flipped, it might go HTHTHTHTHTH or HTHTHTHTHTHT or some variation thereof. At no point during the flip will there be more tails than heads in the sequence, assuming Heads was up at first.

But this is neither here nor there for the mathematical ideal of a coin flip, which is really what we're talking about.
[/sblock]

...which is why you use these things called averages (it's something you should look into). It's been a while since my mind has been boggled. I appreciate that, I suppose, in some strange way.

Dudes... No... You can't average probability! It's individual for each roll

 

[DracoSuave]

Dudes... No... You can't average probability! It's individual for each roll

:facepalm:

Probability IS an average.

It's an average taken by dividing qualifying outcomes by the total number of outcomes.

 

[Dr_Ruminahui]

What Dracosuave said.

Basically, a simplified version of what we are doing is this:

- what is better? 8 static damage, or 2d6?
- what is better? Hitting on an 11+ for 6d6 damage, or hitting on a 16+ for 6d12 damage?

The only way of answering either of theose questions is by averages. And that's what statistics is all about. Saying that 2d6 is better than a guaranteed 8 damage because on a particular roll one rolled a 12 is missing the point.


PS - if you took the second answer for either (from a straight damage standpoint), you are doing it wrong.

 

[UngeheuerLich]

What Dracosuave said.

Basically, a simplified version of what we are doing is this:

- what is better? 8 static damage, or 2d6?
- what is better? Hitting on an 11+ for 6d6 damage, or hitting on a 16+ for 6d12 damage?

The only way of answering either of theose questions is by averages. And that's what statistics is all about. Saying that 2d6 is better than a guaranteed 8 damage because on a particular roll one rolled a 12 is missing the point.


PS - if you took the second answer for either (from a straight damage standpoint), you are doing it wrong.

1. Individual events are statistically independent.

2. Average damage doesn´t telly you a lot on low hp enemies.

3. a difference between 17 damage and 150-200 is so significantly that going with the average is wrong in so many ways.

A single crit ends the combat against most monsters immediately. Every point of damage done before the crit happens is wasted. (In most cases) And then there are combats that last forever.

Even 2d6 may be more optimal than 8: If you are facing monsters that have 9 hp. In this case dealing 8 damage is equal to doing 4,5 damage. If you however are facing 16 HP monsters, dealing 8 static damage is even better than 2d6+3 (with an average of 9.5.

So just looking at the average and not taking variation into account on anything that is less than a big enough number on rolls is very wrong.

@Obryn: There is nothing as unlimited number of throws with an average of 50/50. If you had any number of throws where you had exactly 50/50, the next throw will unbalance it certainly.
It is just that the variation is going against zero when you have done enough throws.

A side not: Average and variation plays a big role when sepculating on the stock market.

 

[UngeheuerLich]

@ obryn: no it is not always better to chose head if head is up. (only fo this single event when you have no other information) You also have to look if the person always has the same side up before he throws, and throws in a comparable way (same height etc.). If you notice that the coin lands 16/20 times with side up that was down before the throw, i would always bet on the side which was down before the throw when its my turn to bet.

Compare it to the "Why does my bread always land on the marmelade side?"-problem...

And before someone comes and says: but it is better to bet on the other side, because it came up only 4/20 times until now: No, "gamers fallacy"....