new death save == leave your buddy on the floor for 3 rounds?

[komi]

Instead, the question I tried to answer is: What is the chance of dying on the THIRD round assuming you have been unconcious for TWO.

I came to the same answer as you, but by a different means. If you have a computer, it's easy to do these problems by just looking at all possible rolls. For three rounds, there's 27 possible combinations.

(X = Strike; - = nothing; O = stabilize)
--> X X X  729 / 8000
--> X X -  810 / 8000
--> X X O   81 / 8000
--> X - X  810 / 8000
--> X - -  900 / 8000
--> X - O   90 / 8000
    X O X   81 / 8000
    X O -   90 / 8000
    X O O    9 / 8000
--> - X X  810 / 8000
--> - X -  900 / 8000
--> - X O   90 / 8000
--> - - X  900 / 8000
--> - - - 1000 / 8000
--> - - O  100 / 8000
    - O X   90 / 8000
    - O -  100 / 8000
    - O O   10 / 8000
    O X X   81 / 8000
    O X -   90 / 8000
    O X O    9 / 8000
    O - X   90 / 8000
    O - -  100 / 8000
    O - O   10 / 8000
    O O X    9 / 8000
    O O -   10 / 8000
    O O O    1 / 8000

I marked with arrows, the conditional case you're looking at. So the answer is:
729 / 7220 = 10.1%

Which matches what you came up with. You're solution was more elegant. But with this technique, you can extend it to any number of rounds, do any amount of conditional probability you want, and come up with precisely the right answer.

 

[Sir Brennen]

Yes, on a 20, you stand back up with 1 hp.

With 1/4 of your full hps, actually.

 

[FireLance]

With 1/4 of your full hps, actually.

I think there's some confusion on this point. What I heard from someone who played the preview during D&D XP is that if you roll a 20, you can spend a healing surge if you have any left, and if you have not yet used your second wind during the encounter.

 

[catsclaw227]

With 1/4 of your full hps, actually.

Yea -- this is what I thought too.

Well, just as an anecdotal example, when I ran a 4e playtest last week, the paladin went into the negatives in the first encounter, and the cleric immediately used a healing word to put him back up. I believe the players were not thinking "We can relax - he's got three rounds left." They were thinking "We're outnumbered by kobolds and one of our defenders just went down? We need him back in the fight STAT!"

And this was my point. All these statistical models leave out the critical variables such as:

• Who went down?
• Where is the downed PC in relation to the rest of the beasties?
• Where is the downed PC in relation to the rest of the party?
• Is there a PC in range that can generate a healing surge?
• If there is no PC in range, can there be one within (1, 2, 3, 4) rounds?
• What are the Players attitudes about PC death?
• What level are the PCs? If 1st level, then maybe a dead PC isn't a big deal, but if they've been playing it for a year, then has that attitude about death changed?

Unless these models can reflect a wide number of variables such as these, then the question "How often will a PC die if they go down, and in what round?" can't really be answered with much accuracy.

But then again, I am not a statistician. What do I know...

 

[loisel]

By the way, I liked your solution to the problem. I came up with the same answer (see post #28), but I did it with brute force. I haven't used Markov Chains much.

I hadn't thunk of your solution. Yours is probably easier to explain than mine.

 

[cr0m]

Internet connection: $25
Computer: $600
Browsing a thread on Enworld about the death save, and finding a perl script that calculates the odds: priceless

thx stripes! made my day!

 

[Delta]

In 4e, it looks like you've got two things: You're likely to die in three rounds without help, and combat takes a lot longer than two or three rounds to resolve.

But you're not likely to die in 3 rounds without help. The chance of that is only 9%.

 

[Sir Brennen]

I think there's some confusion on this point. What I heard from someone who played the preview during D&D XP is that if you roll a 20, you can spend a healing surge if you have any left, and if you have not yet used your second wind during the encounter.

Hmmm... I'm looking at the fan created 4E Lite PDF, which is compiled from official sources, and the way it's described there is a DC 20 Heal check by someone will basically trigger your Second Wind as you describe. For rolling a 20 on the Death Save, it only mentions "you stabilize", with no clear definition of what "stabilize" actually means.

The 1/4 hit points is based on the Design and Development article, which presents the rules ported over for 3.5, so it may not be entirely the same as 4E. Still, I haven't seen anything which directly contradicts this.

 

[catsclaw227]

The 1/4 hit points is based on the Design and Development article, which presents the rules ported over for 3.5, so it may not be entirely the same as 4E. Still, I haven't seen anything which directly contradicts this.

This is a good point. I saw this in the Des&Dev article too, assumed it would be the same port into 4e.

 

[Alnag]

There is one serious flaw in the logic of letting be for 3 rounds. The fact, that enemy can still attack the helpless character during those rounds. And what than? If I would want to be nasty as a DM, I would concentrate all at least two or three foes to finish that guy so he wouldn't get a chance to enter the fight again...

No statistics for that, I guess.