So let me get this straight. What the two of you are essentially saying is that, because 18 Dex has an 83% chance of success, he's essentially guaranteed to succeed? That we can disregard the 17% chance 3 Dex has to win?
No. Statistics doesn't say it will absolutely win every time (or even ever). What you should do with statistics is always choose the side with the better number. bacause statistically, you will win more often.
Let's say we're arena fighting. 2 players are running the fighters (or maybe it's a computer doing all the rolling as a straight stand and swing fight).
Would you rather bet on Dexter or Limpy? Knowing that according to our math, Dexter should win 83% of the time. Let's assume our math is correct (because that's a whole 'nother ball of correctable wax).
While technically Limpy COULD win, I would rather bet on Dexter. Put a little more on the table, and assume you live or die based on who you pick.
Yes, it would totally suck if I picked Dexter and he lost. But we don't KNOW the outcome going in, we only know the chances. And Dexter has better chances than Limpy.
Look at those odds for a second and consider what they imply. In almost 1 out of 5 cases, the slowest (clumsiest) man in the world will be faster on the draw than the fastest (most agile) man in the world. This is in a scenario that has already ruled out the element of surprise. Does that strike you as realistic?
In real life, I have a black belt. Let's assume you do not have any combat training. Statistically, Danny will bet on my beating you. Because I have more levels than you (the virtual non-combat you). But, I recognize that in a fight, anything could happen. While I want to remain as confident in my ability as I can, technically, I might fumble, you might land a lucky hit. And the longer that fight goes on, the more chances you have to get lucky over my "sure thing"
In the real world, would you take a drug that had a 17% chance to kill you? After all, you have an 83% chance to live! Those are great odds, right?
Like cocaine or heroin? Which apparently has a pretty high addiction rate? And people apparently take quite willingly.
I can't tell you that D&D is realistic. I can tell you what the math shows. And that if you know the odds, you should do what the odds tell you and ignore the freak impossible successes.
Statistics and odds often run into people's intuition about what to do.
It's easy for a newbie poker player to ignore the odds and say his way works because he just won playing 2-7 offsuit (the worst hand in Texas Hold'em).
A pro-player knows the odds and plays them instead. He also uses that knowledge for when to fold (like folding pre-flop on 2-7).
When you win at something improbable, it reinforces your decision that you were right. That doesn't mean you were, it just means you got lucky.
When you do what the odds say and win, in a way, you didn't get lucky, you simply didn't get unlucky.
Another important thing to remember is choice as it applies to odds.
In our arena, Dexter and Limpy don't get to choose to fight. They have to. So if you could choose which one to be, you'd want to be Dexter.
But more often in life, you can choose to fight or not. So if you are Limpy, and have a choice on who to fight, you would avoid Dexter, and wait until a weaker opponent came along.
This is what happens in Texas Hold'Em. The pros fold A LOT of hands. They wait until they have good hands. At a 10 man table, they don't like to play with every man in on the flop. Because that worsens any "good" hand's chances that something weak might get lucky.
The point then, if you know your chances of victory are poor, you use that knowledge to avoid the conflict until your chances are in your favor.
It makes sense for a 1st leve fighter: you fight a goblin, run from a bugbear.
