Communities

Writing
Writing
Codidact Meta
Codidact Meta
The Great Outdoors
The Great Outdoors
Photography & Video
Photography & Video
Scientific Speculation
Scientific Speculation
Cooking
Cooking
Electrical Engineering
Electrical Engineering
Judaism
Judaism
Languages & Linguistics
Languages & Linguistics
Software Development
Software Development
Mathematics
Mathematics
Christianity
Christianity
Code Golf
Code Golf
Music
Music
Physics
Physics
Linux Systems
Linux Systems
Power Users
Power Users
Tabletop RPGs
Tabletop RPGs
Community Proposals
Community Proposals
tag:snake search within a tag
answers:0 unanswered questions
user:xxxx search by author id
score:0.5 posts with 0.5+ score
"snake oil" exact phrase
votes:4 posts with 4+ votes
created:<1w created < 1 week ago
post_type:xxxx type of post
Search help
Notifications
Mark all as read See all your notifications »
Q&A

Comments on Are 3 10% chances better than one 30% chance (when penalized by a variable for failures)?

Parent

Are 3 10% chances better than one 30% chance (when penalized by a variable for failures)?

+7
−0

I'm playing a computer game in which you can spend 100 coins to be given a 10% chance to upgrade a weapon, or spend 200 coins for a 20% chance, or 300 coins for a 30% chance and so on, up to 1000 coins for a 100% chance.

There is no penalty for failure, only that you lose the coins spent.

My theory is that if you only have 300 coins to spend, it would be better to do 3 attempts since 10% + 10% + 10% = 30% and you have the added bonus of having a chance to spend less coins if it succeeds before the 3rd try.

Not sure if my theory is correct because a 100% chance is not the same as 10 10% chances. Also after getting failures, the odds remain the same. But I don't have the logic/math knowledge to have an answer and explanation.

Note:
For the purpose of the question, let's assume the game is somehow computing true random numbers and therefore able to give true 10%-100% odds
Upgrading weapons is something done very very often. Let's say the average player wants to upgrade ~1,000 times.

History
Why does this post require attention from curators or moderators?
You might want to add some details to your flag.
Why should this post be closed?

1 comment thread

General comments (3 comments)
Post
+4
−0

Let's assume, that there are only the two options from the header: 3 × 10% or 1 × 30%. But the following calculations should be applicable to all other cases too.

Next, we have to make an assumption about how much you can lose or win in every round. Since you are playing in two different currencies (weapon level vs in-game money), this isn't that easy. For purposes of simplifying the game, I'm going to assume that a weapon upgrade is about 200 coins worth.

Now, we have to calculate the expected value of your game. The expected value tells you, how much you win on average if you play the game very often. If the game is fair, it should be 0. If it's positive: good for you. If it's negative, good for the other one. We want to find out, which game option gives you the higher EV.

You calculate the expected value by multiplying the net win of every possible outcome with the probability for that outcome and then summing all those products.

For the 1 × 30% case that's easy. You have a 70% chance of losing 300 coins (-300 coins). And you have a 30% chance of losing 300 and earning 200 coins (-100 coins).

This gives us as expected value for the 30% case:

$$ E = 0.7 * (-300) + 0.3 * (-100) = -240 $$

This means, that on average you will lose 240 coins.

Now, let's look at the 3 × 10% case. Your question contains two different possible interpretations of that case:

  • A: Once you win, you will stop playing.
  • B: You will always play two rounds.

Let's consider each interpretation. I'm gonna start with A. First we have to look, which possible outcomes there are.

----x-------# win in 1. round (0.1)
    | loss in 1. round (0.9)
    |
    ------x----------# win in 2. round (0.1)
          | loss in 2. round (0.9)
          |
          ------x-------# win in 3. round (0.1)
                | loss in 3. round (0.9)
                #

As you can see, we have four different outcomes ("#"): win in 1st, 2nd or 3rd round or loss. They have these probabilities and net wins:

var 1st 2nd 3rd loss
prob 0.1 0.9 × 0.1 = 0.09 0.9² × 0.1 = 0.081 0.9³ = 0.729
net win -100 + +200 = +100 -200 + +200 = 0 -300 + +200 = -100 -300

Based on that, we can calculate the expected value for the A interpretation:

$$ E = 0.1 * (+100) + 0.09 * (0) + 0.081 * (-100) + 0.729 * (-300) = -216.8 $$

Last, but not least, we're going to look at the interpretation B. Since you are always playing three round, you're always going to lose 300 coins. The question is, how often you win. If you win three times, you'll end up with +300 (-300 + +600), if you win two times, you'll end up with +100 (-300 + +400) and if you win one time, you'll end up with -100 (-300 + +100). If you lose every time, you'll end up with -300.

The probability to win three times is $0.1 * 0.1 * 0.1 = 0.001$. The probability to win two times is $3 * 0.9 * 0.1 * 0.1 = 0.027$. Note the 3, which is, because you have three options for choosing, which round you lose in. The probability to win one time is $3 * 0.1 * 0.9 * 0.9 = 0.243$. The 3 is here, because you have three options for choosing the one round you'll win in. The probability to lose every time is $0.9 * 0.9 * 0.9 = 0.729$.

Given that we can now calculate the EV for the last case.

$$ E = 0.001 * (+300) + 0.027 * (+100) + 0.243 * (-100) + 0.729 * (-300) = -240 $$

Well. That's interesting. The expected value for "1 × 30%" and "3 × 10%" is the same. Therefore, your choice between those doesn't really matter. However, the best strategy (of these), is to play up to three times, but stop if you win once.

History
Why does this post require attention from curators or moderators?
You might want to add some details to your flag.

1 comment thread

General comments (1 comment)
General comments
luap42‭ wrote almost 4 years ago

My notation and my names are probably a bit weird, because I learned math in a different language. If there's an ambiguos or incorrect name/notation, that's likely the fault of that and please correct (make or suggest an edit or add a comment).

Skipping 2 deleted comments.