Dark Orbit - finally a "real world" application for math

Dont worry, I did something like that. That is why I know for a fact that the average damage of a LF4 laser is 175. The range is from 150 to 200 with a uniform distribution. The distribution remains uniform even if more than one laser is used, which is quite counterintuitive from a mathematical point of view. But otherwise, you would never see the theoretical maximum damage with several lasers equipped, which would cause many complaints from players although the average damage would be the same.
By the wax, the LF3 had a range between 125 and 150 damage with an average of 137.5. The distribution was not uniform, to say the least.
The average damage per second is the average value multiplied with the hit chance. I assumed 80% here, but it does not make much of a difference for the results because both the player and the opponent have the same hit chance.

 

2. What is a good configuration
2.2 Goliath vs Goliath

The following figures hold the difference between the time it takes the opponent to shoot down the player and the time it takes the player to shoot down the opponent. If this time is positive, the player wins, which is indicated by a green background.
The last column holds the sum of all times in each row. The higher this time is, the better a configuration performs against all kinds of configurations the opponent can come up with.

We see that the defensive configurations (using a shield design for the ship and hercules drone designs) are superior to offensive configurations.
And while there appears to be a "best" configuration, the differences between this configuration and one with a slightly different number of lasers in the drones are very small. So you can still choose if you want more damage or more shield.

Terminology:
Defense: Hercules drone design, shield design for the ship, heart formation
Offense: Havoc drone design, damage design (5%) for the ship, heart formation
Slow: 15 shields in the ship
Fast: 0 shields in the ship

Defense vs Defense; slow


Defense vs Defense; fast


Defense vs Offense; slow


Defense vs Offense; fast



Offense vs Defense; slow



Offense vs Defense; fast


Offense vs Offense; slow



Offense vs Offense; fast

 

Very nice charts. Thanks again for all the effort you put into this thread

You didn't state it but I guess offense = havocs and damage design, right?

 

Exactly, thanks for the input. This information was hidden somewhere in the introduction, but it really should occur here.

 

No fast vs slow configs?

 

Sure, here you go

 

How/where did you get the data for these charts?

 

That should be in the introduction. Basically, I calculated the average dps and the total amount of HP/shield for each configuration and divided the two values for the player and the opponent.

 

Oh...WOW...these graphs are good or GODly.

 

I think they are MATHly

 

Sorry to go into the topic of GGs again, but here's my latest results (today):

Used 268 Extra energy and exactly 20k uri. Went from 44/111 parts for zeta to 74/111 parts, exactly 30 parts. Used every multiplier at x2. Does that fall into the 99% confidence interval in your GG graphs?

EDIT:I Forgot, I missed a x2 multiplier because I clicked too fast.

 

Your thread is amazing, I have always wondered about starting to make these calculations but have always failed to start with it I love math and especially probabilities! Gratz for finding the time. Gonna take a closer look at those config thingies, I wonder have my head calculations been somewhat correct

 

I made a quick table about how efficient is upgrade an item when the Discount hour event is active.
Basically, you need 3 events for a full discount.
Green: Will finish in event
Red: Will finish after
Yellow: Likely you'll never reach this point
Blue: Discount

 

None of the graphs I posted can answer this question. They are only valid for complete gates. I hope you understand that I wont run and publish simulations on request for any combination of turns performed and parts recieved from the GGG.
30 new parts from 468 turns with a gate half filled sounds reasonable though.

Thanks for the kind words. Feel free to report back when you compared your results to mine.

Thanks for your contribution. Although I cannot say I fully understand your table, this gave me some ideas for future topics to be discussed.

 

The damage range is based on total damage not damage per laser. So it would not be affect by the number of lasers used or type of lasers. Damage is total maximum damage to 75% of maximum.

 

This happens to be true for the LF4 laser, because 75% of its maximum damage of 200 is 150. This coincides with the minimum damage of a LF4.
But for the LF3, the maximum and minimum damage per laser are 150 and 125 respectively, which does not follow your 75% rule.

 

Hi, first of all very nice work with all these simulations, statistics and numbers in here
I have one more suggestion (unfortunately I didn't read through all the pages, so I'm sorry if it has already been mentioned) to add to your thread:
How does a ship with full Nano Hull change the rewards you get from the GGG. How are the original 4% probability of getting Nano Hull are split between the other possible rewards? Do they all add on top of the 13% probability for getting a gate part or are they spread equally between all other rewards?

 

This was already discussed in some of the last 136 posts
Unfortunately, it would take an enormous amount of samples (in the order of 100000 turns) to find out how the percentages change.
Lets just conclude that it is better to use a ship with full nano hull, because whatever you get in exchange is better.

 

100k turns could be done if someone buys $1k+ of uri in MHH

 

The Problem is not so much the amount of Uridium or Extra-Energy that is needed.
With the optimal strategy for the Alpha, Beta and Gamma Gates the total amount of Uridium needed to perform 100k turns is surprisingly low.
What keeps me from doing this is the time it would take to sample 100k turns.