Author Topic: Effective Armor And Percentile Armor: Definition and Implications  (Read 1063 times)

noble_resonance

  • Airman Recruit
  • *
  • Posts: 2
    • View Profile
TLDR

Most vessels in leviathans are made of paper because their own profiles negate the protection provided by their breach numbers.  Smaller ships tend to be better armored because of their smaller profiles and so swarms of destroyers are disproportionately effective.

Abstract

Profile dice and breach numbers can be combined to give a more concise idea of how hard a given vessel is to damage.  The first and easiest measure is Effective Armor.  This is calculated by subtracting the average roll for a facing's profile dice from the lowest breach number on that facing.  This is the number that your opponent must roll in order to cause a hit.  The second is Percentile Armor, or the percentage chance that your armor will provide no armor and any shot fired at the vessel will produce a hit.


Anatomy of an Attack Roll

Any attack in Leviathans is a sum of many dice against a static number.  The range of the dice are pulled from many source, namely:

  • The size of the attacking gun
  • The skill of the attacking crew
  • The profile of the defending ship
  • The breach number of the defending ship

The final equation looks like this:

Dg+Dc+(Dp1+Dp2) >= Bn

Where Dg is the result of Gun Die, Dc is the result of the crew die, Bn is the breach number, and Dp1 and Dp2 are the results of the profile dice.


This is all well and good, if unexciting.  The big problem is that over the set of all leviathans, these numbers will vary quite a bit. However, two of these numbers, the profile dice and the breach number, are specific to a single vessel.  Isolating these factors together gives us a rough idea of how this section of the attack equation will look like.  Some algebra later:

Dg+Dc >= Bn-(Dp1+Dp2)

Effective Armor

 The Bn-(Dp1+Dp2) half of the equation is what I would call the Defense Term.  This is only dependent on the vessel in question and theoretically firmly under the control of the vessel designer.  Of course, all these values are actually ranges. A decent approximation of this term can be create by simply taking the average result of the profile dice and breach numbers using them as static terms. 

As an example, the port side of the HML Evesham:

Breach Numbers:
14 - 14 - 14 - 14 - 15 - 16
Bn =~ 14.5

Profile:
Blue (d6: Avg 3.5) + Yellow(d8: Avg 4.5)
Dp1+Dp2 =~ 8

HML Evesham Effective Armor: 6.5

Compare this with the port side of the HML Charger, a smaller, more lightly armored ship

Bn =~ 12.33(3)
Dp1+Dp2 =~ 6.5

HML Charger Effective Armor: 5.833(3)

Despite having much lower breach numbers, the charger doesn't actually lose all that much protection because it's smaller profile helps to make up for it.

As a third example, let's look at the French Destroyer Pelletier, again at the port side

Bn  =~ 11.33(3)
Dp1+Dp2 =~ 4.5

Pelletier Effective Armor: 6.666(6)

This destroyer is tougher than the armored cruiser.

Percentile Armor

While Effective Armor is a good, off-the-cuff measure of how well defended a ship is, it's hardly a complete look.  Because both breach numbers and the results of the profile dice vary what we really want to know is how often our armor protects us.  We can ask ourselves how often will our armor provide no protection at all, such that any gun fired will damage us.  This event is called a Profile Breach.  The math for calculating the probability of a profile breach is:

Dg+Dc = 1 (Any die will roll at least a one)

Dg+Dc+(Dp1+Dp2) >= Bn

::Bn-1 <= Dp1+Dp2

Dp1 and Dp2, being dice, are relatively easy to calculate odds for.  I have a small program on my computer that does the brute force math of counting all possible sums of an arbitrary set of dice and determining how many of them exceed a given number.  I'm going to be using that to calculate the probabilities of the profile dice although I'm not 100% sure it is bug free.

Compensating for the fact that breach numbers vary is also relatively straightforward.  We simply calculate the odds of the profile dice breaching each breach number on a side and then calculate the probability of any breach as 6 independent events (based on the roll of the slot die).  This gives us the Percentile Armor, or the chance that our armor will simply fail to protect us against any incoming fire.

Again, the HML Evesham:

Modified Breach Numbers:
13 - 13 - 13 - 13 - 14 - 15

Profile:
Blue (d6) + Yellow(d8)

Percentile Armor - 13:  0.0625
Percentile Armor - 14:  0.02083(3)
Percentile Armor - 15:  0

(4/6) * 0.0625 + (1/6)*0.02083 + (1/6) * 0 = 0.0451 = 4.51%.

The Evesham appears to have a golden BB problem.

How about the HML Charger:

Modified Breach Numbers:
10 - 10 - 10 - 12 - 13 - 13

Profile:
Black (d12)

Percentile Armor - 10:  0.25000
Percentile Armor - 12:  0.08333
Percentile Armor - 13:  0

(3/6) * 0.25 + (1/6)*0.08333 + (2/6) * 0 = 0.1388 = 13.88%


Charger has a real golden BB problem.

If we cross over to the other side of the channel and look at the Pelletier

Modified Breach Numbers:
10 - 9 - 9 - 11 - 12 - 12

Profile:
Yellow (d8)

Percentile Armor: 0

Pelletier mocks your golden BB a second time!

Conclusion

The balance between profile and armor numbers needs to be looked at.  The increase in firepower from larger profiles more than negates the armor advantage of larger ships and turns them into floating deathtraps.  When it comes time to play, I will field fleets of the theoretical Napoleon class destroyer.  It will have a tiny profile, decent armor, and mount a single turreted 240mm cannon in it's nose, which I imagine would be roughly equivalent to the Evesham's 9.2in guns.  Short, common, ugly, and exceptionally effective.

Tonbo Karasu

  • Able Airman
  • ***
  • Posts: 305
  • Dragonfly Clan Ambassador
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #1 on: May 08, 2010, 04:21:48 AM »
I'd been vaguely thinking along similar lines for a little while, and had already produced a table of effective armours.  I've spent a little while and thrown on the Percentile Armour as well (although I've flipped your definition: 100% percentile armour protects you from yourself 100% of the time).

NameBowSideSternBowSideStern
HML Anfield77.36.7100%100%100%
HML Beagle6.56.26.2100%100%100%
HML Raven6.86.86.7100%100%100%
HML Charger5.25.84.796.7%86.1%91.7%
HML Courser5.76.54.8100%93.1%90%
HML Hertfordshire6.27.25.3100%98.6%86.7%
HML Evesham6.35.55.7100%95.5%94%
La Gloire6.57.36.3100%100%97.2%
Montpelier6.56.26.2100%100%100%
Pelletier6.36.86.3100%100%97.2%
Agile96.57100%97.2%100%
Ardent978100%97.2%100%
Pontbriand6.76.55.7100%90.3%100%
Philip II Augustus5.26.75.799.5%97.9%91.7%

Yup, I'd say there's possibly something slightly skew about this.

However, the SI values do help.  The minimum possible damage for taking an SI check is 1 slot destroyed on the current side.  This converts into a test at 2dR+2.  What's the chance of a catastrophic breakup after that sort of thing?

The Evesham needs to get 24 or more on 2dR+2.  That's impossible, 0%.

The Charger is less robust at 20, but it's still quite a small chance of 6%

The Pelletier, on the other hand, needs to get 18 or more, which comes to 15%

What about when there's a bit more damage, say 3 slots on one side (or more, but spread out)?

Here's the table below.

NameBoom 1Boom 3
HML Anfield15%45%
HML Beagle15%45%
HML Raven6%28%
HML Charger6%28%
HML Courser15%45%
HML Hertfordshire1%15%
HML Evesham0%6%
La Gloire15%45%
Montpelier15%45%
Pelletier15%45%
Agile10%36%
Ardent6%28%
Pontbriand6%28%
Philip II Augustus0%1%

So, effectively, destroyers are hard to damage but, once you start damaging them, easy to destroy.  Armoured Cruisers are easier to damage but can take a lot of punishment before disappearing.

Which is, I think, relatively representative of the era.

Also, it will make sense to defensively mask with your detroyers: not only are they more expendable, but the enemy is less likely to damage them.
Karasu



Bronzite

  • Airman Recruit
  • *
  • Posts: 10
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #2 on: May 08, 2010, 08:02:20 PM »
I think the metric we really care about is either number of hits taken vs. percentage of chance of explosion, or more likely, number of shots taken vs. percent chance of explosion. 

There's also an interesting question with heavier ships of at what point do you stop shooting at them?  My playtesting has suggested its realtively easy to disable an Armored Cruiser with a destroyer or even a fast cruiser (like HML Charger.)  Rear shots wreak havoc on a cruiser.  They hurt any ship, but the cruisers we've seen so far are vulnerable to having their MP reduced below turning speed or easily loosing all guns in their rear arcs with relatively few hits.  The initative system exacerbates this issue.  The result is that a destroyer can realistically expect to stand toe-to-toe with an armoured cruiser twice its points, and come out victorious. 

Its possible this all balances out in fleet actions.  I haven't done the math, or the playtest, but given the number of rounds it takes to sink even a small opposing vessel, smaller ships seem to be the way to go.

First Sea Lord Randall Bills

  • Administrator
  • Airman
  • *****
  • Posts: 94
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #3 on: May 21, 2010, 05:56:30 PM »
We've actually been looking at that question on our end for some time...but that's some pretty awesome math to potentially back up that we do need to do some tweaking on the numbers.

Thanks!


Randall

Alius

  • Airman
  • **
  • Posts: 116
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #4 on: June 05, 2010, 10:38:33 AM »
I think the metric we really care about is either number of hits taken vs. percentage of chance of explosion, or more likely, number of shots taken vs. percent chance of explosion. 

There's also an interesting question with heavier ships of at what point do you stop shooting at them?  My playtesting has suggested its realtively easy to disable an Armored Cruiser with a destroyer or even a fast cruiser (like HML Charger.)  Rear shots wreak havoc on a cruiser.  They hurt any ship, but the cruisers we've seen so far are vulnerable to having their MP reduced below turning speed or easily loosing all guns in their rear arcs with relatively few hits.  The initative system exacerbates this issue.  The result is that a destroyer can realistically expect to stand toe-to-toe with an armoured cruiser twice its points, and come out victorious. 

Its possible this all balances out in fleet actions.  I haven't done the math, or the playtest, but given the number of rounds it takes to sink even a small opposing vessel, smaller ships seem to be the way to go.

In my tests, it doesn't matter if there's more fleet around to choose from. If anyone is out to sink a ship and remove guns from the equation, the cruisers are the easiest casualties. Even if it is immobilized, is it really worth screening to save it if you have a more valuable vessel, like a battleship on the board? Not really.

Toolian

  • Able Airman
  • ***
  • Posts: 342
  • Your Forum Heel
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #5 on: June 07, 2010, 05:05:03 AM »
uuuuuum  :-X

2 + 2 = 4

pianobar

  • Sky Captain
  • Able Airman
  • *******
  • Posts: 203
  • Royal Statistician and FRSS
    • View Profile

pianobar

  • Sky Captain
  • Able Airman
  • *******
  • Posts: 203
  • Royal Statistician and FRSS
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #7 on: June 07, 2010, 03:11:21 PM »
On a serious note, I'll echo Randall's comments. Keep the math discussion coming if you think you've noticed something...everything is being reviewed, verified, and weighed.

Tonbo Karasu

  • Able Airman
  • ***
  • Posts: 305
  • Dragonfly Clan Ambassador
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #8 on: June 14, 2010, 06:27:14 PM »
I think the metric we really care about is either number of hits taken vs. percentage of chance of explosion, or more likely, number of shots taken vs. percent chance of explosion. 

I've done some number bashing on this and it really is quite a lot of numbers and a fair bit of Excel work.
Would there be any interest in my posting it here, or would people prefer to just get the entire spreadsheet?

Alternatively, I've just been doing it for the fun of spreadsheets...

Karasu
Karasu



pianobar

  • Sky Captain
  • Able Airman
  • *******
  • Posts: 203
  • Royal Statistician and FRSS
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #9 on: June 17, 2010, 09:30:42 AM »
I think the metric we really care about is either number of hits taken vs. percentage of chance of explosion, or more likely, number of shots taken vs. percent chance of explosion. 
I've done some number bashing on this and it really is quite a lot of numbers and a fair bit of Excel work.
Would there be any interest in my posting it here, or would people prefer to just get the entire spreadsheet?

I for one am interested, and would appreciate a look at the sheet itself if possible. However, please feel free to post a copy or summary here for community discourse.

Tonbo Karasu

  • Able Airman
  • ***
  • Posts: 305
  • Dragonfly Clan Ambassador
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #10 on: June 17, 2010, 05:22:27 PM »
Okay,

So here's what I did. 

Assumptions:
1)  The average attack is 2Dblue.  I used this as the standard attack dice, so that it was a constant.
2)  The armour value is the same for all locations.  This is the only way I managed to get the number-crunching down to manageble proportions!
3)  All attacks are against the same facing.  Again, if you start to consider attacks against different arcs, things quickly multiply up at geometric rates.

Process
I put together a worksheet with various formulae. 
The first set were working out what %age chance an attack against a specific die type had of breaching each armour level.
The second table took as input the %age from above and an SI value.  It also took into account how damaged a vessel already was. Then it works out what the chances are that a shot will cause the target to be destroyed, be damaged, or be unaffected.
The final table is an array of a number of shots down the side and the condition of the target across the top, giving a %age chance of each.
Then, after some thought, I went back and did amended versions of 2 and 3 taking into account one or two MISS slots.

With that done, I plugged numbers into the sheet and then copied the results out to somewhere else.

Below is a table with the chances various combinations are dead after some shots
SI18181818
Armour11111212
Misses0101
10000
20.20.10.10.1
310.70.60.4
43.22.21.91.3
57.95.14.42.9
615.39.98.55.6
725.516.814.59.6
837.725.522.114.8
950.435.231.121.3
1062.345.240.828.5
1172.554.850.436.2
1280.763.559.644

Want more?

Karasu
Karasu



Alius

  • Airman
  • **
  • Posts: 116
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #11 on: June 19, 2010, 01:33:37 PM »
That's pretty narrow for my tastes. Granted, dBlue is the common close range attack die for most ships. dBlue is also the front/rear go-dice on most destroyers, as well.

But, there are many more ships with different dice involved.

One thing I've noticed though, when playing around with dice combinations for other games, the number of combinations is always based on the smallest polyhedral.

So, when you're looking at a dYellow/dBlue combo (without crew) there is a wide range of numbers in around the average that all have the same chance of showing up. Makes for some really wild randomness there. (Almost as if you were rolling a single die in some cases... ;)  )

When you go to post your results, I'd also like to see the equation you used to formulate the results. (Another reason I hate spreadsheets. No way to doublecheck the results.)

Tonbo Karasu

  • Able Airman
  • ***
  • Posts: 305
  • Dragonfly Clan Ambassador
    • View Profile
Re: Effective Armor And Percentile Armor: Definition and Implications
« Reply #12 on: June 19, 2010, 05:01:26 PM »
Personally, I restricted myself to the 2dBlue because I was more interested in seeing what effect the variation of defensive factors was having on the statistics.  It shouldn't be too hard to produce a whole range of arrays for different attack/defense combos.

This is for a bit more info on the maths involved, so feel free to ignore it.

The Formula for calculating whether a shot damages, destroys or has no effect is as follows:
Damage = ((w-x)/6+(x/6)*((w-x)/6))*y = A
Destroy = x/6*x/6*y*[chance of 2dRed meeting z-2x] = B
No Effect = 100% - A - B = C
w = Number of non-miss slots
x = Number of slots damaged already.
y = %age chance that a shot will cause damage. [Calculated by simple combinational probability]
z = SI

The bit where I work out how much chance a ship has of blowing up works out as follows.
The array has 8 columns (0 - 6 damaged and destroyed) and as many rows as you want, each row being for the number of shots fired.
The 0 damage column is nice and straightforward C0 raised to the power of the number of shots.
The other columns are all worked out similarly.  Each value is a combination of two calculations: the value of the cell above times Ccolumn+ the value of the cell 1 up and 1 to the left times Acolumn-1.  This probably needs an example.  To work out the chance that a ship is going to have taken 2 slots of damage after 5 shots you do ((Chance of 2 slots after 4 shots)*(Chance of shot doing nothing to 2 slots damaged state))+((Chance of 1 slot after 4 shots)*(Chance of shot doing damage to 1 slot damaged state))
The destroyed column is what you need to fill up the row to 100%.

Having put together the formulae into a spreadsheet, all I need to do now is plug in the SI and %age chance to damage (for whatever combination you want) and the numbers pop out the other end.

I hope that made sense.  If you want any more clarification, I'll see what I can do.

Karasu
Karasu