Return of the Fallen

I have seen neither work.

On that note, can someone tell me where the resurrecting at your homeworld part is mentioned? Not only do I think it's a bad way of implementing a resurrection ability, I just can't seem to find the tooltip that says it works that way.

Well, Return of the Fallen is Free, that's going to be a big factor. And Resurrection is for 5-10 minutes, so thats enough to take part in a battle and they should be excellent sacrifice fodder for the Rebel titan.

 "I mean, for every 10 you lose, you get 1 back"

Not true. Gambler's fallacy. Just because you have ten ships does not mean any will come back. Every ship is a coin toss independent of the others. Every ship that dies has a 10% chance of returning. Those are bad odds no matter how you look at it. 

You're right, it's not in the tooltip... odd... I thought it was. It is in the entity though, if that counts: 

hudIcon "HUDICON_RESEARCHSUBJECT_RETURNOFTHEFALLEN"

modifierType "ChanceToRespawnFrigateAtHomeworldUponDeath"
baseValue 0.000000
perLevelValue 0.050000
artifactPicture ""
uniqueOverlayBrush "RebellionResearchOverlay"

Interesting, it only mentions 'frigates'... so this may be even more restrictive than we thought.

Or all ten could come back. I think he realizes that over a great many ships destroyed, 1 out of every 10 will come back.

I very much know what the odds are and how all that works. Anyone that took my statement to mean "build 10 and only 10 ships, let them die, and you're guaranteed 1 back" is an idiot. I was merely trying to contextualize the average over many trials.

In retrospect, I could have been a bit clearer with out extraneous exposition by saying "you get about 1 back"...

Statistically: Yes
Realistically: Probably not.

I agree with you, but the odds are bad.  

I said out of a great many ships destroyed. With large samples something close to that ratio will be observed, even in reality. And I've had more than a few such games with a sample that would probably be large enough.

Well that's not quite Gambler's fallacy(it's related to be sure) but technically Gambler's Fallacy is "I'm overdue for a win so I probably will win".  It's  more directly gambler's fallacy if 9 ships have already died with no revive and you're sure you'll get a revive on the tenth.  When all 10 ships are still breathing it's still a fairly safe bet to expect 1 revive(not anywhere near certain of course, it's somethng like a 65.2% chance of getting at least 1 revive in 10 ships dying)

but expecting on average about 1 in 10 ships to revive is very much proper probability.  Yes you may not get exactly that(in fact most time you won't), but the wonderful thing about probability is the more trials you run the closer on average the result will generally come to the average. 

For example suppose we have a 50 ship fleet?  With some simple probably concepts(and a dash of combinatorics) we find the odds of getting:

o revives: is about .5%

1 revive: is about 2.8%

2 revives: is about 8.8%

3 revives: is about 13.9%

4 revives: is about 18.1%

5 revives: is about 18.5%

6 revives: 15.2%

7 revives: about 10.8%

and so on.  This does not take into account the chance of revived ships reviving a second time when they are killed, which makes the calculation a great deal more work and the difference is rather marginal, so I opted for omitting this detail(these results are a fairly good approximation).

In otherwords you have about an 88% chance of getting at least 3 ships in a  50 ship fleet to revive(so you can pretty much rely on a 6% revive rate).  Moreover you have about a 64% chance of at least 5 ships reviving.  So about 2/3 of the time you will get the expected "1 ship in 10" or better....2/3 are pretty solid odds.

People often vastly underestimate the reliability of small percentages with a large number of trials.  One wonderful thing about probability is that with a sufficiently large number of trials pretty much any % chance of success will result in a total number of successes fairly reliably close to the expected value(in this case 1 ship in 10).  ALso remember that since the advent rebels are going to be reviving both friendly and enemy ships, in reality we should be running this calculation based on the combined fleet size of both combatants, in which case a mere 50 is not farfetched even for mid-game battles.

Of course with smaller fleets it will be less reliable, but in larger fleets the odds of scoring close to the average gets even higher.

The point is in late game fleet battle, a majority of the time this technology with be plenty reliable.

Sorry if this turned into a bit of a rant, I'm studying to be a Math major and the misunderstandings about probability which abound are a bit of a pet peeve of mine 

Well thats the thing, will these abilities combo with each other? We won't know until they actually work.

Interesting how I start a topic with no math or science evident, and somehow people insist on that direction 

Just the other day I was telling my friends "You're on a game show and there's 3 doors..." and they didn't believe me!

Actually, those are 2 different techs... I'm not sure if the probability of Reanimation (found the name) is greater than or equal to the Return of the Fallen tech. I would hope so, as the Reanimated ships only last a short while.

Doesn't matter if they combo(well it does but that isn't what I meant), I was just referring tot he fact you have a 10% chance for both enemy and ally units.  As I said, any of the calculations involving repeated revives is more work then I was willing to put forth at the time of posting.

I was merely saying that for the purpose of a single battle the difference between reviving 10% of your own 50 ship fleet isn't much different then having a 10% chance of reviving from both sides if both sides have 25 ship fleets.

The synergies(which is to say multiple revives for a given ship, whether it works or not) add a whole additional layer of complexity to the problem(I'd have to throw in calculus in addition to probability & combinatorics).  Which is more work then the rather small difference it makes warrants IMO.

God, I need to learn to not type a paragraph long response every time I am slightly misinterpreted 

Oh I know.  I was less arguing the value of a single technology in a  void and more simply argueing that all considered, these techs are going to have a fairly reliable impact on mid through late game fleet battles and not simply be "hit or miss" because of the "low" 10% probability.

Now I think you're misunderstanding me. I'm not commenting on the math, its just that return of the fallen may only apply to your normal ships and not the copies you acquired through reanimation.

Sold. I'll be back tomorrow.

I don't see my misunderstanding and agree with everything said in this thread.

My point was:
It is very likely you with get no ships back.

In hindsight I should have elaborated: 
It is likely you will get less then 1/10 back.
It is possible you will get 1/10 back.
It is unlikely you will get more then 1/10 back.
It is very unlikely, but possible you will get all of your ships back.

I just do not like chance when strategizing >.< 

(I was never considering Reanimation because the topic was solely on Return of the fallen.)

This got a bit off topic because of my comment, my apologizes.

We should have a Math convention. 

Methinks I must be articulating myself poorly.  That is precisely what I thought you meant.   In all honesty I hadn't even considered that possibility before you mentioned it.  I thought it was possible ships revived by return of the fallen could revive again, but I'd actually never considered that clones from reanimation might be able to resurrect on death via return of the fallen(after all they are already clones & clones of clones sound rather silly to me).

In my OP my comment about running the calculation on the combined fleet size of your opponent and yourself was merely mentioning the fact that you have a 10% chance of creating clones from enemy ships that AND you have a 10% chance of resurrecting your own ships when they die, meaning you have a 10% of ressurecting/cloning each of the ships in both player's fleet at the start of the fight(provided they die). 

As I said any sort of multiple Resurrections of a single unit(regardless of which tech grants the revive) scenario is really an unnecessary consideration for the point I was trying to make(small relative difference).  Anyway, regardless of where the misunderstanding is methinks we're starting to derail the thread, so I'm going to get back on topic.

Did you even look at my numbers?  This is nowhere near true.  With 50 ships:

"very likely you will get no shops back"- the chance of getting no ships back in a 50 ship fleet is only .5%- that's 1 in 200 such battles. Even in a  mere 10 ship fleet the chance of getting no revives is only 34%(so the chance of getting at least one revive is about twice as high).  That said, there's no point in dividing a fleet up into 10 ship segments- the simple fact of the matter is the larger a fleet gets, the more reliable this tech will become.  And by the stage of the game you're learning it 10 ship battles are less then common.

but I digress back to the the 50 ship fleet example:

"It is likely you will get less then 1/10 back"- 36% of the time you will get less then 1 in 10 back.  I suppose 36.1% is still fairly likely, but you're almost twice as like of getting more then 1 in 10 ships back

"It is possible you will get 1/10 back"- 18.5% of the time(just short of 2/3) you will get back AT LEAST 1 in 10 ships.  That's more then "possible", that's probably

"It is unlikely you will get more then 1/10 back"- 46% of the time you will get more then 1 in 10 back that's not "unlikely", that's nealy a 50%-50 chance.  You're calling the 36% chance of getting less then 1/10 revives "likely" and the 46% chance of getting more "unlikely"

All put together you have about a 64% chance of reviving 1/10 or more.  The point is you are incorrect in your assumption that you will generally get less then 1 in 10 ships back.  it's exactly the opposite, odds are you will get at least 1 in 10 back.  It is still partially random to be sure, but nowhere near as undependable as you make it sounds.

The level of discourse over what "Chance to resurrect: 10%" means over a limited sample space of ships is all said and good. The tech is, in my mind, a 10% hp boost to your cannon fodder, though we'll all be much happier to see an expensive frigate being resurrected over a scout ship.

Do you mean the Reanimation research tech that directly follows from Return of the Fallen? I thought Return of the Fallen didn't have a time limit...

Yes, I meant the Resurrection from Reanimation. The ships from Return of the Fallen are permanent.

So, wait.

Has somebody besides the OP tried to increase the probability on Return of the Fallen to 50% per level?  If modding this crashes Rebellion, that's a big bug.

Bilun, fyi, your numbers look slightly off... are using some other method than standard Bernoulli?

For 2 revives, my math says 7.79%, all the rest look right except for a rounding error on 1 revive.

Yeah, I didn't get the answer today... I'm working on it. Seems like a good opportunity to apply some of the calc I've been learning.

uh, so... was that a random number? a guess? I actually did modify it to be exactly 50% (didn't want to try 100%, because I figured that might be an issue), so if it was a guess, you win a prize:

hmm good catch.

Just reran the calculations for 2 ships and you are correct, it is indeed definitely 7.79....  Methinks it was a transcription error on my part.  I probably got 7.79 before, rounded it to 7.8 and then proceeded to type 8.8 changing both 7s to 8s instead of just the 7 being rounded .   

Or possibly I just scratched one of the inputs on the calculator.  In all honesty I ran my calculations rrather recklessly fast to get my post finished; even if I goofed up wasn't expecting anyone to call me on it  .

In anycase Bernoulli strikes me as one of those names I heard back in the day when I actually took statistics & discrete math classes and have since forgotten.

 To describe my procedure though I simply ran:

Let x denote the number of ships revived

(.1)^(x) * (.9)^(50-x) for the probability of any given x ships being the two revives, then mulitplied that by the number of distinct  x combinations which may be selected from 50 ships, which is to say (50*49*....*(50-x+1))/(x!).

So while I'm not sure on the name of the method(perhaps it is standard Bernoulli) is I basically brute forced it with general probability & combinatorics knowledge.  

Also if you're trying to find the formula for taking into account the possibility of multiple revives, my previous statement wasn't quite correct(about the extra level of complexity requiring calculus).  At least the method I see of finding said formula is actually more directly tied to differential equations, which are directly related to calculus but technically a different subject.

yeah, that's Bernoulli. I honestly wouldn't have noticed except I started putting things in my calculator just to double check I knew what the hell I was doing and got different numbers :p The fact that it was the 3rd one down really threw me for a loop, because the first I got, but was obvious, then the second has a rounding error, so I thought the 3rd would settle the matter.... Anyhow, all is good

Yes, I realized it's a differential equation, which is why I was interested in it. My trouble is that everything I found on differential equations is how to solve them, not so much on setting them up. There's a friend I have that I had been trying to get to explain it to me for another problem... He's been busy. That's kinda why it got delayed...

Anyhow, it's made me realize that my Calculus book I have is not enough... I need one on trancendentals and all that jazz. Linear algebra seems dumb in comparison...

All this talk of probability made me wonder, if the dev's can't fix the techs with probabilities, maybe the Return of the Fallen should just activate on every 10th or 20th cruiser you lose (similar activiation for reanimation) depending on research level.  It would be more predictable and we wouldn't have to talk about gamblers or fallacies.

But perhaps Sins wouldn't be Sins if you didn't have deceptive percentages and fuzzy math. 

Well to be fair the percentages only seem deceptive because players are accustomed to the other way most games present percentages.

And there certainly are advantages to presenting it the way sins does, such as:

1).  More intuitive interaction between percentage increases & decreases.  In most games a 33% reduction is equal to and consequently "cancels out" a 50% increase.  In a sins a 33% increase  is the same as a 33% decrease.  This also makes balancing buffs vs debuffs easier then in most games. 

2). In many cases sin's method of presenting percentiles actually tells you more about the benefit of the ability.  For example while a sins "25% reduction in max speed" actually only reduces max speed by 20%, that translates into "you have 25% more time to beat on the ship before it gets away" and "you increase the gap between you and this ship when fleeing 25% faster".   Likewise while 200% reduced AM costs only actually reduces the AM costs by 33% it means that you'll have enough AM to use 200% more abilities before you run out of AM.  In many such cases the percentage sins displays is actually more useful when determining the total benefit/yield of the ability then the actual % reduction of the stat.   

3).  Debuffs can be scaled up indefinitely without nasty increasing returns.  In a conventional system, once a percentages reduction gets up around 60-70%, the benefit of each additional 1% is 2.5-3x as great as it was back around 0-10%(IE when increasing a damage reduction % from 0 to 1% results is a 1% reduction in total damage.  Increasing a damage reduction from 75% to 75^ results ins a 4% reduction in total damage).  These increasing returns  easily led to a much less flexible range of variance in which debuffs could give, as any high values would start drastically out performing things even just a bit smaller by a large margin

4). This is related to #3, but the "weird" way percentages work is precisely the reason it's balanced that all effects of the same type in the game more or less stack with each other.  Stacking negative percentages can easily get out of hand when their stat effects have increasing returns.  Admittedly this last point can also be managed by making sure all separate effects stack multiplicatively.

In all fairness as well this presentation just saves the player single quick calculation, since in other games you could find this same information simply by taking the inverse the complement of the fraction/%(a single simple step).  But as far as intuitive/useful tooltips are concerned I actually think the sins presentation is a better method.  it just seems strange because we're accustomed to seeing the other presentations.

That turned into a bit of a rant, but my inner math geek has been making me want to say for that I rather like the way percentiles are handled in this game despite the predominantly negative regard most people seem to hold them in.

That was my thought as well.  Another thing I do is have a couple dozen scout frigates in my fleet.  I use them to repair my titan and to use their matyrdom ability.