Which fleet wins this battle? » Forum Post by ShuShu62

To spell it out for the newer players,

what exactly have we gained when it comes to the essence of "playing a game"?

Just a little more insight

Yes - in this case, make sure the ships you want to lose first are going to be the ones shot at.

According to the specs you provided, above, the Terran BB's are going to be the 1st targeted ... The Iconians look like they will stand up pretty good against the Terran Mediums afterwards.

The BB's ratio was just below 1 (~112/116)

(Total offense/(Total Defense + Hit Points))

while the medium ships were around 28/94, or around 1/3. If you took some of the Zero Point Armors off the mediums and replaced them with weapons to adjust the ratios, not only would you have more attack value, you would lose the smaller attack value vessels 1st, leaving the BB's to take out the Iconians and gain experience.

(If the numbers above are off a little for 1.4, bear with me - I'm still finishing off a 1.3 game - but you get the idea....)

Successfully updated karma reason!

No offense taken.

My observations were defintely geek stuff. I do think, however, that the results are very contrary to what most folks would expect given the forums genral tenor on the matter.

My motivation was actually the post on ship defenses that basically summed up the reason to use ship defenses is strictly because you don't want to micromanage swarms of tiny ship fleets. They justified their results based on extensive trials using flight sim. Since their observations were so contrary to in game actuals that I was seeing, I felt it was worth posting.

Quite frankly, I was shocked at how many people picked the Iconians. Guess I'm not as smart as I thought I was.

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

According to the specs you provided, above, the Terran BB's are going to be the 1st targeted ... The Iconians look like they will stand up pretty good against the Terran Mediums afterwards.

I didn't notice that you had predicted this. You definitely deserve credit for the best response then.

Though, I thought he meant that only the BB's had equal number of ZPA to weapons. If all ships had equal ZPA to weapons (7) then the total defensive capability of the fleet should have been at least 9*7*10=630. Since this is more than the fleet statistics show, I felt it was impossible. However if it was only the BB's that had the 7 ZPA's then this would be 210 and with a not too unreasonable defense bonus and a limited amount of defense on the mediums this would make sense, but then the mediums would have been targeted first.

Anyway, ShuShu, can you give us the precise specs of the ships involved (after the relevant weapons and defense bonuses)?

Successfully updated karma reason!

The key is the way the humans ships get thier totals divided up amongst all those ships. I should have posted that I thought so before, but wasn't sure.

My personal experience has always been a few strong ships will wipe out hords of weaker ships mostly because of defense. I personally also think stardock should do more research in this area, because for the most part the defensive large hull ship strategy is almost always the best. (except when they're ahead of you in tech, or techs are maxed out)

Having said that, with 1.4 I did run into the Thalians using defense very effectively. (forced me to build some fleets of disposable all attack ships)

Successfully updated karma reason!

SrGalen

That observation isn't just for newer players, unless your lumping me into that category.

I thought I had made a lot of observations, but that wasn't one of them!

ES

We have concluseively established that I do not have all the answers. I need to go and cry now.

Successfully updated karma reason!

Mumble,

I will check and see. I accidentally overlaid the save file I was using, but I may have another save out their somewhere. I am assuming all you need are the total Atack point, hitpoints, and defense points for each ship.

If so, I may be able redirive it from what I know.

Successfully updated karma reason!

FYI - here are some binomial probabilities.

Probability that Iconians True Superiority is
10% 99 to 1
74% 9 to 1
38% 4 to 1
15% 7 to 3
5 % 3 to 2
1 % fifty/fifty

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

I will check and see. I accidentally overlaid the save file I was using, but I may have another save out their somewhere. I am assuming all you need are the total Attack point, hitpoints, and defense points for each ship.

If so, I may be able redirive it from what I know.

Actually, attack points per category, defense points per category and hit points, all after any applicable bonuses, for both Terrans and Iconians. This is what is input to fleetsim.

It's not the end of the world if you no longer have the data and use your best guess.

FYI - here are some binomial probabilities.

Probability that Iconians True Superiority is
10% 99 to 1
74% 9 to 1
38% 4 to 1
15% 7 to 3
5 % 3 to 2
1 % fifty/fifty

From whence cometh this data and what does it mean? Is this a 74% chance that the Iconians superiority is 9 to 1 or is it 9 to 1 that the Iconians superiority is 74%. And what does a superiority of 9 to 1 or 74% (whichever is applicable) mean? And why do your answers only generate new questions (this one is rhetorical)?

They justified their results based on extensive trials using flight sim

Finally, what is flight sim?

Successfully updated karma reason!

errr, fleetsim?

10% probability that the results ocured because the real Iconian probability of winning the battle was 99%

74% probability that the results ocured because the real Iconian probability of winning the battle was 90%

etc.

There is only a 1% probability that the Terrans realy were the superior fleet and were very unlucky.

Comes from the binomial probability on where there are only two results (success p and failure q) p + q = 1. I googled binomial probabilities and it sent me to Vassar's site. That had a nice little toy to get the answers.

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

Comes from the binomial probability on where there are only two results (success p and failure q) p + q = 1.

Oh. Essentially the confidence factor that a test result of 9 out of 10 resulted from a true odds advantage and not just random luck.

errr, fleetsim?

Oh. OK.

Sorry about the twenty questions, I'm actually not doubting your methods or results, I just want to understand them. It is interesting that the answer that I got from fleetsim and reality was so different. What I did was to "assume" 9 identical Terran ships each with 1/9 of the fleet total hitpoints, attack and defense. In the "balanced" case this shouldn't be a bad assumption but the points you've already made about that are most certainly valid.

I guess that there would be a real "tipping point" at the point where the BB's become the first targets. The more unbalanced the ships the bigger this effect. Above this point by the smallest possible margins, the Terrans are the overwhelming favorites, while at the smallest possible amount below this, the Iconians become the overwhelming favorites.

I would suspect that if all the terran medium ships had 1 hit point, no attack and no defense, but if for some reason were targeted first by the Iconians, then the Terrans would be the overwhelming favorites. Perhaps, what you've proved is that it's good to have "fodder" ships around and that you only should spend enough on them to ensure they are targeted first. Certainly, you should never put defense on them. This was SrGalen's point

If what you were trying to prove was that fleets of a few dreadnoughts can beat fleets of higher ship counts consisting of a distribution of sizes, then you've done that, but only as long as the AI makes the mistake of building it's smaller ships such that it's large ships are targeted first. I think that's a bad assumption, but you've seen what can come from my assumptions.

Successfully updated karma reason!

I am actually curious about fleet sim, because I think the average ship terrans would be easier not harder to beat so it I am curious to know if fleet sim is calibrated correctly.

Assuming it is, then I think the IMPORTANCE of targeting sequence is a surprise to many of us.

Off___Def_____Hit points
G128__G112____hp46
G112__G98_____hp46 (2)
G28B6_G24_____hp2
G28___G24_____hp24 (4)
G21___G17_____hp15

M53___B12_M0__G13______hp163
M49___B12_M13_G27______hp70
M49___B12_M13_G27______hp95
M49___B24_M4__G13______hp135

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

Interesting, I input this data into fleetsim and get the following results which are virtually identical to my assumed 9 equal ships. This says the Terrans win with 4 or 5 ships lost 89.2% of the time.

Just to be sure of your nomenclature

M49___B24_M4__G13______hp135

means 49 missile attack, 24 beam attack, 4 missile defense, 13 gun defense, 135 hitpoints.

This seems to imply a definite problem with fleetsim. I've previously tried swapping the attacker and defender and it makes no difference.

Average Losses: 4.35

Lost 7 ships: 0.3%
Lost 6 ships: 4.5%
Lost 5 ships: 30.7%
Lost 4 ships: 58.5%
Lost 3 ships: 5.9%
Lost 2 ships: 0.1%

As a test I resimulated with no defense on the medium terran ships. I think everyone would expect that this should certainly cause the mediums to be targeted first and cause a different result. But the result was virtually identical.

I suspect that fleetsim is not correctly targeting the BB's as is done in the game.

Successfully updated karma reason!

Targeting is working properly at first glance. I'll have to examine it more deeply tomorrow. Here's a dump of the data for this battle though, and you can see that the battleships are calculated higher than any other terran ship aside from the one with 2HP.

Successfully updated karma reason!

Hi!

I still think Iztok's comment that there wasn't sufficient information to truly answer the question turned out to be the best response.

That's exacly the answer I had in mind.

When presented with the problem I started modelling attack&defenses, but quickly realized that I HAD TO ASSUME lots of things. Esp. the unknown human ships were the biggest obstacle, so I became upset, and dropped the problem first. Then I realized I HAD to point to the main problem and posted the "not enough data".

BR, Iztok

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

the IMPORTANCE of targeting sequence is a surprise to many of us.

I must admit it *was* a surprise to me even though it shouldn't have been. Perhaps it's because I only ever build fleets of identical ships, in which case which ship is targeted first makes no difference whatsoever.

In my games there are really only two types of fighting ships that I build. When I'm behind in technology or in production capacity I use swarms of fighters. I try to max my logistics and wear down my opponent. This is my desperation fall back position. I use it whenever I'm in trouble. I find it has the best chance of success against whatever I'm up against.

I only go to the huge hull massive defense massive offense once I get to the point of technological or manufacturing superiority. I've said this before, but I've always felt that fleets of a small number of huge hull ships are at a disadvantage against the more distributed fleet. But it's certainly less micromanagement to deal with so I look at using these types of fleets as a convenience that I can afford if I otherwise have the advantage in technological and/or manufacturing capacity.

Though this experiment is shedding doubt on the wisdom of this strategy, every in game battle that I’ve ever had certainly seems to support this. Since I have yet to play a v1.4 game, is there any chance that there’s been some change to the targeting algorithm or anything else related to fleet battles that could be the cause?

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

ShuShu, a slight off topic digression. This thread has been an unqualified success. I've no idea when I last saw an actual game related thread go above the 50 post mark. Usually, it's only the political topics or one of the dreaded three topics of death (carriers, tactical combat and multiplayer, oh my) that can occupy the collective consciousness of the community for so long. Congratulations and thanks.

Kryo, I don’t expect a response, but it appears you really do have to read every stupid little post that goes by here. Talk about being a fly on the wall. I’m sure some of the drivel that goes by must make you want to scream. I hope they pay you well.

Successfully updated karma reason!

M49___B24_M4__G13______hp135

means 49 missile attack, 24 beam attack, 4 missile defense, 13 gun defense, 135 hitpoints

The Iconians don't have any beam attack. must be 24 beam defense.
These values are correct in Kryos post.

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

OK. I'll try this again. Here's what I entered into fleetsim.

Attacking Fleet
A9. 15/15 HP 0-0-21 0-0-17
A8. 24/24 HP 0-0-28 0-0-24
A7. 24/24 HP 0-0-28 0-0-24
A6. 24/24 HP 0-0-28 0-0-24
A5. 24/24 HP 0-0-28 0-0-24
A4. 2/2 HP 6-0-28 0-0-24
A3. 46/46 HP 0-0-112 0-0-98
A2. 46/46 HP 0-0-112 0-0-98
A1. 46/46 HP 0-0-128 0-0-112

Defending Fleet
D4. 135/135 HP 0-49-0 24-4-13
D3. 95/95 HP 0-49-0 12-13-27
D2. 70/70 HP 0-49-0 12-13-27
D1. 163/163 HP 0-53-0 12-0-13

And these are the results.

Simulation Complete: Attacker Wins!

Average Losses: 4.48

Lost 8 ships: 0.1%
Lost 7 ships: 2%
Lost 6 ships: 11.5%
Lost 5 ships: 30.6%
Lost 4 ships: 44.2%
Lost 3 ships: 11%

A little bit different towards favoring the Iconians but not by much.

And this is the data.

Terrans

[priority] => 1.3076923076923
[number] => 4
[hp] => 2
[maxhp] => 2
[starthp] => 2
[beam] => 6
[missile] => 0
[mass] => 28
[shield] => 0
[point] => 0
[armor] => 24
[dead] =>

[priority] => 0.81012658227848
[number] => 1
[hp] => 46
[maxhp] => 46
[starthp] => 46
[beam] => 0
[missile] => 0
[mass] => 128
[shield] => 0
[point] => 0
[armor] => 112
[dead] =>

[priority] => 0.77777777777778
[number] => 3
[hp] => 46
[maxhp] => 46
[starthp] => 46
[beam] => 0
[missile] => 0
[mass] => 112
[shield] => 0
[point] => 0
[armor] => 98
[dead] =>

[priority] => 0.77777777777778
[number] => 2
[hp] => 46
[maxhp] => 46
[starthp] => 46
[beam] => 0
[missile] => 0
[mass] => 112
[shield] => 0
[point] => 0
[armor] => 98
[dead] =>

[priority] => 0.65625
[number] => 9
[hp] => 15
[maxhp] => 15
[starthp] => 15
[beam] => 0
[missile] => 0
[mass] => 21
[shield] => 0
[point] => 0
[armor] => 17
[dead] =>

[priority] => 0.58333333333333
[number] => 8
[hp] => 24
[maxhp] => 24
[starthp] => 24
[beam] => 0
[missile] => 0
[mass] => 28
[shield] => 0
[point] => 0
[armor] => 24
[dead] =>

[priority] => 0.58333333333333
[number] => 7
[hp] => 24
[maxhp] => 24
[starthp] => 24
[beam] => 0
[missile] => 0
[mass] => 28
[shield] => 0
[point] => 0
[armor] => 24
[dead] =>

[priority] => 0.58333333333333
[number] => 6
[hp] => 24
[maxhp] => 24
[starthp] => 24
[beam] => 0
[missile] => 0
[mass] => 28
[shield] => 0
[point] => 0
[armor] => 24
[dead] =>

[priority] => 0.58333333333333
[number] => 5
[hp] => 24
[maxhp] => 24
[starthp] => 24
[beam] => 0
[missile] => 0
[mass] => 28
[shield] => 0
[point] => 0
[armor] => 24
[dead] =>

Iconians

[priority] => 0.4016393442623
[number] => 2
[hp] => 70
[maxhp] => 70
[starthp] => 70
[beam] => 0
[missile] => 49
[mass] => 0
[shield] => 12
[point] => 13
[armor] => 27
[dead] =>

[priority] => 0.33333333333333
[number] => 3
[hp] => 95
[maxhp] => 95
[starthp] => 95
[beam] => 0
[missile] => 49
[mass] => 0
[shield] => 12
[point] => 13
[armor] => 27
[dead] =>

[priority] => 0.28191489361702
[number] => 1
[hp] => 163
[maxhp] => 163
[starthp] => 163
[beam] => 0
[missile] => 53
[mass] => 0
[shield] => 12
[point] => 0
[armor] => 13
[dead] =>

[priority] => 0.27840909090909
[number] => 4
[hp] => 135
[maxhp] => 135
[starthp] => 135
[beam] => 0
[missile] => 49
[mass] => 0
[shield] => 24
[point] => 4
[armor] => 13
[dead] =>

ShuShu, in what version of the game did you run your test? I can't help but think this might be some difference in v1.4 assuming, of course, v1.4 was the version you used.

Successfully updated karma reason!

Just want to point out, that nobody changed there stance after this post, so I think the ..."Well I didn't realize the fleets weren't balanced statements" are really, "I didn't realize that unbalanced fleets maded a difference until I saw the results

Personally i only read the initial post before making my predictions and only scanned the rest of the posts after that -- there were quite a few before i noticed this thread.

However obviously the exact stats for each ship are needed to make an accurate prediction.

Also, was I wrong in thinking the Terrans are the attacking force? I based that thought on the fact that the terrans showed 0 movement left and the iconians showed full movement left which made me believe the terrans attacked on their last movement for the turn.

I know you were trying to make a point out oif all of this too. Funny thing is that if either fleet was built the way I would have built it with the MP used, It would have won every time.

Successfully updated karma reason!

Hi!
I still think Iztok's comment that there wasn't sufficient information to truly answer the question turned out to be the best response.

That's exacly the answer I had in mind. When presented with the problem I started modelling attack&defenses, but quickly realized that I HAD TO ASSUME lots of things. Esp. the unknown human ships were the biggest obstacle, so I became upset, and dropped the problem first. Then I realized I HAD to point to the main problem and posted the "not enough data".

BR, Iztok

I graciousily tip my hat to you.

Successfully updated karma reason!

Mumblefratz

Join Date 07/2006

+139

I still think Iztok's comment that there wasn't sufficient information to truly answer the question turned out to be the best response.

Actually, I had said that prior to noticing that SrGalen had specifically pointed out that the Terran BB's would be the first targeted and because of this picked the Iconians. He probably deserves the top credit for noticing this crucial fact. I know I didn't notice it.

According to the specs you provided, above, the Terran BB's are going to be the 1st targeted ... The Iconians look like they will stand up pretty good against the Terran Mediums afterwards.

Iztok's response, that there was insufficient data to prove either case, was probably the most scientifically correct one.

But I don't think that it makes much sense to make too big a fuss over who "won" or "lost". I know that ShuShu's not a real big fan of the winner/loser thing either. I'd say the same thing even if my guess had been correct.

The winner is the one that learns something that helps them improve their game, regardless of their "guess". The loser is the one so obsessed with winning that he learns nothing when he inevitably does lose.

Successfully updated karma reason!

ShuShu, if you can replicate the battle again (or a similar one) in-game, can you record all of the hits and damage dealt? I'm curious if luck/critical hits had any significant effect, and also about the distribution of damage rolls, as the fleetsim doesn't take either of those into account.

Successfully updated karma reason!

Mumble

The Iconians are the attackers. They are the ones with enough movement points to initiate the attack.

This was 1.4 but I sw the exact same behavior in 1.3.

Kryo,

I didn't see how your fleetsim accounts for ship level. There is a thread out there that established that ship level increases more than just hit points.

Unfortunately, I cannot reccreate the battle. I was able to post the values from some notes I had lying around. I crossed checked all the values and the totals all match so I trusted my notes.

This is the best I can do for a description.

The Iconians win.Except for the one battle they lost, they had over 30 hp left. Twice, they came away with TWO ships. Quite frankly, I was suprised at the one loss. The terrans were very lucky and got two Iconians in the first Round.

The other battles all follow the Same pattern. The terrans take out the 70 Hp Iconian and damage the second. The Terrans lose their top battleship and suffer great damage to the second in the first round. Both sides lose a ship in the second round. The terrans lose their last battleship in the third round. At that point, it is a mop up operation as the medium ships do little or no damage. The third Iconian ship only dies if the third terran battleship has taken out 3/4ths of the hitpoints before it goes

Statistically speaking, the odds that the Terrans had a 55-50 chance is 1 in 100. The odds that your results and mine are because of chance has got to be more than 1 in a million.

All,
I really did post this thread because I wanted to post the fleet evolution thread, and was afraid to do it without any empirical evidence that the scenarios I outlined were even possible, much less playing out the way I claimed. And since I'm an instigator, I couldn't just post the answers could I?

I just got lucky that the fleet compositions split the community right down the middle. I originally thought everyone would choose Terran, and worried that would kill the thread. Once the Iconian backers showed up, I felt the challenge could roll along without me tipping my hand.... too much at least.

Successfully updated karma reason!

Good Job I enjoyed seeing how it turned out, and some interesting discussion afterwards.

Successfully updated karma reason!

There is a thread out there that established that ship level increases more than just hit points.

As I recall there were many people positing that, but when someone finally obtained objective results, the numbers showed that rolls covered the same ranges and fequencies regardless of experience on otherwise identical ships. Levels only affect HP.

Successfully updated karma reason!