That would have to be the ONLY way to get more than 1 HP on a cargo hull: I've built "raiders" in the early stage of a game before using a cargo hull. I then attacked troopships, freighters and small/weak enemy ships with them. This was back when the attacker got to shoot first and if you got the other guy with your first shot he never got to shoot back. Even after gaining substantial experience, these ships never got more than one HP,

As stated earlier in the thread, ships have to take damage before gaining levels. So with a 1 hp ship it's an immediate nonstarter!

I may have to correct what I said earlier although it could be the levels do in fact increase weapons damage from exp upon upgrading, I forgot I had a planet that had a 49 percent bonus to starships wich could be the real reason it had a better attack even though it had the same components.

Hi!

With experienced ships I seem to get 0 damage more often than I get 0 damage on inexperienced ships.

Because experienced ships with more hit points are LESS attractive than rookie ships, so rookies are usually targeted first. You can test that: merge a rookie ship to a fleet of experienced ships with more hit points. Send that fleet in battle and watch where the opponent's fire is going.

Ships got more firepower with experience in GalCiv-1. In GC-2 they got hit points.

BR, Iztok

Because experienced ships with more hit points are LESS attractive than rookie ships, so rookies are usually targeted first. You can test that: merge a rookie ship to a fleet of experienced ships with more hit points. Send that fleet in battle and watch where the opponent's fire is going.

Yes, I know this. When I say that experienced ships seem to get less damage than inexperienced ships it's based on fleets consisting of only experienced ships compared to fleets consisting of only inexperienced ships, both versus equal opponent fleets. In my case I'm also sure there is no starship bonus involved.

I also think there's a comparable difference in the attack value of experienced ships versus inexperienced ships. However, the attack value differences are more difficult to measure because reported damage is capped at the hit points of the ship being attacked.

Alright here my origanal game become corrupted- So I started over- This time with 3 ships from the same planet-- 2 ships where fresh rookies, 1 was a level 7- dreadnaught. Maso Level (if it matters)

ship base configs - Huge hull, 75 base hitpoints- Weapons: 75 guns- Defense: 316 point/28 armor.

I gifted 1 rookie ship to the Altarians, the other rookie and the lvl 7 then where used to get the numbers. My defensive mines upped my ships numbers to:

attack: 174 gun
Defense: 624 point defense/56 armor

Rookie ship hitpoints: 75
Lvl 7 ship hitpoints: 100



After 15 fights against the ships clones- the score stood at this -

Rookie - 28 rounds/1125 damage inflicted/272 damage recieved
which to me means the rookie dished out about 40.2 points a round while recieving 9.4 points of damage a round.

LVL 7 Ship - 27 rounds/1125 Damage inflicted/202 damage recieved
So - the level 7 ship dished out 41.6 points a round while recieving 7.5 points of damage a round.

Neither ship was defeated. So from what I can gather a lvl 7 ship, has a slight defensive roll bonus over its fresh rookie counterpart

Going to start a new game- work up a handful of high level ships, (my goal is to get a group of the same ships from the same planet between levels 20+ and 5+, and do this again...) This time thou I'm doing small map (cut down my load time), abundant, normal- Just in case the level has anything to do with ship rolls.

After 15 fights against the ships clones- the score stood at this - Rookie - 28 rounds/1125 damage inflicted/272 damage recieved which to me means the rookie dished out about 40.2 points a round while recieving 9.4 points of damage a round. LVL 7 Ship - 27 rounds/1125 Damage inflicted/202 damage recieved So - the level 7 ship dished out 41.6 points a round while recieving 7.5 points of damage a round.

This data is probably sufficient except you need to provide each sample point, not just the total. With each sample point you can calculate the mean and standard deviation and from these get pretty much everything else. But you need the individual data points. For example, a damage of 10, 10, 10, 10, 10 and 10 has the same mean and total as damage of 5, 15, 0, 20, 0 and 20 but have very different standard deviations.

Actually from the data as you provided it, the fact that the damage inflicted on the rookie ship is about 30% more than the damage inflicted on the experienced ship, is pretty significant. The attack values also show a significant advantage for the experienced ship. The attack values of the experienced ship is limited to the 75 hit points of the rookie ship whereas the attack value of the rookie ship is limited to the 100 hit points of the experienced ship. Even with the 33% higher cap on rookie ship attack values, the experienced ships managed to equal the total damage inflicted by the rookie ships. Overall this is not very rigorous, but I would conclude that besides having more hit points the experienced ships are about 30% more effective on both offense and defense.

Plus all this is with level 7 ships. I often get ships in the level 20 range.

a damage of 10, 10, 10, 10, 10 and 10 has the same mean and total as damage of 5, 15, 0, 20, 0 and 20 but have very different standard deviations.

Oops... I just recorded total rounds and total damage, etc. I'll correct this with the next test... Thanks for all the help Mumblefratz, you are a regular wealth of knowledge-

Oops... I just recorded total rounds and total damage, etc. I'll correct this with the next test... Thanks for all the help Mumblefratz, you are a regular wealth of knowledge-

I don't know as if you really need another test. As I said your data pretty much shows a 30% offense and defense advantage for the level 7 ship. This is by no means a mathematical "proof" but it's good enough for me.

As they say on TV. This myth is confirmed.

Doing another test and giving the individual sample results would allow such a proof, if anyone feels it's necessary. If you do, one thing would be to make sure that both the rookie and experienced ships attack value was less than the hit points of either ship. This would eliminate the artifical cap of attack values that otherwise skews the results. You would also not want to count the final attack that kills a ship as part of your sample, because there's no way to tell if it gave 5 points of damage because that's the roll it got, or if it got 5 points of damage because that's all that was left to kill the ship.

I don't think this data proves anything much at all. 27 sample points is not a lot. Also, having such heavy defences relative to attack is going to skew things pretty badly...7.5 to 9.4 *sounds* like a large difference, but the *real* relative difference in rolls will probably be much lower

I'd actually say your data supports there being no significant difference between levels!

I'd actually say your data supports there being no significant difference between levels!

There's a quote, don't remember all of it but goes something like "There's lies, damned lies, and statistics".

Anyway, with the individual sample points, someone that knows what they're doing can calculate the mean, standard deviation and confidence factor and be able to make a judgement based on something other than opinion.

[edit]

I would like to get a definitive answer for this just because I'm interested in how things work, but in some sense it only really matters if you would do things differently dependent on the answer.

Clearly, experienced ships are valuable, if experienced ships fight and defend better as well as have more hit points, I don't know as if there would be anything I would do differently, I already try to protect and repair experienced ships as much as possible.

It is my subjective opinion is that experienced ships fight and defend better than inexperienced ships. In order for me to give up this opinion I'd need definitive data that "proved" me wrong.

[/edit]

Just post the numbers and I'll run a test with my statistics software (SPSS).
What I'll need looks something like this:

Rookieship: Advancedship:
Attack Damage Attack Damage
1. x1 x1 x1 x1
2. x2 x2 x2 x2
3. x3 x3 x3 x3
4. .. .. .. ..
n. xn xn xn xn

Where all the x's are numbers. From this table the total number of samples can be determined, the averages for each of the four variables AND their standard deviations (measure of variation). From this I can run a T-test, thus comparing the averages of Attack and Defense between the 2 ships.

There will always be some difference between the averages. This test will say what the chance is that the difference is coincidence (p-value). If this chance is smaller than 5%, scientists claim that it's not coincidence and that there is a SIGNIFICANT DIFFERENCE. The test will also give confidence intervals.

It would be nice to keep all the values lower than certain 'caps' because for a T-test to give valid results the pattern of the values need to be 'bell shaped' (normal distribution). This means basically that a lot of the values are around the averages with fewer values extemely low or high.

If you managed to read through all this statistics jargon I congratulate you. Statistics is a very slippery subject. I kinda like it, and it would be fun to do some serious science on GalcivII, just for the fun of it.

EDIT:
hmmm, the table looks a bit screwed up, I hope you get the point though. 4 columns of numbers, one for each of four variables: RookieAttack, RookieDefense, VeteranAttack and VeteranDefense.

Alright SideMancer- I can do that. (Still absorbing how you want the numbers - if I have a problem I'll im you if thats all right)
So if I gather this correctly you need me to:

1) Build ships that do not have overwhelming attack values
2) Ditch the kill attack.

Question:

1) If I ditch the kill attack, do you want me to leave the last shot fired from the enemy ship, unless it depletes the ships hitpoints to 1?

2) How many tests would you like? And against how many diffrent ships?

I would like to get a definitive answer for this just because I'm interested in how things work, but in some sense it only really matters if you would do things differently dependent on the answer.

I'm with you Mumblefratz, understanding game mechanics is really the key to taking ones game to the next level. And if the answer turns out to be "Higher Level Ships are more effective", then I am going to do something different. Right now to me, I really don't care about more hitpoints. I use my huge ships to clear everything out, and only throw my small hull ships in front when I need something to take the damge that I can't avoid. This strat really doesn't build up to many high level ships, since the only ships that get built up are my huge and large ones. (And I'll throw those away too if my production side out strips my opponent(s) by 4 to 1 or more) or if I have a Mumblefratz eco and I can just buy 50 a turn...

I'd actually say your data supports there being no significant difference between levels!

Well, Random50- if you can think of anyway to run the tests that would make you happy let me know. Anyway, I'm off to finish up that game so I'll have a good test bed for these tests

if I have a problem I'll im you if thats all right)

Sure, no problem

1) Build ships that do not have overwhelming attack values 2) Ditch the kill attack.

Yes

1) If I ditch the kill attack, do you want me to leave the last shot fired from the enemy ship, unless it depletes the ships hitpoints to 1?

Just leave out the last enemy shot will be fine.

2) How many tests would you like? And against how many diffrent ships?

I imagine you could build ships with low attack and big defense so that they actually do win the battle but there's no cap effect. I don't think the damage taken has a cap problem, but correct me if I'm wrong. A straightforward test would have 2 identical ships, one is a veteran (say level 8) and one is a rookie. Then in the game you manoeuvre the 2 ships in striking distance of an enemy ship and save the game.

Then you attack the ship with one of your ships, note the numbers from the fight. If you don't have 15 samples from one fight, you load the game and run the same fight, taking more samples. Repeat this for each ship. Then post the numbers (in four columns/variables: RookieDamInflicted, RookieDamReceived, VeteranDamInflicted and VeteranDamReceived with 15 numbers in each column) and I can run the test.

The EXACT number of samples doesn't matter much, about 15 per variable should be enough. It could be that there is a difference between the ships, but the difference is very small. In that case we would need some more samples or a higher level veteran (larger effect).

1) If I ditch the kill attack, do you want me to leave the last shot fired from the enemy ship, unless it depletes the ships hitpoints to 1?

Any shot that kills a ship (in either direction) needs to be excluded because it's value would be capped by the ships remaining hit points. For this same reason it would be good if the hit points of both types of ship were large (as reasonable) with respect to the ships attack value.

2) How many tests would you like? And against how many diffrent ships?

I think keeping the test as simple as possible is best. I'd limit the ships to only one weapon and defense type, it doesn't matter which. Also, I'd guess that a defense level half that of the attack level would give meaningfull results. I'd go with an experienced and inexperienced ship of a single type. It might also help if the experienced ships level were a bit higher say around level 10, so that there's potentially a more noticable difference between the ships.

if the answer turns out to be "Higher Level Ships are more effective", then I am going to do something different.

This is actually a whole seperate topic but I use the repair via upgrade method. I pretty much use fleets of a single ship type. I generally try to put at least one rookie ship in a fleet of experienced ships. This makes the rookie the target and protects my experienced ships. Also, I make a copy of my current ship by selecting it in the shipyard for upgrade and saving it to a new name. For example, Fighter I to Fighter IA. Now I have two identical ship types. At the end of every turn I remove any ships with damage from my fleet and upgrade it to the other ship type i.e. from Fighter I to Fighter IA or vice versa. Because the ships are identical the "upgrade" is inexpensive. The side effect of the upgrade is that all hit points are repaired. I keep the ships being upgraded in the same parsec as my fleet so that they're protected during the upgrade process. As long as I'm in my own influence area or not very far from it the upgrade takes a single week and the repaired ships are ready for action next turn. Using this method my average experience level is around 20 with some ships close to level 40. It's really nice to get a fleet of medium hull ships with over 100 hit points each.

I'd go with an experienced and inexperienced ship of a single type.

This is very important, the ships must be the same type, the only difference must be their experience. I agree also with Mumblefratz's other suggestions.

Alright, the test bed game is ready:

Ship type: Small Hull Class: CDD Nova M4
Base Hitpoints: 11
Attack: 8 gun
Defense: 4 Armor


Rookie Ship: LVL 0 XP 0 HP 11
Vetern Ship: LVL 40 XP 243 HP 34

Target Ship: Medium Hull Yor Frigate
Hitpoints: 16
Attack: 4 Missle
Defense: 4 Shields

Game Type: Level Normal/Map size Small/all abundant

I will record all combat totals except the last 'kill' round
I will run these samples off the same saved game, load game, kill target record results. Repeat
Then I'll post the results back here and let you have fun...

Exellent, thanks for making the effort!

However, it is now 19:30 in the Netherlands, and I can't access my statistics program at work from my home computer, so I won't be able to compute any results until tomorrow afternoon.

(actually, with this few samples you could calculate the results by hand, but the computer program gives additional info and graphs I would like to see as well).

Vetern Ship: LVL 40

Level 40?! That's a lot better than level 8!

Sidemancer how I understood you wanted the data was 1. Damage Inflicted Damage Recieved so this combat sequence, Rookie Inflicted 4 points damage, Rookie recieved 1 point Damage would look like this 01- 4 1
Which is what I did- Hope I didn't screw it up...

lvl 40 ship

01- 7 3 2 1 5 0
02- 2 1 5 3 6 1 0 1
03- 5 1 3 4 4 2
04- 3 3 5 3 2 2 4 2
05- 2 1 8 1 3 0 0 1 3 2
06- 0 3 6 2 5 3 5 1
07- 4 1 3 1 2 2 5 3 0 2
08- 0 1 1 1 6 4 5 0 2 0 1 3 0 2 0 1 0 1 0 2 1 2
09- 6 2 8 3
10- 5 1 2 3 4 0 4 2
11- 3 0 7 2 5 1
12- 5 2 2 0 1 3 2 2 5 0 1 1
13- 2 0 1 2 2 1 2 2 8 2 0 2 1 2
14- 1 0 4 3 6 2 3 1
15- 6 2 1 3 5 0 3 0 1 2

lvl 0 Ship

01- 0 1 4 0 0 2 5 0 6 2
02- 6 0 0 1 3 0 3 1 4 2
03- 3 3 6 2 2 1 1 1 1 0
04- 4 1 7 2 1 1
05- 3 2 5 2 0 4 5 2 (Rookie Destroyed)
06- 1 1 1 1 8 3 5 2 1 0
07- 4 0 3 0 3 1 5 3 0 4
08- 3 3 3 4 3 2 3 1 4 0 (Rookie Destroyed)
09- 0 2 1 1 2 2 3 0 2 1 2 3 3 1 (Rookie Destroyed)
10- 6 2 1 2 3 0 1 2
11- 3 2 0 3 1 1 5 1 7 3 (Rookie Destroyed)
12- 3 3 5 4 2 1
13- 6 0 2 1 5 2
14- 5 2 6 3 3 4 (Rookie Destroyed)
15- 5 0 6 3 3 3

Also would like to thank Mumblefratz and Sidemancer for all there help

lvl 40 ship 01- 7 3 2 1 5 0

Not sure I'm reading this correctly. This means in round 1 the level 40 ship, gave 7 damage and rcvd 3 damage, gave 2 rcvd 1, gave 5 rcvd 0?

lvl 0 Ship 01- 0 1 4 0 0 2 5 0 6 2

Then this would mean in round 1 the level 0 ship, gave 0 damage and rcvd 1 damage, gave 4 rcvd 0, gave 0 rcvd 2, gave 5 rcvd 0, gave 6 rcvd 2.

What I'm unclear about is that the damage given by one ship should be the same as the damage received by the other so these sequences should be reciprocals of one another. Unless something else is going on here.

Okay- just figured out exactly what Sidemancer wanted- (Thanks for the email clearing it up Sidemancer) Hope this helps Mumblefratz - I'm getting me a statistical education

Damage Infilicted BY ships

lvl 40 ship 70 rounds

1- 7 2 5 2 5 6 0 5 3 4 3 5 2 4 2 8 3 0 3 0 6 5 5 4 3 2 5 0 0 1 6 5 2 1 0 0 0 0 1 6 8 5 2 4 4 3 7 5 5 2 1 2 5 1 2 1 2 2 8 0 1 1 4 6 3 6 1 5 3 1

Lvl 0 Ship 65 rounds

2- 0 4 0 5 6 6 0 3 3 4 3 6 2 1 1 4 7 1 3 5 0 5 1 1 8 5 1 4 3 3 5 0 3 3 3 3 4 0 1 2 3 2 2 3 6 1 3 1 3 0 1 5 7 3 5 2 6 2 5 5 6 3 5 6 3

Damage Recieved by ships
lvl 40 ship 70 rounds

3- 3 1 0 1 3 1 1 1 4 2 3 3 2 2 1 1 0 1 2 3 2 3 1 1 1 2 3 2 1 1 4 0 0 3 2 1 1 2 2 2 3 1 3 0 2 0 2 1 2 0 3 2 0 1 0 2 1 2 2 2 2 0 3 2 1 2 3 0 0 2

lvl 0 Ship 65 rounds

4- 1 0 2 0 2 0 1 0 1 2 3 2 1 1 0 1 2 1 2 2 4 2 1 1 3 2 0 0 0 1 3 4 3 4 2 1 0 2 1 2 0 1 3 1 2 2 0 2 2 3 1 1 3 3 4 1 0 1 2 2 3 4 0 3 3

I've done some math on it, but I think I'll let Sidemancer or Mumblefratz run the numbers on it before I comment

Bit late unfortunately, but my suggestion would be to test the damage inflicted part against targets with *zero* defence. That way you know *precisely* what the damage roll was for every attack. "Zero" damage inflicted actually involves quite a wide possible range of rolls from the defender and attacker.

Mumblefez,

The two test ships are attacking a control ship, they are not attacking each other. If they attack each other, not only do you have random50's issue (is it the defence or the offence) but you have the ship issue (was ith the Experienced offense or the rookie defense)

Much better to run the two test samples against a control.

Bit late unfortunately, but my suggestion would be to test the damage inflicted part against targets with *zero* defence. That way you know *precisely* what the damage roll was for every attack. "Zero" damage inflicted actually involves quite a wide possible range of rolls from the defender and attacker.

This is correct and I did think about it a bit. The point being made is that you're testing two random variables at the same time, one is the attack value of the attacker, the other is the defense value of the defender. This test is really measuring a single random value which is the attacker's attack value minus the defender's defense value. From my rememberance of statistics this is a common occurance. You can't necessarily intuit what are the unique random variables in an arbitrary problem. If I remember correctly, it doesn't really matter unless the random variables in question are correlated or have some dependancy upon each other. I would suggest that the person with the relevent knowledge is SideMancer and I would defer to his opinion. On the face of it I would suspect that this effect would tend to accentuate the differences between experienced and inexperienced ships if such differences exist. If these differences don't exist it dosen't really matter.

The problem with it is it reduces the effective sample size because you have no way of telling the zeros apart.

Taking it to an obviously absurd extreme to illustrate the point, if both a 100 attack vessel and a 10 attack vessel go up against a 1M defence vessel you are going to need a ridiculous number of sample points before the difference you *know* exists becomes statistically "proven" because obviously almost every shot is going to do zero damage.

Meanwhile, if you look at a small sample size and the low attack vessel happens to get lucky first, it's going to look significant. This is why I put much more store in the original attack data which showed very similar values rather than the defence data which *appeared* to show a significant difference...the attack was rarely doing zero damage, whilst the defence was so high it looked like it would produce zeros more often than not.