Trying to Decipher Army Battle Damage [UPDATE 3-28-22]
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Trying to Decipher Army Battle Damage [UPDATE 3-28-22]
FINDINGS SO FAR (summary):
The base damage formula is [ { (Unit Attack Power x 100) + constant } x modifier ]
The modifiers are:
- Charge gets x1 vs. Charge
- Charge gets x1.3 vs. Bow
- Bow gets x0.8 vs. Charge
- Bow gets x0.9 vs. Bow
- Bow gets x1.2 vs. Magic
Enemy Charge uses the player's Unit Attack Power instead of their own vs. Bow
Enemy Bow uses the player's Unit Attack Power instead of their own vs. Charge
Magic deals a straight [200 x Unit Attack Power] vs. Charge or Magic
Poisonous Pollen defeats 80% of your soldier total
Armored Cavalry and Teo's Charge use the player's Unit Attack Power instead of their own
BEFORE FIRE SPEARS
- Takes no damage
- Damage is x1.3 vs. Everything
AFTER FIRE SPEARS
- Can only be damaged by Charge (which gets x1.3 against them)
- Damage is x1.3 vs. Magic & Bow
- Damage is x0.4 vs. Charge
- Armored Cavalry has a higher constant than Teo's Charge does
- You get one dude on the battlefield for each 400 troops you have (1 - 400 is one dude, 401 - 800 is two dudes, etc.)
- Don't know how the game decides who wins when Magic fights Magic
- Still scratching my head on how the constant is derived and how it changes from round to round. Sometimes it goes up, sometimes it goes down. I need to do another playthrough specifically focusing on that. It may have something to do with troop count, the attack power of your strongest unit, and/or whether you won or lost the previous round.
The base damage formula is [ { (Unit Attack Power x 100) + constant } x modifier ]
The modifiers are:
- Charge gets x1 vs. Charge
- Charge gets x1.3 vs. Bow
- Bow gets x0.8 vs. Charge
- Bow gets x0.9 vs. Bow
- Bow gets x1.2 vs. Magic
Enemy Charge uses the player's Unit Attack Power instead of their own vs. Bow
Enemy Bow uses the player's Unit Attack Power instead of their own vs. Charge
Magic deals a straight [200 x Unit Attack Power] vs. Charge or Magic
Poisonous Pollen defeats 80% of your soldier total
Armored Cavalry and Teo's Charge use the player's Unit Attack Power instead of their own
BEFORE FIRE SPEARS
- Takes no damage
- Damage is x1.3 vs. Everything
AFTER FIRE SPEARS
- Can only be damaged by Charge (which gets x1.3 against them)
- Damage is x1.3 vs. Magic & Bow
- Damage is x0.4 vs. Charge
- Armored Cavalry has a higher constant than Teo's Charge does
- You get one dude on the battlefield for each 400 troops you have (1 - 400 is one dude, 401 - 800 is two dudes, etc.)
- Don't know how the game decides who wins when Magic fights Magic
- Still scratching my head on how the constant is derived and how it changes from round to round. Sometimes it goes up, sometimes it goes down. I need to do another playthrough specifically focusing on that. It may have something to do with troop count, the attack power of your strongest unit, and/or whether you won or lost the previous round.
Last edited by wataru14 on Tue Mar 29, 2022 1:23 am, edited 10 times in total.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
I reloaded and tried again
Turns 1 and 2 were the same.
Turn 3 enemy used charge.
Remainder was 312 and I took 1411 damage when I charged.
The remainder was 35 points higher than it was in round 1. I suffered no casualties in Rounds 1 or 2.
Remainder was 252 when I used bow. 60 point difference, as expected.
I have no flipping idea how this constant is calculated. It will be some time before I can get to a battle where I can test further. The next one is Armored Cavalry, which probably has its own formula. Then there isn't another until Northern Checkpoint, but the troop count is so lopsided I might not be able to collect good data. I'll try again when I get to Shasarazade.
Turns 1 and 2 were the same.
Turn 3 enemy used charge.
Remainder was 312 and I took 1411 damage when I charged.
The remainder was 35 points higher than it was in round 1. I suffered no casualties in Rounds 1 or 2.
Remainder was 252 when I used bow. 60 point difference, as expected.
I have no flipping idea how this constant is calculated. It will be some time before I can get to a battle where I can test further. The next one is Armored Cavalry, which probably has its own formula. Then there isn't another until Northern Checkpoint, but the troop count is so lopsided I might not be able to collect good data. I'll try again when I get to Shasarazade.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
Armored Cavalry (no fire spears) do 130 with a remainder of 364
Teo's Charge (no fire spears) do the same but with a remainder of 154
Neither takes damage from anything
Armored Cavalry & Teo's Charge (with fire spears) both deal 130 with a remainder to Bow & Magic and take nothing
They deal 40 + remainder to Charge.
They take nothing from Bow & Magic and take 130 + remainder from Charge.
The remainder went from 493 to 550 to 102 (Teo's Charge) to 402 and it stayed there.
I am getting no closer to figuring out this damage constant. I'll have to play around more with Shasarazade.
At least the base damage values are consistent across battles.
Teo's Charge (no fire spears) do the same but with a remainder of 154
Neither takes damage from anything
Armored Cavalry & Teo's Charge (with fire spears) both deal 130 with a remainder to Bow & Magic and take nothing
They deal 40 + remainder to Charge.
They take nothing from Bow & Magic and take 130 + remainder from Charge.
The remainder went from 493 to 550 to 102 (Teo's Charge) to 402 and it stayed there.
I am getting no closer to figuring out this damage constant. I'll have to play around more with Shasarazade.
At least the base damage values are consistent across battles.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
I finished Shasarazade and am no closer to figuring out how the base damage is calculated. There has to be a pattern I'm not seeing. If nothing jumps out at me in the Final Battle, I'm probably going to give up and call it a "static damage bonus." Let someone else try to figure it out.
But I have managed to figure out a lot of the battle mechanics. I updated the first post with my findings and change it as new info is uncovered.
But I have managed to figure out a lot of the battle mechanics. I updated the first post with my findings and change it as new info is uncovered.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
And I spoke too soon. I did notice something.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
I think I'm going to give up on figuring out how the constant works. It seems to always be 40 points lower for bow than it is for charge, except when the constant is low. In that case it is 20 points lower. it comes out to a roughly 77-80% difference, but it doesn't work out exactly evenly. The fact that its a clean slat difference and the % is weird makes me think that is how it is supposed to be derived.
It goes down and up from round to round. Whether you won, lost, or tied in the previous round seems to have an effect on how much it changes, as does whether you used a strong unit or a weak one. But it doesn't change proportionally to the unit strength or to the change in troop counts.
Maybe I'll revisit it at some later point, but my motivation to keep reloading battles has ebbed and the frustration is getting to me.
It goes down and up from round to round. Whether you won, lost, or tied in the previous round seems to have an effect on how much it changes, as does whether you used a strong unit or a weak one. But it doesn't change proportionally to the unit strength or to the change in troop counts.
Maybe I'll revisit it at some later point, but my motivation to keep reloading battles has ebbed and the frustration is getting to me.
- TwilightSky
- Posts: 11
- Joined: Mon Dec 12, 2016 1:58 pm
Re: Trying to Decipher Army Battle Damage
Perhaps this constant is being changed by how long the fights take?
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
It seems to go steadily down from round to round, but then occasionally jumps up. I started a new playthrough and have savestates at the beginning of all the major battles so far (Pannu Yakuta, Garan, Scarleticia 2, Shasarazade, and Last Battle). This playthrough I only recorded rounds 1 and 2, but I can reload and go farther to test things if desired.
I have some data formatting to do, but I can post my numbers as an Excel file on Google Docs if people are interested.
I have some data formatting to do, but I can post my numbers as an Excel file on Google Docs if people are interested.
- Celes
- Global Admin
- Posts: 1422
- Joined: Sun Nov 12, 2006 6:25 pm
Re: Trying to Decipher Army Battle Damage
Maybe this guide that I have never published is of any help:
http://www.suikosource.com/games/gs1/guides/war.php
http://www.suikosource.com/games/gs1/guides/war.php
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
Thanks! I've been using that to help figure out the enemy constants. It has helped a lot! But I'm still no closer to figuring out the secret behind the constant. Someone with more skills than I may have to rip into the source code to find it.Celes wrote: ↑Thu Aug 13, 2020 4:36 pm Maybe this guide that I have never published is of any help:
http://www.suikosource.com/games/gs1/guides/war.php
My most recent research now leads me to believe the constant is more complicated than previously thought. The difference between Bow and Charge is always a round number. Always. That can't be coincidence. It is usually exactly 40 points, but that is for when the overall damage remainder is is around 200. For example, in Pannu Yakuta Round 1, the constant for Charge was 189 and Bow was 149. But when the remainder was 104 for Charge, it was 84 for Bow. Only a 20 point difference. In both cases, though, it was roughly 80% of the charge number.
This leads me to believe that the constant is really two parts. A base constant (which in this example cases would be 200 and 100) and a small +/- variance (possibly random, like with weapon damage). But that now takes one unknown value and makes it TWO unknown values. I still have not been able to find any pattern in how the constant changes.
Sometimes winning makes it go down, other times it goes up. Losing does the same. As does tying. It seems to change at different rates if you used a strong unit as opposed to a weak one in the previous round, but not all the time. My motivation to fiddle around with it has waned recently due to life stuff, but I have my save states so I can pick it up again. For now, I'm just going to say refer to the first post in the thread for base equations and leave the constant as a Question Mark for now.
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
I started playing a new game and I'm just about to recruit Kimberly. I have a new idea about the unknown damage constant and I'll try testing it out when I get to the second Scarleticia Battle.
My theory is this: 400 soldiers is the threshhold for a "dude" on the battle map, my theory is that the damage is based on the number of "dudes" killed.
For example: you have 8,000 soldiers. That's 20 "dudes" on the map. You take 1235 damage.
Your soldier count is now 6765, which is 16.9125 "dudes." So 16 at 400 and 1 at 365.
My thought it that the game does something at this stage. Either auto-killing the non-full "dude" and adding it to the expected damage or distributing your soldiers evenly among the "dudes" you have left. In the first case, the next attack against you should have a constant of 365. In the second case, your 17 "dudes" are each now 397. So the number of "dudes" killed in the next round should be off a little.
Both are working theories, so I'll have to see what the numbers show.
My theory is this: 400 soldiers is the threshhold for a "dude" on the battle map, my theory is that the damage is based on the number of "dudes" killed.
For example: you have 8,000 soldiers. That's 20 "dudes" on the map. You take 1235 damage.
Your soldier count is now 6765, which is 16.9125 "dudes." So 16 at 400 and 1 at 365.
My thought it that the game does something at this stage. Either auto-killing the non-full "dude" and adding it to the expected damage or distributing your soldiers evenly among the "dudes" you have left. In the first case, the next attack against you should have a constant of 365. In the second case, your 17 "dudes" are each now 397. So the number of "dudes" killed in the next round should be off a little.
Both are working theories, so I'll have to see what the numbers show.
-
- Posts: 20
- Joined: Wed Jan 05, 2022 3:09 am
Re: Trying to Decipher Army Battle Damage
This is actually really helpful!Celes wrote: ↑Thu Aug 13, 2020 4:36 pm Maybe this guide that I have never published is of any help:
http://www.suikosource.com/games/gs1/guides/war.php
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage
It is. Knowing what the enemy unit strength is is helpful to a great degree. But what I'm trying to do is figure out the formula for how damage is calculated. I almost have it (see 1st post in this thread) but there is a static modifier that is added to each attack that I cannot figure out for the life of me. I have tried everything and I cannot figure out how the game derives it.alberto90 wrote: ↑Mon Mar 21, 2022 8:14 pmThis is actually really helpful!Celes wrote: ↑Thu Aug 13, 2020 4:36 pm Maybe this guide that I have never published is of any help:
http://www.suikosource.com/games/gs1/guides/war.php
- wataru14
- Guide Writer
- Posts: 727
- Joined: Thu Aug 05, 2010 10:37 am
- Location: Las Vegas
Re: Trying to Decipher Army Battle Damage [UPDATE 3-28-22]
I had a breakthrough.
Watching longplay videos and using the data from those battles, I figured out some things about the constant.
It is always a round number (100, 200, etc.)
It is determined by your soldier count.
In the battle of Pannu Yakuta, the Liberation Army had 7015 soldiers. The constant was 200 in Round 1.
In one video, they made it to Round 2 without losing any soldiers. The constant was still 200. In another video, they suffered losses and began Round 2 with 4195 soldiers. The constant was 100.
In the second Teo Battle, the Liberation Army began with 12515 soldiers. The constant was 400 in Round 1. They began Round 2 with 11313 soldiers and the constant dropped to 300.
In the Final Battle, the Liberation Army started with 18710 soldiers in Round 1. The constant was 600.
After these constants were applied, there was a remainder ranging from -18 to +30. That can easily be attributed to random variance.
I think this is it. I'm going to watch more videos in the morning to get a larger sample size, but I'm 99% sure this is how the damage constant is calculated.
As a rough estimate:
* 1-4195: 100
* 4196-9744: 200
* 9755 - 12514: 300
* 12515 - ?: 400
* ? - ?: 500
* ? - 18710: 600
I'm sure I can narrow that down watching more playthroughs.
It's possible that the constant is set differently for each battle, but that doesn't explain why it went down in cases where the Liberation Army lost soldiers. It has to be derived from soldier count. My hypothesis is that it is 100 for every 3000 or so soldiers you have. So:
* 1 - 3000: 0
* 3001 - 6000: 100
* 6011 - 9000: 200
* 9001 - 12000: 300
* 12001 - 15000: 400
* 15001 - 18000: 500
* 18001 - ??? : 600
There are some slight discrepancies, so the tiers might not be even, but I think I'm on the right track with that.
Watching longplay videos and using the data from those battles, I figured out some things about the constant.
It is always a round number (100, 200, etc.)
It is determined by your soldier count.
In the battle of Pannu Yakuta, the Liberation Army had 7015 soldiers. The constant was 200 in Round 1.
In one video, they made it to Round 2 without losing any soldiers. The constant was still 200. In another video, they suffered losses and began Round 2 with 4195 soldiers. The constant was 100.
In the second Teo Battle, the Liberation Army began with 12515 soldiers. The constant was 400 in Round 1. They began Round 2 with 11313 soldiers and the constant dropped to 300.
In the Final Battle, the Liberation Army started with 18710 soldiers in Round 1. The constant was 600.
After these constants were applied, there was a remainder ranging from -18 to +30. That can easily be attributed to random variance.
I think this is it. I'm going to watch more videos in the morning to get a larger sample size, but I'm 99% sure this is how the damage constant is calculated.
As a rough estimate:
* 1-4195: 100
* 4196-9744: 200
* 9755 - 12514: 300
* 12515 - ?: 400
* ? - ?: 500
* ? - 18710: 600
I'm sure I can narrow that down watching more playthroughs.
It's possible that the constant is set differently for each battle, but that doesn't explain why it went down in cases where the Liberation Army lost soldiers. It has to be derived from soldier count. My hypothesis is that it is 100 for every 3000 or so soldiers you have. So:
* 1 - 3000: 0
* 3001 - 6000: 100
* 6011 - 9000: 200
* 9001 - 12000: 300
* 12001 - 15000: 400
* 15001 - 18000: 500
* 18001 - ??? : 600
There are some slight discrepancies, so the tiers might not be even, but I think I'm on the right track with that.
-
- Posts: 95
- Joined: Wed Jan 29, 2014 9:17 pm
Re: Trying to Decipher Army Battle Damage [UPDATE 3-28-22]
Any reason to not just check the game code?
I don't have a save near a battle right now so I can't check, but I can try to get one there to validate.
I don't have a save near a battle right now so I can't check, but I can try to get one there to validate.