As a French person I feel like it's my duty to explain strikes to you. - AdrienIer

Create an account  

 
Caster of Magic Release thread : latest version 6.06!

So, if heroic shout 12 damage per figure for 12 mana and fireball is 15 and 15? does this end up making heroic shout nearly useless for chaos mages? I know 8 was op, but maybe 11 mp would've made it still better than fireball? or is there some reason I'm not comprehending that makes fireball worse?
Reply

(September 28th, 2016, 00:03)namad Wrote: So, if heroic shout 12 damage per figure for 12 mana and fireball is 15 and 15? does this end up making heroic shout nearly useless for chaos mages? I know 8 was op, but maybe 11 mp would've made it still better than fireball? or is there some reason I'm not comprehending that makes fireball worse?

Fireball is 10 for 15 and it counts as Chaos and Fire damage.
Reply

I keep quoting the threads rather than checking in game, whenever I'm feeling lazy. I guess a lot of information in them is out of date.

http://www.realmsbeyond.net/forums/showt...p?tid=8090

4. Fireball
Cost : 15
Research : Normal
Effect : Hit all figures in the target unit with a strength 15 magical fire attack.
An excellent direct damage spell to use on multi-figure units, it can usually wipe out midgame or weaker normal units entirely.
Reply

I think that's a typo, don't remember it ever being 15.
Reply

Eek when I tried 2.9 twice, in 1401 my magic spirit is attacked by the computer but the game does not go to battle scene, just to next turn with both units still alive... user error?
Reply

(September 29th, 2016, 09:37)coldsteel Wrote: Eek when I tried 2.9 twice, in 1401 my magic spirit is attacked by the computer but the game does not go to battle scene, just to next turn with both units still alive... user error?
If the AI is not hostile but you somehow end up with units on a tile they were planning to move to (not attack) they cancel that move, the same way as you do when you click "No" on the "You have a treaty with this wizard, attack anyway?" window.
The difference is the AI can click that "no" even if they don't have a treaty.
This most likely happened because you walked onto the tile the AI designated for gathering units, building a city or board a ship from or something like that.
It can be a bit misleading as the "taking back" portion is not shown unfortunately but I rather have this than the AI losing settlers because your spirit is on the tile they want to build on.
Reply

(September 29th, 2016, 09:59)Seravy Wrote:
(September 29th, 2016, 09:37)coldsteel Wrote: Eek when I tried 2.9 twice, in 1401 my magic spirit is attacked by the computer but the game does not go to battle scene, just to next turn with both units still alive... user error?
If the AI is not hostile but you somehow end up with units on a tile they were planning to move to (not attack) they cancel that move, the same way as you do when you click "No" on the "You have a treaty with this wizard, attack anyway?" window.
The difference is the AI can click that "no" even if they don't have a treaty.
This most likely happened because you walked onto the tile the AI designated for gathering units, building a city or board a ship from or something like that.
It can be a bit misleading as the "taking back" portion is not shown unfortunately but I rather have this than the AI losing settlers because your spirit is on the tile they want to build on.

Hmm I would be surprised if my game's behavior is what you intended. It is something completely new compared to 2.81. My single spirits get attacked over and over, for years, chased every turn, but never killed. This preoccupies the enemy's heroes and stacks so those troops get nothing else done. But heck I'll give it a go tonight and see where it leads. Maybe I can win on impossible with my rush slinger stack heh. I'll try lots of spirits and paralyze the enemy.
Reply

(September 29th, 2016, 15:14)coldsteel Wrote:
(September 29th, 2016, 09:59)Seravy Wrote:
(September 29th, 2016, 09:37)coldsteel Wrote: Eek when I tried 2.9 twice, in 1401 my magic spirit is attacked by the computer but the game does not go to battle scene, just to next turn with both units still alive... user error?
If the AI is not hostile but you somehow end up with units on a tile they were planning to move to (not attack) they cancel that move, the same way as you do when you click "No" on the "You have a treaty with this wizard, attack anyway?" window.
The difference is the AI can click that "no" even if they don't have a treaty.
This most likely happened because you walked onto the tile the AI designated for gathering units, building a city or board a ship from or something like that.
It can be a bit misleading as the "taking back" portion is not shown unfortunately but I rather have this than the AI losing settlers because your spirit is on the tile they want to build on.

Hmm I would be surprised if my game's behavior is what you intended. It is something completely new compared to 2.81. My single spirits get attacked over and over, for years, chased every turn, but never killed. This preoccupies the enemy's heroes and stacks so those troops get nothing else done. But heck I'll give it a go tonight and see where it leads. Maybe I can win on impossible with my rush slinger stack heh. I'll try lots of spirits and paralyze the enemy.

This sounds like a serious bug then. Can I have the save file?

Edit : confirmed, I started a new game and the same thing is happening to me. I need to find the source of this bug ASAP.
...found it. When I edited the intercontinental attack procedure, for some reason one line of code got deleted which was "if hostility off do not attack"... No idea how that happened, didn't even change that specific part of the code.
Reply

Update!!!
Thanks for the bug report, considering the severity of the bug, instant new version!
Not much changes this time, but the AI speedup is quite significant. I tested each AI procedure separately for time consumed and it's unlikely to improve speed any more than this, time taken is divided evenly between various movement types, and most likely it is consumed by the actual moving of the units, not the AI's thinking. The "shortest path" the moving units have to find to their destination is a huge bottleneck because it has to process a 1000 tile map and this cannot be solved in linear time. I don't think I can make it run any faster, I tried other solutions but they were many times slower. I'm actually surprised the one I made runs this fast.
tl;dr : This is as fast as AI turns can be, unless we get a faster DosBox or a faster CPU.

Quote:2.91
-When the AI checks if a target can be reached by land movement for any purpose, instead of calling the “shortest path to destination” procedure, it'll only check if the tile can be reached. This speeds up AI turns significantly. (A 5 minute long turn was observed to now complete in only 2 minutes)
-AI settlers will no longer try to go through a tower if that player already has 9 units in it. (If the other side of the tower has no military targets, units going through can get stuck inside)
-Fixed 2.9 bug : AI intercontinental stacks try to attack without hostility.
Reply

When memory is cheap and cpu time is expensive pathfinding nodes are pre-computed. As in you could pre-compute a large number of paths when the landmass forms and the game starts.
This is how most modern games do it, iirc. Or at least 90s games like total annihilation or baldur's gate.
I know research into real-time ai pathfinding is still cutting edge (for things like second life/minecraft/voxel games where by obstacles can be added after the initial pathfinding grid is computed.)


That said I assume the code already does this and it's still not linear to use it, or there's no way to add such a huge section of new code?
Reply



Forum Jump: