WHY God - is there no decent Sailing Ship war game?!?!

[mwsheeler]

Can't wait to see the progress...

[tater]

Will have to try to mess with this tonight. Saturday completely got away from me, kid events all day, then I forgot that we rescheduled our anniversary dinner to Saturday night (wife was on call the night of our actual anniversary, Thursday), so that was a no-go. Today occupied by mother's day glop til after dinner.

[bix]

yeah i noticed that that particular ship, which happens to be # 0 in my array, seems to have a bit of it's own agenda. that is why intelligence on a computer is artificial. that ship also happens to be the first ship to execute a tack in its collision avoidance scheme. i'm looking at her closely too.

i'm gonna think about how i might represent each ship's target, at least for debugging purposes at first.

oh and regarding collisions from a previous post - as i mentioned, just like in real life, some ships WILL collide, others WILL run aground. i am fully aware of the event, just not handling it yet. i think when foes collide, in auto mode, i'll put them into a "grappled" state. when friends collide, i'll have each ship lose a couple stays.

i have bounding box collision enabled on all ships, so i can make them "bounce" off each other, and not pass through each other, but thus far i am not satisfied with how the "bouncing" looks so i'm ignoring the event. it is again mostly an issue of time and scale

best,
bix

[con20or]

Looks really promising. Most of the ships lined up beside an enemy

[troopie]

Awesome work bix! Thanks for the release, was great to see 'version 0.0.1'

I am seeing ships in irons when facing into the wind so it's all good, though most do seem to make headway a little too close to the wind.

As you've noted there is an awful lot of tacking going on, but I assume this is merely due to the restricted map size and a lack of an agenda/orders? I'm sure you'll get on top of it.

My fps was steady around 62.5 with plenty going on in the background so all good there, the horizon looks awesome btw!

I can't wait to see this senario with a northerly wind!

Do you have plans/ideas for implementing a 'line of battle' (depending on the era/nationalities you're representing I 'spose) or would that just be upto the orders issued by the player?

[troopie]

I left this running for an hour or so and they are slowly creeping off into the distance, are they free to roam the whole world?

Most have paired up.

[troopie]

Actualy they are all hangin around, just farther afield from the grid.

[bix]

any sim fully driven by computer with little or no random deviation (such as this version) will eventually devolve into a zero-sum state. it all grinds down to nothing if you don't shoot some energy into it every now and then.

so, i've fixed the "irons" they no longer make windward progress, but once in irons, they are stuck - no handler yet. soon enough. i think i'll have them boxhaul, or maybe even put out some boats to pull her around. we'll see.

i also reduced head to the wind to 30 degs, it was 25, too generous. now the tacking in more visibly correct.

re all the tacking - part of the ai is always trying to gain a windward advantage, so often they are overly zealous with gaining it.

just implemented initial gunnery!! will do a quick test and push it to my server in a few minutes. stay tuned!
bix

[bix]

okay, don't laugh!

and no comments about the cannon shot being, umm, how shall i say... "disproportionately sized" in relation to the ships :P

same url, same install instructions:

http://www.cityarts.com/signal/signal.zip

the latest will always be this url, i back up the previous release numerically, so the old one is now signal1.zip

so what we have here -

for the sake of demo, the range is extreme, and the ships quickly start out with their bow chasers. once within appx 1/2 tactical grid distance, they try to helm about for a broadside as best they can, given obstacles and other ships. again the range is extreme. if they cant make a broadside they will fire a bow or stern chaser. i have yet to decouple reloading/firing times between bow/stern and star and lar broadsides. for now one and only one battery fires, depending on how she bares, and reload is from 5-8 seconds.

the ships only target their opposite number and maintain that target throughout. this is helpful to visualize who is chasing whom, for debugging. naturally, the final will aim at the most opportune target (which still may not be the closest, the ai has a big tree of decisions to consider)

btw, any ship in the line of fire not the target is ignored, the balls pass right through (for now).

you will notice that the gunnery is unbelievably accurate, ie, they never miss. i must reiterate that the ai is NOT cheating, it is just insanely good at picking an elevation and powder charge to drop the ball on point. if you choose to man your own guns, you have the same control params, charge and elevation.

because the gunnery is not faked in any way, i will be able to easily modify the shot based on difficulty level, crew skill, wet or poor quality powder, drunkenness, morale, worn guns, hot guns, poor command, what have you, simply by manipulating a few variables.

oh, yes, the trajectory is also bit exaggerated for clarity.

enjoy! lemme know...

best,
bix

[troopie]

Nice work bix!

Their tacking is certainly more believable now, all though they don't avoid heading into the wind. Re the upwind stall; eventually wind and drift should pull them round to a point where they could catch some wind I'm sure, or kedging (draging a light anchor out ahead of the ship with a small boat and then pulling it in) would be an option (also handy if the variable wind becomes becalmed, or if they really feel they need to make headway upwind) but really they should just be programed to avoid heading into a stall to start with.

Re gunnery: looking very promising! Line of sihgt needs attention but I'm sure you'd already be working on it.

Re the download naming; I like how you provide the url but its gonna get a bit confusing/tedious for us mob; when I download and extract it conflicts with the previous files of the same name (I'm not in the habit of overwriting files) so unless I put each version into a new folder (even then, naming is still a problem), I've gotta rename it which then means my version numbering is gonna be inverse to yours which is gonna make communication pretty cumbersome I think. Is there an easier way?

[bix]

hi troopie!

yeah i think i need them to be less aggressive with getting upwind, but that means the action will surely drift downwind, away from the primary tactical area.

if say, the goal was to capture or defend a port, well ya cant really do that if yer all downwind of the port.

in case of defense, a line of ships bow to stern anchored would be a likely tactic, so i'm thinking of dealing with intelligent anchoring very soon.

re download naming, at this point, you can't be too concerned about previous versions, there is little to preserve. just overwrite the old file. the new one is better and besides, i'm maintaining the backups for you!

this way all forum posts point to the same url, to the latest, such that no matter how a person stumbles upon the url, it is fresh.

bix

[bix]

all,

i added a compass rose HUD element, click a direction to change the wind. for now just the 8 points are active - temporary, i need to make a special widget for this.

amazing what an effect the wind has. this is an instantaneous change in direction, i'll interpolate it soon enough (time to make a class just for the wind).

wind has no effect on gunnery yet, but it will soon. this will make the AI gunners less deadly.

the usual url (sorry troopie)

http://www.cityarts.com/signal/signal.zip

best,
bix

[bix]

captains,

gonna spend some time tomorrow resolving the new features, and will begin focusing on scenarios such as "capture/defend port" if anyone has a particularly favorite, well documented historical engagement of a small fleet action, please suggest it, and i'll try to build a scenario that is representitive.

in the mean time, after my rum ration, i was musing with pieter on pirates ahoy, regarding the fact that i made two releases in one day. thought i'd paste it here:

i'm just gonna keep pumping it out. i have a friday release planned, but will prolly push a new build tomorrow as well.

i've worked in this industry for more than 20 years, and the problem, from my perspective as a mathematician, is the reluctance on the side of the "bean counters" to allow "unproven" aspects of game-play reaching the public. i have no fear in this regard. i could not care less if some 16 year old kid doesn't like my new physics. which is why this project (at this point) is truly an "indie" game. all decisions are mine to make, i'm the captain, and i, my crew is 100% loyal.

several years back i worked on a complex procedural human animation system, for almost a year, and only 1/10th of my work ever saw the light of day, because for some reason or another, the avatars were not "cool" enough. screw that! we are all awkward uncool individuals. if i were as cool as brad pitt, i would not be a mathematician, i'd be a friggin movie star.

and that is the problem. people want to play games where they can be brad pitt, they don't want games that make them face who they truly are.

everything about this particular subject, ie an AoS sim, perfectly dovetails with my deepest convictions about the basis of our mathematical assumptions (and misconceptions), and the challenge of representing them. hell, those stalwart explorers had no clue of their longitude! nor do i, but i, like them, are gonna go there regardless.

i am building this sim for myself. i am listening to experts to ensure that what one sees on the screen is congruent to the expectations of those who possess germane expertise in the field. i'm building a casual game first, but the architecture must be physically and historically accurate, and must be scalable to any depth of interest.

these last few releases, with gunnery and variable wind, injected a huge amount of chaos into the representation. the more it becomes "true to life" the harder it is to understand. simply coordinating gunnery with navigation raises questions of "why did the ship go that way?" well, lets leave that to the court martial, an aspect of game play i intend to pursue as aggressively as bribing the port admiral for a refit.

to reiterate, my primary goal is to create an AI that kicks your ass, but never cheats. the challenge in terms of game design is to ensure you understand why the AI just kicked your ass.

i will soon inject real wave physics, currents, and wind gusts that will completely confound our understanding and expectations about 18th century naval actions. i hope we all may conclude that nelson was just bloody lucky! well, and he had balls.

best,
jbkds

[raymond6751]

Man, I love it. I especially love your attitude. Go for it.

I think we all wish we had the expertise to do what you are doing. But, as you said, if we all could do it...we wouldn't be who we are. We would be you.

Don't forget to have fun too.

[troopie]

Thanx again. I got lots to say but not much time. For some reason overwriting isn't working too well; will touch base tomoz.