Cute Warbands Doing Cute Warband Things

 

Smoke Signals, telling Yogi Bear the picnic table is occupied

So after much huffing and rather a lot puffing, and even more debugging ... warbands are running about doing warband things, in both day and night!

Warbands now move around from loot point to loot point, doing the lootening, as loot points suggest may happen, and light the bonfires on said loot points to signal to all nearby that the lootening is occurrening - and that this would also be a great time to raid that said specific loot point and try and slaughter all within to loot their loot.

Did I mention loot? 

 Well the player can now loot corpses, which was a main aim of the gameplay, and not only can he swipe the swag that the fallen is clad in, stripping them down to their underpants, he also buries said unfortunate recently deceased, leaving a small gravemarker to mark their passing. Failure to loot and bury said dearly departed, results in their equipment being lost to eternity, and their corpses rotting away to leave a skull, marking their departure from this earthly plane.

The Freshly Slain

 

The Not So Fresh After Remaining Unburied

I had a play around with the ScatterSky and TimeOfDay objects, and after some faffing in the code, decided the easiest way - at least in network overhead - to get really red dusks and dawns, was to use the built in TimeOfDay callback with a little custom C++ to change to change the rayleigh scattering and ambient colour. This callback also allows for the campfires to light up on sunset and extinguish at dawn, giving more markers as to places the warbands are likely to head. View distance for the warbands changes according to the time of day, and leads to some rather nasty surprise battles during the hours of darkness, when bands or burley warriors bump into each other in the pitch black.

 Archery now works rather nicely, though required considerable faffing due to the blending of animations for aiming, which threw off the IK animations which I had created for use with two-handed pole weapons (hello Dane Axe). After many huntings through C++ to try and find out WTF was going on (I still think it's all down to the laterally animation movement which the engine isn't expecting) it turned out easier to define a specific look/head arm animation for pole weapons and then manually alter the animation to get it to align to the IK by around -4 degrees on the XZ axis. Not the cleanest way of getting a working result but game development, and inface coding in general seems to be very much to be a matter of workarounds and use more duct tape. And duct tape fixes everything! (as long as you use enough)

Here's a test of leading aim with longbows at just short of 300m, which works a lot better on a flat surface than undulating terrain for obvious reasons of consistent direction and velocity.

 Currently all bows look like crossbows, due to this being the only model and animations I have currently made, whilst I decide on how best to animate traditional bows, with their hold in left and draw with right animations. As my current warband timeline only goes up the 11th Century at the moment, the English warbow/longbow has yet to appear, but researching some historical composite bows for the Magyars, it was interesting to  note that they used very light arrows on very strong composite bows to out-range their enemies at the expense of projectiles which were weaker in the strike, and would break when they hit the ground (which of course stops the guy you just missed from picking it up and shooting it back). To this result, the Magyar have by far the longest range (to be honest 400m is kinda pushing it but hey vidya gaemz), but with stopping power only marginly better than a 100m short bow.

Remember kids, when you see the horse archers run away, do not follow. 

 So, here's some edited highlights of archery, melee, point blank night battles, corpses rotting to skulls, and general AI powered gameplay in action.

 So all that leaves us with our not so heroic player, still running about in his underpants armed only with a Bollock Dagger.

Nice lack of kit you have there oh level 1 noob

 But, as shown in the first video far above, the player can actually loot corpses of their equipment, armour and weapons. So the next major job is to come up with a Game User Interface (GUI) so that the player can actually decide what stolen swag they want to equip, to boost their avatar's chances of survival in a good old fashioned scrap. Alongside that, I need to create interfaces for the leveling system and career progression, all of which is already coded, but currently inaccessible due to lack of tailor made input screens.

Behold, The Wonder Of Clothing! Coming Soon™

 

Speaking of input screens, I modified the visual performance settings, so things liek groundCover (that's grass to most people) not only increases and decreases in number, but also expands and ... er ... despans(?)  in range from the camera. I found that there was already a variable for this, originally created in mind to reduce the range, but with a little tweaking of code I could get the radius to which grass is drawn on screen to increase/decrease in line with element numbers.

In other news some dastardly swine stole an hour of my sleep in return for British Summer Time, so I sat out in the sun with a bottle of wine, and it was freezing cold and cloudy and I probably now have hyper/hypo-thermia ... which ever one involves a brass monkey dropping it's balls all over the place.
 

Mulit-Level Long Play Test

 Blast monsters, grab loot, buy swag powerups, unlock nodes, beat boss, run to exit. Rinse and repeat for next level.

Finally I actually got around to blocking out some level design. Nothing fancy, just a plan of walls and obstacles for the player to manoeuvre around (apparently that's maneuver for Americans).

It didn't take long before I realized that the whole thing was way too big and needed to be smaller. But that wasn't the only issue. As the game time increases in a level, so does the difficulty. And as the levels increase, the game gets harder. Obviously this is a tried and tested gaming concept, but I was finding that the difficulty curve ramped up very sharply. A little rejigging later and ... it was far too easy. Minor adjustment and too difficult again. Balancing this difficulty curve has taken up the vast amount of time this month.

It turned out that a lot of this difficulty stemmed from the player's powerups not reinitiating in later levels. I had made various changes to the powerup system and how it gets stored for reference and not all of it carried across correctly, and in one case I found that the variable being stored had received a name change at the other end which was no longer being referenced.

Going through the various powerups I fixed all the broken paths and references, and generally tweaked the strengths and effects. Gradius style "option" followers, or weapon satellites which follow the player and increase firepower were toned down in the amount of damage they did. "Static Heal" powerup was scrapped because the player never has chance to stop long enough to increase healing without getting swarmed by enemies. Enemies and their spawning routines were tweaked to give a more gradual rise in challenge but also to watch out for the player wiping the floor with them over a longer period of gameplay.

Each level has an exit, and each exit requires unlocking via activating nodes with later levels having multiply nodes to unlock. To make sure that the player understands this concept there are visual hints in the bottom right of the screen and the in the first level the node is next to the exit. Once all nodes are unlocked the level boss spawns (currently just a large cube with an aggressive charge attack) and once the boss has been defeated the exit opens. To prevent the player from lingering in the level to "clean up" cash and remaining swag I decided to stop the enemy from dropping loot coins once the boss has been defeated. Difficulty also takes a temporary increase both whilst the boss is released and also when they have been defeated to encourage the player to head for the open exit and progress the game. 

In the end it all seems to work rather well, and I am pleased with the result. Here's nearly an hour of playtesting the system. I also found an error at the end of it, as not statistical information was being passed to the end of level screen when the player was killed and the game was over - now fixed.

So, next up is adding a few more statistics to the end of level screen such as percentage of kills, cash, swag, etc that the player has gained that level. After this it will time to get on to actually designing art for the level, and eventually even the characters with a greater variation of enemies, bosses and enemy attacks. But hey, one thing at a time! ;)

Loot Crates And Swag Parcels

Loot Crates and Swag Parcels ... are actually the same thing. Previously they were just placeholder grey cubes, but now they have a final form.

Top row Large crate (rare swag items), middle row Medium crate (uncommon items) and bottom row Small crate (common and useable items). Each crate type has three models which are placed with a randomized rotation.

When the server loads the level, Loot crates are randomized throughout the game world. Initially I added an impact decal with each crate but realized that previously spawned decals do not get passed to new clients when they join unless they are first saved into the level's decal file. This seemed a bit of a faff, so I simply modeled the impact texture into the actual loot crate mesh.

I also replaced the Level Exit Node (which up until now had been just another Activation Node) with it's own model and updated the on-screen prompts to reflect this.

The post-level status screen has also got (yet another) change, replacing the text with icons though text explanations will pop-up via using mouseover. I was hoping that the icons would be fairly self-explanatory but decided to include the pop-up text for safety's sake. Nothing worse than a player wondering what the hell they've got 30 somethings of.

And here's how it all looks in the demo level I use for testing when the Exit is unlocked and the boss fight against Placeholder Huge Cube Boss begins.

Next up, either start to flesh out the remaining models for powerup special attack events - which will lead onto creating characters and the like, or it will be time to start creating some actual game levels - all of which will no doubt start out as whiteboxing.

Decisions, decisions, the pressures of command ...