25th OCTOBER NEWS
It's time to fill you all in on the progress!
BOARD GAMEBOT SUPPORT PROGRESS
As stated in the last posts, I have been working on adding bot support, and I'm happy to say that this is progressing steadily given the minimal hours I have to work on the game, but it's also as confusing and chaotic as the GIF above (NOTE: All GIF's here are bot controlled players)
Currently I have now got the board game mode on it's own without mini-games functions with bots.
The bots aren't super smart though and still require work, mainly in regards to the avoidance of hazards that shoot projectiles, but also in regards to bots placing hazards in sensible locations that "make sense".
The idea is for bots to look at the same things a player would when deciding firstly IF a hazard should be placed, and then WHERE it should be placed, rather than randomly deciding these things, there needs to be logic behind the decision. For now though, the board game AI is "passable".
The bots also can navigate obstacles that players can, as previously seen, however more variety in their pathing needs to be added, and deviations from their path need to be added to more accurately reflect what players may do, for example, taking shortcuts or a safer route if there's a lot of hazards detected along the path and there is a better, safer route.
Additionally, getting bots to work with drop in/drop out style of gameplay has been a bit of a nightmare and still isn't quite working just yet, but it's almost fully functional so far.
I am yet to do any online testing with bots too, however my bot support has been factoring in the online functionality too, so fingers crossed there's not too much to iron out when I do online testing.
MINI-GAME BOT SUPPORT PROGRESS
Regarding bot support for mini-games. This has been very difficult due to platformer style and skill based nature of all of the mini-games (so far), but when the bots work, it's an incredible feeling, it's progress.
The intention is to get each mini-game to a playable/passable stage so that the experimental build can be updated and you can all play with bots sooner rather than later, but when that occurs, this doesn't mean every mini-game will have smart bots, they are potentially doing the bare minimum.
Here is a list showing the progress of bot support for each mini-game in alphabetical order.
(NOTE: Anything over 65% is considered "passable/playable")
Blasted - 65% (Bots run around randomly, drop bombs and kick bombs, however they don't jump blocks, don't go for pickups, don't try to lead bomb kicks or bounce bomb kicks yet)
Blue Ballin - 80% (Bots against the rotation direction of the ball, and jump when they are far from their destination, much like players would, they also try to jump sweepers and sawblades, but don't avoid the bounce/hex pads, they really may not need to though)
Coallateral - 95% (Bots run to the other side, jump over gaps, and grab coal, then return it, stopping when a bridge is retracted, and trying to move away when coal is shooting from a pipe, they currently aren't grabbing nearby dropped coal though, and the drop-off can be a bit of a cluster with AI navigation, but I've put in measures to reduce AI effectively lining up and waiting to drop the coal off).
Cognitive Collection - 0%
Creeping Creepers - 0%
Dead Still - 0%
Dark Dangers - 0%
Deadly Disco - 95% (Bots currently don't avoid falling tiles, however everything else is fine)
Dodgy - 60% (CURRENTLY WORKING ON) (Bots try to find the safest location on the stage, however I'm currently working on bots avoiding nearby balls and trying to predict nearby safe locations when lots of balls are close to them, they is a lot to factor in)
Falling Apart - 0%
Get Pucked - 95% (Bots try to move the ball towards their goal, they know where they need to hit the puck, they know how to boost, they try to go around the puck to the side they need to hit, but it's not quite where I want the AI to be yet).
Knock Off - 0%
Laser Lunacy - 60% (This is a tough one, like Dodgy, AI check the closest location on the laser and get it's predicted movement to try and determine if it should jump or run away, it's not working as intended and needs more work. AI will try to avoid open hatches etc too, but will often just jump on the spot over and over again when detecting danger, which leads to them being eliminated far too easily).
Melting Memory - 95% (Bots behave like players, I've implemented a memory system for bots that is based on a number of factors, such as what round the mini-game is up to, how many players there are (visual confusion), has a rotation just occurred (creates a higher percentage of failure that cools down over the rounds), has a player just been eliminated (distraction), has a player just run onto the bots location (self gas-lighting/2nd guessing), etc....the bots feel like players, but I just need to add some aesthetic things such as bots spinning around and jumping around on their safe location while they wait for a new round, this is purely aesthetic though).
Oh Balls - 85% (Bots move randomly around the stage, but have a high chance to move towards another player, players generate target scores that the bot selects from. These scores factor in distance from the bot, obstacles between other players and the bot such as pillars or other players, distance from the edge, velocity of the target and/or other players. They also attempt to jump incoming players, avoid them, or roll straight back at other players.)
Roasted - 80% (Bots run around at random, they look for the most lucrative locations, so they tend to prioritize clusters of plectrums over single plectrums on their own, and they avoid flaming vents)
Tap Out - 0%
Termination Trail - 0%
Wheely - 0%
I have been trying to complete support for at least 2 mini-games a week, however some of these mini-games are a bit of a nightmare to program the bots for, especially mini-games where there's a lot of platforming, or a lot of hazards... (oh wait..that's...almost every single mini-game), and with day to day life thrown into the mix, I sometimes have only gotten only 1 mini-game done.
But it is fun learning, and as I learn, things will get better, I may re-do previous bot AI if I learn something new so that I can make the bots smarter for example.
I also intend for bots to behave as much like a human player as possible as mentioned in a previous update, they should have "personalities" as such, which would influence behaviours.
For now though, the goal is to get the minimum viable product out to you all so you can enjoy playing with bots.
MINI-GAME CHANGES AND POSSIBLE FUTURE CHANGES/IDEAS
Since I have been working with mini-games again, rather than being stuck working on the board game 100% of the time, I have been reminded of some of the mini-game changes that I've been wanting to make for a while. Some ideas have already been implemented as part of working on bot support., some have been noted down or considered for the future.
Cognitive Collection - will likely be updated so that player goal locations shuffle every so often. Alternatively the secondary mode may include a randomly lit up cog that triggers the shuffle when a player runs over it, which allows more player interaction.
Get Pucked - Instead of players getting 1 point whenever a puck goes over their goal, the player now continuously get points every 0.5 seconds until the puck goes over another goal.
The puck will now push players out of the way when bouncing out from the edge of the arena if all the players are clumped up. And when a player boosts, they have more push force to knock other players out of the way or bounce the puck. Additionally the boost function now creates particle effects, but it does need a SFX to further emphasize that they are boosting.
I may make it so the goals shuffle every so often, or perhaps there will be a floor section that will rotate the goals around, however I'd rather reduce moving parts due to online movement sync issues.
Blue Ballin - I've done some aesthetic changes by adding fog so that it's easier to see the silhouette of the ball now, however at a later stage I intend to update the background. It's not super important at this stage however.
Laser Lunacy - I may introduce plectrums again, or introduce this as a secondary mode, the plectrums would serve as a balance against the time alive. So if there were 2 players left in the round, but one of them have 50 plectrums and the other has 20 plectrums, but the one with 20 plectrums is the last player alive by 0.5sec, then player 2's plectrum count would factor in, perhaps each plectrum would increase the alive time by 0.05sec each, effectively giving player 2 another 2.5sec alive, and allowing them to win by 1 sec (30 plectrums difference = 1.5, and player 1 was alive 0.5 sec longer, therefore player 2 wins by 1sec).
This was previously in the mini-game, but it caused a lot of confusion, so we'll see what happens...it just adds more incentive for players to run around.
Alternatively, I could add power-ups like shields, double jump etc.
Melting Memory - The arena will likely change so it doesn't look like Roasted, but green. This is yet to be determined what will be done, but I have a few different ideas, I just need to make sure they are fool proof. Additionally, I have idea's for eliminated players which would allow them to still participate to reduce boredom while waiting for a winner as this mini-game can be a bit slow.
Oh Balls - I'd like to add power-ups to this at some point, effectively these would be kind of rare to spawn, they would spawn in the place of lowered pillars, and players would need to keep the ball on the power up location for 2sec to collect it. These power ups may be "Lead Ball" (heavier ball, no bounce), "Time Bomb" (after a period of time, the ball releases a pulse that knocks nearby balls away), "Double Jump" (self explanatory), "Air-Ball" (Allows greaters air control when jumping), "Ghost Ball" (Can pass through players and pillars upon activation), "Teleport" (Allows the player to teleport to a random location on the stage upon activation).
Roasted - Various pickups could be added to Roasted, such as immunity to fire, weapons to knock players off etc, or I could add fighting mechanics to this mini-game.
Tap Out - More variety of obstacles and environment.
Creeping Creepers - Various Pickups that allow invisibility to the creepers, faster movement, random teleport etc, and of course fighting mechanics could be implemented to mess up other players.
FIGHTING MECHANICS
I do want to add some "fighting" mechanics to a lot of the mini-games, as I think player to player interaction is paramount, but this would be a feature that should be able to be toggled on/off as some people don't like it at all.
The fighting mechnics would be super simple, there would be a head bounce, a telegraphed slap, and a jump kick if the player were to run at another player, jump and press the slap button.
ADDITIONAL NOTES
I'll continue to update you all on the progress as much as I can during one of the busiest times of the year, I want you all to see where things are at, and I encourage everyone to provide their feedback and suggestions, I listen to everything, a game like Metal Heads NEEDS that feedback from you all in order to succeed.
Please also let me know what you would like to see once bot support is added? Do you want more mini-games? Improved mini-games? more boards? character customization? etc
THANK YOU ALL FOR YOUR CONTINUOUS SUPPORT, YOU ROCK!
DISCLAIMER FOR NEW METAL HEADS SUPPORTERS
For those who are new to Metal Heads, I am just one solo dev, working on this in what little spare time I have between my full time job, relationship and life in general, but eventually if Metal Heads does become financially successful, this will enable me to work on the game again full time like I used to.
But I have to re-iterate this as to not create false expectations that the progress will be super fast like a fully funded indie game, or an indie game where the developers live in a situation where they are funded by family etc.
Then the bots would try to long jump this section of the board, rather than landing on the safe spot in between...as you can see by the little coloured sphere's, there wasn't one in the safe spot. This was due to not understanding my own programming...
Then there is a nice colourful sphere in the right place, the green sphere's represent an AI navigation point before a gap is detected, but as you can see, the sphere is a long way away from the edge, so the AI would jump too early.
Now as you can see, after the fix, our little Metal Heads can now traverse this gap very majestically!
Aside from these problems, there were a bunch of other problems too, such as the AI continuously jumping like a maniac once they got across the gap, or sometimes they'd get half way across, then float backwards and fall into the slime (this took about 2.5 weeks to fix), there was also the issue of frame rates dropping significantly, the AI only making it across the first 2 gaps, then failing 9/10 at the 2nd gap, and that one bot that made it past, would somehow fall into the slime at the 3rd gap or 4th gap....
