Dear forum users! In compliance with the new European GDPR regulations, we'd just like to inform you that if you have an account, your email address is stored in our database. We do not share your information with third parties, and your email address and password are encrypted for security reasons.

New to the forum? Say hello in this topic! Also make sure to read the rules.

Status Update 2018-10-25

Status updates from MythoLogic Interactive about the game progress.
Forum rules
By using the forum you agree to the following rules.
Post Reply
User avatar
Gurt
Lead Programmer
Lead Programmer
Posts: 1884
Joined: Sun Feb 28, 2016 3:22 pm
Title: Lead programmer
Started SFD: Made it!
Location: Sweden
Gender:
Age: 34

Status Update 2018-10-25

Post by Gurt » Thu Oct 25, 2018 8:09 pm

10 days ago I said we needed about two more weeks to work on the game. Since then we have decided to add yet another few things into the game before the release which means we're going to need yet another two weeks of work. We concluded that it was a must to add these last minute things into the game.

If everything goes well we might be able to release the game mid November or late November, which I want as some sort of ultimate deadline.

I made a video over bots fighting and they are getting better and more challenging to fight. Click here to watch it on YouTube. The video is only bots fighting.

-------
The game is planned to be released during Q4 2018.

We will continue to work on the game after the release to bring more content and features to the game as time goes on. As we're working on this in our spare time the update frequency might be similar to the Alpha period.

The MythoLogic Interactive accounts will not be required, nor useable, in the Steam version of the game, which will be the platform we release the game on. We will most likely sell the game on other PC-distribution platforms in the future but we're focusing on Steam right now. Steam will be required to play the game.
7 x
Gurt

User avatar
Del Poncho
Superfighter
Superfighter
Posts: 262
Joined: Sat Mar 19, 2016 8:57 pm
Started SFD: July 2012 (1.0.5)
Location: Poncholand
Gender:
Age: 28

Post by Del Poncho » Thu Oct 25, 2018 8:17 pm

In your final view of the game, will the bots be able to detect dangerous objects such as barrels, canisters, c4s and such?
2 x
The risk I took was caluclated, but man.....am I bad at math.

User avatar
Gurt
Lead Programmer
Lead Programmer
Posts: 1884
Joined: Sun Feb 28, 2016 3:22 pm
Title: Lead programmer
Started SFD: Made it!
Location: Sweden
Gender:
Age: 34

Post by Gurt » Thu Oct 25, 2018 8:28 pm

Del Poncho wrote:
Thu Oct 25, 2018 8:17 pm
In your final view of the game, will the bots be able to detect dangerous objects such as barrels, canisters, c4s and such?
We have plans to let bots analyze dangerous zones and objects such as barrels, canisters, mines and crushing elevators. This is planned for next year at some point and will come as a later update to the game.
4 x
Gurt

User avatar
piterskiy
Superfighter
Superfighter
Posts: 60
Joined: Mon Oct 30, 2017 12:12 pm
SFD Account: PiterskiyReal
Started SFD: PreAlpha 1.4.X
Gender:
Age: 20

Post by piterskiy » Thu Oct 25, 2018 9:35 pm

damn, these bots can do some nice combos now!
0 x

User avatar
KliPeH
Moderator
Moderator
Posts: 914
Joined: Sat Mar 19, 2016 3:03 pm
Title: [happy moth noises]
SFD Account: KliPeH
Started SFD: Pre-Alpha 1.4.2
Gender:
Contact:

Post by KliPeH » Sat Oct 27, 2018 9:08 am

Gurt wrote:
Thu Oct 25, 2018 8:09 pm
[...]

If everything goes well we might be able to release the game mid November or late November, which I want as some sort of ultimate deadline.

[...]
My birthday is late November so that would be a nice birthday gift indeed! Don't let me down Gurt :heart:

Reading the status update I see you've somewhat attempted to tweak bot behavior patterns in order to make them more intelligent and versatile, but there's a pattern I see time and time again which stumps your progress and hasn't been getting any attention. It's been problematic for me ever since I started using bots in my games and that is especially true for 1-player-with-7-bots lobbies. It is prevalent in your video too which is why I'm bringing this up now.

Bots more often than not will split into pairs at the beginning of the round and engage each other in melee combat until one dies, at which point they will seek out the nearest live target and employ the same tactic again. Due to all bots being on the same level of melee combat "expertise" the winning fighter always emerges victorious at extremely low HP (if not straight up dead), posing a very low threat to other bots or, in my case, myself. The fighting not only weakens it but also takes up most of the time it would otherwise use to loot the battleground. Since no matter the amount of bots I set only one stray bot will ever go after me (even number) or none will at all (odd number), picking the low-healthed under-equipped fighters off at the end of the round always earns you very easy wins.

Will the bots ever have an internal "nah fuck that" switch that they toggle at random(?) in order to escape combat and actually equip themselves with decent weaponry that could contest minutes of undisturbed looting on a real player's side? Will they ever have a loot "threshold" that won't allow them to engage other players until it is reached? Will the AI ever elect to attack a group of fighters (or no fighters at all) instead of focusing one fighter down at a time? Perhaps the Expert AI could adopt lower-difficulty-level AI strategies from time to time to break the monotony of "I will slap you to death or die myself" that persists throughout each one of my playthroughs?

I feel allowing bots to occasionally use "runner" tactics would only do good. I know you've said bots will never attempt to play passively to bait/camp a real player out, but choosing to play passively/aggressively is different to straight up playing sub-optimally; if the AI does not give itself a few breathers in combat it will continue to exhaust itself and provide easy kill for the better-equipped players. After all the current AI opponent will become the highest preset AI difficulty in the game so I think making it slightly more flexible could make it as unpredictable and thus nearly as strong as real, thinking players are.
0 x
 
Image

User avatar
Gurt
Lead Programmer
Lead Programmer
Posts: 1884
Joined: Sun Feb 28, 2016 3:22 pm
Title: Lead programmer
Started SFD: Made it!
Location: Sweden
Gender:
Age: 34

Post by Gurt » Sat Oct 27, 2018 10:09 am

KliPeH wrote:
Sat Oct 27, 2018 9:08 am

Reading the status update I see you've somewhat attempted to tweak bot behavior patterns in order to make them more intelligent and versatile, but there's a pattern I see time and time again which stumps your progress and hasn't been getting any attention. It's been problematic for me ever since I started using bots in my games and that is especially true for 1-player-with-7-bots lobbies. It is prevalent in your video too which is why I'm bringing this up now.
If you want more of a challenge consider setting the bots in the same team and play against them. The bots will NEVER be as good as a real player. Just to clarify: The bots will NOT react to the gameplay as a real player or how you would expect a real player to perform. Bots are stupid! What you're talking about is to allow the bots to analyze some high level abstraction of the current combat situation - and based on this decide what's the "better" option at the moment. This is super hard and time consuming to even get working at some level and isn't part of our bot design philosophy.

You have the bot feature list available to read in the beta section and there you can see we have plans on some of the things you have mentioned but once again: Bots will NOT react like real players that can decide to run away and grab an item that might help them win (or camp/hide behind some barrel or try to make it a draw at all cost...)

We have designed the bots to target a close opponent and try to eliminate them with the tools/items available nearby. They might grab a weapon on the way but other than that they will keep getting closer to their target. Two bots like this will gravitate towards each other and start brawling if they don't have any items nearby. They will not run across the map just to get a better item as that goes against our design philosophy for the bots. This is OK.

Don't get any unrealistic expectations on the bots that they will react like real players - because they will never do that. The bots are stupid compared to a real player. They will keep attacking you even if that's a certain death for them. This is OK.
1 x
Gurt

User avatar
KliPeH
Moderator
Moderator
Posts: 914
Joined: Sat Mar 19, 2016 3:03 pm
Title: [happy moth noises]
SFD Account: KliPeH
Started SFD: Pre-Alpha 1.4.2
Gender:
Contact:

Post by KliPeH » Sat Oct 27, 2018 10:37 am

Gurt wrote:
Sat Oct 27, 2018 10:09 am
[...] What you're talking about is to allow the bots to analyze some high level abstraction of the current combat situation - and based on this decide what's the "better" option at the moment. This is super hard and time consuming to even get working at some level and isn't part of our bot design philosophy.
All I would require of them is to look at their weapon slots to see what equipment they are missing, then to decide - perhaps even at random, for max unpredictability - to stop fighting and make a bee-line towards a crate that contains a weapon type matching that of an empty slot. I'm not asking them to decide which weapon would suit the opponent's health and environment best, just to refill weapon slots on occasion. That way they're not left completely defenseless when I get to them. They can already path-find their way to weapon crates so I don't see why that would be a problem; it certainly wouldn't make them weaker if the chance to stop fighting is low, and even more so if you make them do this only when within certain proximity to other combatants or any other complex calculations of this sort.

I can't really argue with "this isn't on our to-do list" but consider the fact that in order to make the bots more difficult to deal with you're proposing to turn all the odds in their favor, i.e putting all of them on the same team and trying to fight them that way. No, I'm not going to do that, of course I'm going to lose against multiple robotic input-perfect melee patterns and marksman-like aim. What I do want is a fair fight - bots utilizing all the resources to their advantage and fighting me the way I fight them, and also fighting their own kin that way. Fighting brain-dead brawlers is no fun once you realize they kill themselves so fast and in such a reliable fashion there's practically nothing left for you to do. It looks cool in the trailers because player input isn't required, but once you win against the unarmed last-standing 5-HP'ed bot 7 rounds in a row with all weapon slots filled to the brim it stops being an enjoyable experience.

Sure they're never going to think like real players do but you can strive to improve the AI to that extent. I'm not trying to bash on the AI's performance here, it's still very fun to play against it, I just find it hard to accept the fact the bots will always be inferior "because they're bots". You can program them to play better.
0 x
 
Image

User avatar
Sh4d0w
Superfighter
Superfighter
Posts: 279
Joined: Sat Mar 19, 2016 11:06 pm
Title: the xD guy
SFD Alias: too many aliases
Started SFD: before sfd
Location: in a distant land
Gender:
Age: 22

Post by Sh4d0w » Sat Oct 27, 2018 10:39 am

KliPeH wrote:
Sat Oct 27, 2018 9:08 am
Gurt wrote:
Thu Oct 25, 2018 8:09 pm
[...]

If everything goes well we might be able to release the game mid November or late November, which I want as some sort of ultimate deadline.

[...]
My birthday is late November so that would be a nice birthday gift indeed! Don't let me down Gurt :heart:
If the game releases in mid November, this is going to be a nice birthday gift for me instead. Oh, well, I'm just going to wait.
Gurt wrote:
Thu Oct 25, 2018 8:09 pm
If you want more of a challenge consider setting the bots in the same team and play against them. The bots will NEVER be as good as a real player. Just to clarify: The bots will NOT react to the gameplay as a real player or how you would expect a real player to perform. Bots are stupid! What you're talking about is to allow the bots to analyze some high level abstraction of the current combat situation - and based on this decide what's the "better" option at the moment. This is super hard and time consuming to even get working at some level and isn't part of our bot design philosophy.

You have the bot feature list available to read in the beta section and there you can see we have plans on some of the things you have mentioned but once again: Bots will NOT react like real players that can decide to run away and grab an item that might help them win (or camp/hide behind some barrel or try to make it a draw at all cost...)

We have designed the bots to target a close opponent and try to eliminate them with the tools/items available nearby. They might grab a weapon on the way but other than that they will keep getting closer to their target. Two bots like this will gravitate towards each other and start brawling if they don't have any items nearby. They will not run across the map just to get a better item as that goes against our design philosophy for the bots. This is OK.

Don't get any unrealistic expectations on the bots that they will react like real players - because they will never do that. The bots are stupid compared to a real player. They will keep attacking you even if that's a certain death for them. This is OK.
That is completely understandable for me. I also did a 1 v 7 challenge a few times and I must say - it is pretty difficult, possibly harder than in Superfighters. Bots alone aren't really that of a hussle, but if they team up, they can be dangerous. Getting them into the same team means you're probably going to die. A lot.

I'm pretty content about the AI in the present. Sure, it does some mistakes, but it really has some devastating potential if ignored. It definitely has an accuracy of a sniper and it is challenging in melee, probably as challenging as the professional players, I'm just saying.

I guess I'll just give those bots a couple of books and they're going to outsmart even the likes of Albert Einstein.
0 x
Image

User avatar
Gurt
Lead Programmer
Lead Programmer
Posts: 1884
Joined: Sun Feb 28, 2016 3:22 pm
Title: Lead programmer
Started SFD: Made it!
Location: Sweden
Gender:
Age: 34

Post by Gurt » Sat Oct 27, 2018 3:10 pm

KliPeH wrote:What I do want is a fair fight - bots utilizing all the resources to their advantage and fighting me the way I fight them...

I can't stress this enough: Bots will not play like real players. Bots are not supposed to be replacements for real players. Bots will always be inferior over players when it comes to tactical thinking - because they're bots!. We still have planned work for the bots for at least another half year to come (perhaps they will provide more of a challenge you seek by then but time till tell). We simply can't just tweak a few things for a few hours and have bots randomly do stuff a player would do in certain situations. I think you underestimate the challenge and complexity to create bots in any game.

If you want a fair fight, consider playing against someone equal to your skill level. Perhaps other real players. Consider playing in teams with bots against other bots (or just you vs one bot). Not all games have to be free-for-all games where you wait out the bots to kill themselves which is what they will do in a free-for-all game if you're absent from the fights (until you're the last one remaining). It's up to you how you want to setup the games.
2 x
Gurt

User avatar
JakSparro98
Superfighter
Superfighter
Posts: 530
Joined: Fri Jul 15, 2016 7:56 pm
Started SFD: PreAlpha 1.0.5
Location: Rome, Italy
Gender:
Age: 25

Post by JakSparro98 » Sat Oct 27, 2018 5:57 pm

I would not neglect the scripting potential to this kind of things, if something is not in the developer's vision, also called to-do list, that's fine, focusing on a single particular aspect for a considerable amount of time it's not something a duo group can do with a game of this caliber, they need to balance work on all fronts, I would say "too many players, too few hands to please all".

But it could be wise to open a way in favor of these people and suggestions, think of a mod in a game, if it's considered good to add in the vanilla game it will be moved to the game by its devs as standalone feature, but the initial work was started by a third party, maybe because the idea wasn't taken into account due for other development priorities and subsequently was revised, here in SFD we do not have mods but a sort of analogy, if only the same building blocks used for developing the bots were available for scripting (pathfinding, raycasting, etc.) everyone (capable of scripting) could be free to re-imagine his ideal bot.
1 x

Post Reply