Seriously? No Non-Player Character Survivors in Single player mode?
6 6

95 posts in this topic

Recommended Posts

1 hour ago, ughughugh said:

I know this game is indefinetely more complex than Go, but Alpha Go (that AI that beat Lee Sedong, the world champion) didn't win through brute computational power, but by learning on its own.

Having taken numerous courses on AI and machine learning, your not even close.

Yes the 'rules' of PZ are more complex then Go, but so are the rules of chess. Alpha Go didn't brute force the problem because you can't brute force it. The number of possible combination of moves is insane. Its also damn hard to measure your success rate in Go until late in the game, which is where the real problem with ML and Go comes in.

PZ on the other hand, is relatively simple from a Machine Learning perspective... Am I still alive? How long have I been alive?  Assign a reward score for each day of survival.

Did the AI get injured? Lose points, learn not to do that.

Hungry or thirsty and no food/water? Lose points, learn not to do that.

Got chased by a massive horde? Lose points, learn to avoid that happening.

When something negative happens, subtract points. When the only real goal is survival and the AI has been told to try and maximize its score, it will learn.

Go on the other hand does not fit well into the score/reward system for machine learning. The effects of a single move in the beginning of the game can have drastic consequences at the end game. This is why go was considered a impossible problem for AI, until Alpha came along ;)

Share this post


Link to post
Share on other sites
2 minutes ago, Fenris_Wolf said:

Having taken numerous courses on AI and machine learning, your not even close.

Yes the 'rules' of PZ are more complex then Go, but so are the rules of chess. Alpha Go didn't brute force the problem because you can't brute force it. The number of possible combination of moves is insane. Its also damn hard to measure your success rate in Go until late in the game, which is where the real problem with ML and Go comes in.

PZ on the other hand, is relatively simple from a Machine Learning perspective... Am I still alive? How long have I been alive?  Assign a reward score for each day of survival.

Did the AI get injured? Lose points, learn not to do that.

Hungry or thirsty and no food/water? Lose points, learn not to do that.

Got chased by a massive horde? Lose points, learn to avoid that happening.

When something negative happens, subtract points. When the only real goal is survival and the AI has been told to try and maximize its score, it will learn.

Go on the other hand does not fit well into the score/reward system for machine learning. The effects of a single move in the beginning of the game can have drastic consequences at the end game. This is why go was considered a impossible problem for AI, until Alpha came along ;)


Just gonna drop a thought here, If someone wants to see machine learning how to play video games go watch "MarI/O - AI playing Super Mario Bros" stream on YouTube, look at the description of how long it took it to beat certain levels via basic trial/error.

An AI working that way would probably become sentient long before it could play PZ believably enough to pass off as your average video game NPC. Likely sometime around 2099 and in all likely hood the "2100 IS THE END" craze would be caused by everyone freaking out about that Video Game AI thing figuring out how to make zombies so it can test it's survival skillz in the real world.

Share this post


Link to post
Share on other sites
3 minutes ago, Svarog said:

An AI working that way would probably become sentient long before it could play PZ believably enough to pass off as your average video game NPC.

That's really going to depend on a number of factors..the cpu power thrown into the problem, and would also require a customized PZ engine that runs insanely faster so the AI learning could do years of game time in fractions of a second...watching something learn Mario is limited by the speed the game is running at. But I wasn't going to comment on the actual feasibility of it, just the difficulty of the PZ vs Go problem itself :lol:

Share this post


Link to post
Share on other sites
37 minutes ago, Fenris_Wolf said:

That's really going to depend on a number of factors..the cpu power thrown into the problem, and would also require a customized PZ engine that runs insanely faster so the AI learning could do years of game time in fractions of a second...watching something learn Mario is limited by the speed the game is running at. But I wasn't going to comment on the actual feasibility of it, just the difficulty of the PZ vs Go problem itself :lol:


Me I was talking about believability. Honestly, the way some mondoids sell NPCs, the things would have to be sentient at this point to get close to the amount of hype around them. I've already dismissed my imagination and wrote the thing off as deceased because it's impossible they'll be that good, I'll enjoy what I get.

Other than that I'm just laughing. I get kicks out of reading those "Where are muh NPCs" threads where people get triggered because "I spent 15€ on this game!" Where I spend that much on shitty microwave dinners and cigarretes in less than a week.

PZ is a fking great game as is, I'd gladly cease smoking and eating for 3 full days to buy it again in the state it is in. Gave me way more entertainment than many other games (hundreds if not over a thousand hours worth of it, don't mind my steam, I played off steam mostly) and it taught me how to code because I got hooked on modding so I basically got free basic coding class out of it.

And I bet those same people that bitch about losing 15€ then go and buy the next big thing for full AAA price on release, play it for standard 8-16 hours and forget it existed.

So many other things to do than wait for NPCs to come... but no, there is a thread about "Wer muh NPC" every week. And I wish I could really get that angry about something when all I really have as an argument is 15€ at the most and I can't even really accuse the devs of not doing anything because they clearly are doing something, especially when I myself would have ditched the Project out of burn out years ago.

It's just funny to me.

And no, TIS didn't pay me off to shill for them, I'm a russian bot, we shill randomly just to annoy you Американцы.

Edited by Svarog
Less than a week... Ha 15€ is my budget for 2 whole days!

Share this post


Link to post
Share on other sites

I don't wanna micro-manage a bunch of Tamagotchis with issues. 

I can do that on Rimworld and The Sims series already. 

 

On PZ, I wanna fill up my pick-up truck with loot and kill zombies with an axe. Kthx. 

 

 

Share this post


Link to post
Share on other sites
7 minutes ago, Vyn Halcyon said:

I don't wanna micro-manage a bunch of Tamagotchis with issues. 

I can do that on Rimworld and The Sims series already. 

 

On PZ, I wanna fill up my pick-up truck with loot and kill zombies with an axe. Kthx. 

 

 

The originals had a habit of just telling you to f*** off and leave if you tried to micromanage them. Or if a hoarde approached them while they were on guard duty, they'd sometimes bolt, rather than go down swinging.

 

So, much like the idea that NPCs are going to be a complex neural-net rivaling Skynet, I don't think you have too much to worry about there. They just need to be a bit better than what we've gotten before - a little more autonomous and the gateway to scripted events.

 

Then again, who really knows what they'll be at this point.

Share this post


Link to post
Share on other sites
4 hours ago, Vyn Halcyon said:

I don't wanna micro-manage a bunch of Tamagotchis with issues. 

I can do that on Rimworld and The Sims series already. 

 

On PZ, I wanna fill up my pick-up truck with loot and kill zombies with an axe. Kthx. 

 

But why try to micromanage them? Just let them do their own thing, you can still fill your truck with loot and kill zombies with an axe just fine, however throw NPCs at the mix and you have the NPC coming to steal your truck with loot and other NPC trying to kill you for your axe or something similar.

 

In their own they will add more difficulty to the game because you will need to rush to loot the first days.

Share this post


Link to post
Share on other sites

 

19 hours ago, ChildrenCruelty said:

Things changed? I played some time in Single Player-Modus and never met any NPCs. I often hear shootings, shouting etc. but never see anything.

 

In the official build there are still no NPCs, but you can add a mod called "Super Survivors" (or something like that). I tried it a short time, but I like it better without NPCs.

Share this post


Link to post
Share on other sites

The Super Survivors! mod is quite good imo. The NPCs can add a lot, whether it's a bunch of survivors that you're trying to keep alive, or hostile bandits shooting at you. Now and again a survivor might have a fire bomb - most of them have guns. So, you might be happily coming back from a looting trip to your base and there'll be a raging gun battle against a horde of zeds w/ buildings on fire. Good times.

Share this post


Link to post
Share on other sites

Look at state of decay 2.   Backed by Microsoft money and three years in development on top of an older title and they still have NPCs that basically stand around at base.   They appear to fiddle around but it's all cosmetic.   Function wise you can ask one to follow you.    That's it.    They abstractly affect the stats of your base and take no actions.   Like cards in a deck game.    The  "enemy" enclaves are three people standing around.   If you get close they shoot at you.    That's it.       The bar is very low and I expect Zomboid will clear it and retain the crown.     With a solid NPC mechanic the game could be completely unique.

 

With just abstracted off camera NPC actions  SOD 2 could be vastly improved.   But no.

Edited by feral_donkey

Share this post


Link to post
Share on other sites

I often wonder after reading these posts why people can't pick themselves up and move on if they're not happy with a product.

 

Who stands around for 7 years at the Spiffo's counter after their cheeseburger wasn't made to their liking. Who sits in the Spiffo's restaurant after the manager said they'd make improvements, waiting and nit picks each improvement made? Even worse, who goes into the back and rushes the cooks?!

 

I'm not saying no one should eat at Spiffo's, but if you're not happy with the burgers, don't eat there. There are other places to eat, other things to do to pass the time. You can always revisit Spiffo's later.

 

I also wonder why anyone who was mostly happy with a product would leave it a bad review. 80%+ happy, still "do not recommend". Why? 80% happy with this game is more happy than I am with life most of the time and I still recommend life over the alternative 100% of the time.

 

Also I have started to wonder if we should start calling ourselves "Yelpers" and not millenials.

 

/endwondering

Share this post


Link to post
Share on other sites

I remember way back when, when NPCs were in. They were little more than zombies with shotguns. To be honest, I was GLAD when they were removed because they had a tendency to shoot first and ask questions later. The first time you break into a house and have your face blown off by an NPC is cool and all, but the 4th or 5th time is annoying. And if they weren't hyper aggressive, they stood around getting eaten.

If NPCs never come back, i'm cool with that. Honestly, I'd like them to come back, and if they're even half as good as TIS says they will be, i'll be happy. But i bought the game to survive the zombie hordes. I've had fun every time I've played in the last 7 years, and I certainly don't think TIS owes me anything after all this time. I bought the game knowing it was in development, and i knew what that meant.

I have a feeling that a lot of people who are solely focusing on NPCs are those bored with the game, and intentionally forgetting everything the game has, because its not new anymore. They get angry because they're bored with the game - but that's not TIS's fault... no game will hold your attention for years and years, that's just life.

 

Share this post


Link to post
Share on other sites
On ‎2018‎-‎01‎-‎21 at 9:50 AM, DicheBach said:

(...)

This game has been in development since at least 2011. It has an estimated ~875,000 units sold according to SteamSpy (at an estimated median unit price of say . . . $10, that is $8,750,000 in gross revenue since the game was released on Steam, and that would mean ~$6.125 million pre-tax/pre-customs take for Indie Stone after the distributor [Steam] takes their cut . . .).

(...)

There is always one factor people that brings the argument you're bringing always forget. I'm glad however that you considered taxes and other fees that would lessen your rough figure of 6 and a eight million revenu because that is important as well at some point in the maths. Also, you have to take into account the double taxation effect on corporation and dividends as well as integration mechanism that alleviate part of that effect (in effect, the company's profit is taxed yearly, then it's taxed when it comes out as dividends to the shareholders a second time with a usually partial credit to account for the first yearly taxation).

 

One very important factor that people almost always forget in the equation, the one that's almost always the most significant expense during development is : salaries.  People don't work for free full time for years, and we can't assume that they are all shareholders or partner living on the promisse of cash to come.  So there are always very significant yearly puncture to account for salaries in EA titles. In non-EA titles this is part of the development budget that has to be balanced against estimated sales so it will return a profit... sadly this often result in products being shipped before they'd be fully ready due to budget limits, requiring numerous post release patches and in some case DOA products that'll never be finished.

 

But in the case of EA titles where the funding does not come (mostly) from investors, but rather a form of crowd funding, yearly sales have to provide the budget for development.

 

Now, this being said, let's go back to the maths and "forget" other factor like licences, hardware, utility (electricity, water, etc.) and rent.  Let's just take that big chunk of money that are salaries.

 

Let's assume we are talking about a fairly small team of 15 people, for simplification purposes. Let's also assume that each cost roughly 50K/year (including mandatory fringe benefits, I don't know about the US but here part of the taxation from the government is in the form of programs that employers need to contribute to based on their salary mass). That'd be 0.75 millions per year or an estimated 5.25 millions of expenses for the duration you mentioned. Now a bigger team might get things done in a shorter period, but, without financing or investors, that might not always be an option.

 

Even without factoring licences, hardware, utility, rent or yearly taxes, that only leave 0.875 millions over 7 years. In reality, it's probably more of a daily struggle to make ends meet then we'll ever know. Regardless, they keep on adding content at a pace I'll simply not discuss without having real hard data to compare (as opposed to assumptions).

Share this post


Link to post
Share on other sites

One key thing people forget is that the most volume is moved during sales, of which most were 40% off. 

 

The game cost $5 full price for its first three years. It went up to $8 on its steam release. It became $10 a year later and has stayed there since. This is just in usd, ignoring inflation, witch other currencies exhibit on the Steam store.

 

This also ignores that the second largest group that bought the game are Russian, paying considerably less last time I checked.

 

It ignores that the first few years of development had to be funded somehow, which likely means loans and their own personal savings.

 

Finally, it ignores that indie games have become increasingly expensive. It's not uncommon to see companies throwing 10s of millions into a zombie fps and having first week revenues of similar denominations ... 

Share this post


Link to post
Share on other sites
On 6/5/2018 at 9:16 AM, Silverfox said:

I remember way back when, when NPCs were in. They were little more than zombies with shotguns. To be honest, I was GLAD when they were removed because they had a tendency to shoot first and ask questions later. The first time you break into a house and have your face blown off by an NPC is cool and all, but the 4th or 5th time is annoying. And if they weren't hyper aggressive, they stood around getting eaten.

 

Sounds just about right for a PvP server to me :)

 

NPCs are nice, but to be honest playing multiplayer will give you a much more immersive 'other survivors' feel.  No AI will ever be able to truly mimic the experience of having to interact with a real player and try and decide if he's going to trade with you...or blow your face off :)

Share this post


Link to post
Share on other sites

FYI, unless I'm mistaken, OP jumped ship 2 pages ago (right after getting feedback from devs and moderator).

 

I do love to dissect a topic and point out what I feel is lacking, but at this point it seems like we're all discussing in a circle.

 

Keep you energy for the next such topic, make notes and refine your ideas, stay strong friends :)

 

 

Share this post


Link to post
Share on other sites
5 hours ago, Dryke said:

(…) No AI will ever be able to truly mimic the experience of having to interact with a real player and try and decide if he's going to trade with you...or blow your face off :)

In the long run, as to where loyalties lie I'm on you with this (human vs AI)… as to shooting in the face, AI aggressive response might do that quickly enough.

 

The trick in human relations is not the initial response, which is almost always "mild" (from guarded reservation to "ok, let's check these guys out"). The trick is on how it evolves according to human interactions, this does need a lot of fine tuning if it want to boast any sort of : we have a decent AI. Of course, open agression will always be met with one of two things : Fight or Flee. (edit : and I dare say,  if they are smart survivors, odds are they'll flee more often then they'll fight).

Edited by MyTJ

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
6 6