SpongeBob NoPants

The Forest

2018, PC, Single Player (Steam/ Proton on Linux)

Bad combat is aplenty in the video game industry. Depending on the perspective, it can often be masked as “difficult” game that “noobs” don’t get. With support of a few keyboard warriors rushing for their 5 mins of self-righteousness by downplaying any neutral gamers’ concerns, the devs do not see why they need to “learn” how to make a good, engaging game.

The cheapest way to make a game “difficult” without understanding how those Castlevanias and Elden Rings became fan favorites, is to make the enemies spongy.

Basic Game Design 101: in every turn of combat, there is a damage by the player to the enemy, and vice versa.

Skyrim had the simplest way of explaining it, by using the ratio of enemy-induced damage to player-induced damage. A value of 1 meant the player and enemy will both induce same damage to each other, whereas a ratio higher than 1 will make the game more difficult, lower than 1 would make it easier.

Designing this ratio for different difficulty options is not exactly rocket science. At the very least, a game should follow the 1:1 ratio in what they define “normal” mode (not hard, not easy). Of course, narcissist developers often consider their game to be “above the fold”, and if it helps them get good sleep at night, sure they may put this ratio higher in the normal mode itself, but in that case the least we can expect is a lower mode that “does” have the unit ratio, especially if there is no learning curve like FromSoftware games.

The Forest, an open world “tactical” survival game, has neither. It has 4 difficulty modes – Peaceful (no enemies), Normal (easiest mode to actually have enemies), Hard (more difficult enemies), and Hard Survival (more difficult survival added to the previous option). So people wanting to play “with” enemies have to select at least “normal”. What if now, Endnight games the developer, make the ratio ridiculously high so that an enemy can kill you in 3 hits, but you need at least 10? A fan-made wiki promptly defines the combat as “aims to be realistic”! Because as we all know, realistic combat must have this ratio higher than 1. As a neutral gamer, if you die on the 10th day, you’re supposed to feel proud and compelled to put more effort to “learn” the game, “strategize” your moves, “upgrade” your weapons, or even follow the “saintly” advice of running away from combat altogether until you become stronger!

Your time has no value. Your enjoyment does not matter. Only matters is your money that you’ll spend in the blink on their “Early Access” game, which they will finish after 2-4 years. They want you to pay* for that “demo” so that they can earn even before offering the complete product, and the only way to make you do so is to build an aura around that you worship. A false sense of non-existent learning curve needs to be induced upon you. And did I mention that this is the only game existing in the world so if you don’t learn, you’re missing out?

What if your katana cannot block but a cannibal with her bare hands can block fine?. Come’on, cannibals are “hardened” right? That is why when you hit them with a burning weapon, the cloth immediately burns out as if you hit a water sack. And even if they burn full body and get charred, they are still hardened enough to come at and smack you for the next 10 mins. What if you spend mana while descending a rope but you don’t spend a dime of it while rowing your raft with 2 hands even in a storm? Surely you can forgive because it is “aiming to be realistic”? Or so much you can carry that you spread them out like a bedouin vendor at the push of a button but you cannot carry more than 5-6 tablet case medkits or 2 Aloe-Vera leaves?

The design doesn’t make sense in so many areas. Why would you build a gazebo or a fireplace (that is useless for anything) instead of finding your son?

What is the point of building a theme park when you can just hide from the cannibals or make faces at them from a distance at sea?

The story does not turn the game into tower defense at any point. Why do skulls not have any impact on keeping the natives away, since they already use it to depict their residential boundaries?

How does meat become edible after drying for just a day? If drying works so fast (perhaps we are in an equatorial island where the Sun is hottest), why is there no concept of dried leaves on the ground that we could have used to extend the fire? Or perhaps use the sticks lying around everywhere for that purpose? And speaking of short-life of the fires, why is there not a fire extinguishing option, if not with water, at least by stomping on it? You cannot sleep arbitrarily. I get that, and in a way prevents misuse of food. But what I don’t get is why we cannot pass the time? Say wait for some time when seated on a sofa in the yacht that is safe enough?

Darkness is stupidly dark. Even if there is a moon shining brightly and human eyes usually get adjusted to darkness, we cannot see jack unless one of the color gradient options are used, that too only marginally better. Inside a cave too, human eye adaptation does not work. It is not an RTX implementation. It is just that lights are set to disabled at places as a lazy way of thumping their chest – look how realistic our game is! The terrain design too, is frustrating. There are gradually sloping cliffs that you cannot climb, even if you can jump that much height from one platform you built to another. There is not much variation in the flora or fauna, so if it was meant to be an equatorial paradise, its residents certainly don’t fit the ecosystem. Maps and documents have a weird angle and do not fill the entire screen. It was quite a few days before I could actually see what was in the map after I had collected it.

Did you know that you cannot make the crashed plane your first home? About 10 rows of 3+3 seats which you cannot rest/ sleep on, but have to build a bed or a shelter to save. You cannot build a wall at the aircraft’s opening, which would have made it a really secure base giving you enough time to expand it if you want. You cannot make a makeshift blanket from so many unpicked clothes in those luggage cases, even if you can make armor from much smaller animal skins by the same principle. You cannot yank one of the protruding iron rods and make it your first weapon. Because you are playing a “realistic” survival game, where realism is defined only by the devs and their cultist disciples.

“They’ll fix it, perhaps already working on it as we speak” you say. Yes, being in Early Access for so long gives them the advantage of not having to bother with accountability. That is what video games have become. Release the first version as Early Access, and perpetually keep patching after getting feedback on even the basic stuff. As long as the developer “listens” to feedback, why would they need to get it right the first time?

Because they did not learn to make a game by following the most basic lesson in gaming – playing it. It is their “learning-by-trial-and-error” experiment which we the customers should sponsor. The Overwhelmingly Positive reviews do not paint the full picture, because combat has been tweaked several times in this game from Early Access days to full release. They provide cheats though, and this might be one of only a few games where you can still get Steam Achievements with them enabled.

The worst thing about this game’s design is that the enemies on the island spawn effectively only after the 7th day, when you are first exposed to the piece of crap combat, by the time you would have passed Steam’s return window and cannot refund. Intelligent, isn’t it?

*Thankfully, I resisted the temptation and bought the game only recently at 75% or so discount. I don’t buy early access games neither do I ever pre-order. But that is just me.

Pros

  • Excellent performance on Linux. Steady 60 FPS with minimal frame drops only when loading new areas
  • Visual fidelity at par with some AAA games. One of the first games to use Unity Engine
  • Environment is designed to deforest and build upon. More versatile than Fallout 4
  • Enemy jump scares and horror moments. That demonic laughter!

Cons

  • Spongy enemies.You’re not though
  • Badly designed melee combat. Enemies can block. You, on the other hand, ha ha ha!
  • Hilarious mutant design. Stupid-legged animals
  • Forced to build new shelter even with usable aircraft. You’re playing “their” game, not yours
  • Grindy resource gathering in single player mode. All sins forgiven with multiplayer support, right?

6/10

Rants on Dirt Rally AI

Dirt Rally 

Single Player, PC (Windows/ Linux), 2015

No matter how much “realism” (what is realism? Is it same as making the game difficult?) exists in a driving simulator, it is still a video game where mistakes are pardoned and accidents forgotten with a restart. It is important to have the distinction between a video game and real life.

In reality, there is no second chance. If I am allowed to restart a race after I have crashed, then irrespective of how realistic the handling and physics are, it “is” unreal.

It is naive for such games to try to outsmart other similar games by synthetically reducing AI stage times to the point of being not achievable and not including a difficulty setting. Case in point – Dirt Rally. I can complete a stage carefully without crashing even once and maintaining fairly decent speed wherever I can, but still at the end I find myself 30-40 seconds behind. And it is not that all AI have that timing – only 1 or 2 at the top. There is a significant delta after that. The timings are inconsistent even without my participation. The randomness that they have put only varies the timings within a tolerance, but the lead split times remain nonetheless*.

On the internet, fanboys of the game feel proud of this feature. Their justification is that the more hours you put, the better your skills would be, and it is ok to win the first competition after 100 or so hours if you are really not that “great”. They somehow have an issue with providing the customer with choice when it is possible to do so without added cost. If you ask for this choice, you are labeled a console peasant, arcade noob, etc.

What their limited maturity prevents them from comprehending is that even after putting 100+ hours and winning perhaps all competitions in the career mode, I would still not drive a car in real life using a gamepad or a gaming wheel in the comfort of my home. I will only be good at a “video game”. No matter how many hours you put in the game and how many stages you win, it will not qualify you for any real life rally stage.

Many games have a “points” system, where you are awarded more for playing at highest difficulty. Even Dirt Rally offers more “credits” at higher difficulty, and the self-righteous “experts” can feel proud of it. But the “I-put-100-hours-to-learn-so-you-should-too-else-this-is-not-for-you” arguments are BS.

*Let us dive a bit deeper into the AI mechanics of Dirt Rally. Here is an excerpt on Dirt Rally AI from Steam Community:

“In short – the fastest time the AI can achieve was set according to a time set by a real human driver from Codemasters. The AI can never be any faster than he was.

“So the simple truth is – if you can’t keep up with the AI, you’re just not on the level of the top guys. You either need to get better or accept the fact that you’re not fast enough.

“BTW, if you want to see a rally game that literally just “spits out a time” for the AI drivers, try WRC 5 or 6. The times of the AI drivers are completely made up in those games based on just how many mistakes you’ve made during the stage. “

https://steamcommunity.com/app/310560/discussions/0/133258092242139442/

If this is true, then Codemasters got AI all backwards. They do not know what AI means.

Artificial Intelligence is NOT random sampling from data. It is about “learning” from data.

And in a single player game where challenge should be specific to that player, the “data” is how the player is doing, not how others have done.

Why would a single player care about how good a Codemasters employee is in the game? He could go online for such challenges, right? WRC actually got it correct when it comes to the definition of AI. They implemented actual AI, because they consider how “you” are playing. In WRC then it would be possible for each player to have a different experience, tailor-made for him – this is true “single-player”.

Also observe that though WRC is made to look like the only one who “spits out the data”, Codemasters also does so. Selecting one timing from a predefined set is also “spitting out the data”, isn’t it?

In a custom championship event I was playing, the topper had the exact same timing (to the millisecond) on a stage in Finland at night and at day (day was a reverse of the same stage). How is this possible if timings were “simulated”?

In the same championship, that topper guy (name was P. Serrano, though obviously it does not matter) came first in each stage at every location. This game truly honors that legendary line from Fallout: New Vegas –

The truth is, the game was rigged from the start

So,

  • Do not code the AI
  • Spit out data nonetheless (select randomly from a list of values around a top human time),
  • Brag “difficulty == more realism” in a blog for the fanboys.

This does not cover up for poor (or non-existent) AI implementation. If WRC has actually “coded” AI to “learn” about the player, then hats off to them.

AI is actually always for single player; if timings are based on random sampling other humans, then it is multiplayer.

Which is why most games are going the multiplayer way, where they can “crowdsource” the AI thereby reducing actual coding costs.

No, this game does not have a steep learning curve.

It just has professional timings (for 1-2 at the top) in a singleplayer race of “entry-level” difficulty. Let that sink in, before inferring anything from the high ratings the game has received. If you are struggling and only the high ratings keep you going (like yes-I-can-also-ace-it-someday), save yourself precious hours and do something more productive. 

“Learning” about the player is what a single player game should do in order to provide a custom challenge to the player. Each player is different, so it does not make sense to expose all of them to the same challenge even when they are not playing online. If I want to compete against a leaderboard, I could go online or use “Community delta” where Codemasters crowdsource the challenge. If all they wanted was to brag about how good their employees are, they could have just created another “Codemasters delta” as a difficulty option. But they had to cover up for the poor design.

Why restrict the player from having a choice? If you are good, play it at a harder difficulty level; dominate the leaderboard with your achievements. If someone else wants to play at a lower difficulty, then why would it reduce the game’s appeal? It is not about coming first always. It is about making that individual player believe that the fastest time is at all possible for him. It is about customer satisfaction. And competitive strategy – how I can keep my customer from moving to my competition.

To sum up, there is no motivation to keep trying to “improve” your timings. No matter how hard you try, with all sorts of controller settings, not a single crash, correct corners, even Scandinavian flicks, end of the race you are still a 2nd or 3rd at best. There is always a P. Serrano at the top, who will win each stage at every location of your championship. Others will commit mistakes, their cars will fault, but one dude is rigged to win, and it is not you.

No wonder Dirt Rally is almost always on Sale in Steam. If the game was that good, it would have sold without the “80% discount”.

</rant>

Pros:

  • Excellent game engine, controls, and feedback
  • Runs smooth on Linux even on old hardware
  • Choice of classic Rally Cars
  • 80% discount almost always on Steam

Cons:

  • Terrible (actually non-existent) AI
  • Cars do not have any notion of weight, so no “sense” of grip while driving
  • Stages not as exhaustive as WRC

7/10

Game Difficulty vs AI

April 2017

In most games, difficulty does not vary the AI. The only things that change are max health and ammo, ammo pickups, and number of enemies spawning. Some difficulty modes may change the enemy health while some others also reduce number of saves.

If there is no actual AI incentive (i.e. the enemies are suddenly more alert and show more realistic behavior), there is no point in playing at higher difficulty modes. A player looking to relax and navigate the map after a log day’s work will find it frustrating far more than it is enjoyable at highest difficulty. Unless of course he is an “expert” in hand-eye coordination and key “mashing” (like the BOOM! Headshot guy).

However, if the AI is really next-gen and enemies become more aware/ alert with higher difficulty, then certainly there is a case to play in such modes (example: Far Cry series, excluding the enemies-can-now-shoot-through-walls epic fail). In such cases the game may give the player choice of whether to tweak health/ ammo variables too, or just AI.

A truly next-gen AI is one which “learns” how good the player is. But when mostly the objective is to make a more “shiny” copy rather than redesign, that would be too much to ask I guess.

It is really fun to play against advanced AI, but hard-coded AI does not make it “advanced”. Just a half-hearted attempt if all it does is reduce health/ ammo/ saves etc. A game is a game; to some it is about relaxation. They are not looking for a medal by being an expert in it.