XCOM 2 once again proved that Firaxis is one of the few companies that can make good turn-based strategies. Many are still surprised that a genre that was predicted to die at the beginning of the 2000s can look better than its competitors and create surprisingly dynamic situations with no need to rush anywhere.
But go to any discussion of XCOM, and you will absolutely find comments or funny pictures with the same problem – random.
Typical situation in XCOM2. The soldiers stand point blank, but the chance of a miss is surprisingly high for such a situation
———————————————————————————————————————–
85% hit? Most commanders will say that his soldier will miss with such. 100%? You will still laugh, but there are documented cases of misses with such chances.
Not only XCOM: EU and XCOM 2 received a barrage of criticism: someone will say a bad word about the good strategy Battle for Wesnoth about its combat system, which is entirely based on the same hit percentages.
Where does such hatred of randomness come from, how does it work and why is it needed?? And, most importantly, how do we live with it?? Let’s figure it out.
When we talk about random number generators, there are exactly two stools that are divided into many stools: hardware generators and pseudo-random number generators.
In the first case, the device generates a sequence of random numbers based on measured, chaotically changing parameters of the ongoing physical process. Take a simple six-sided cube – you have a simple generator. Take a radioactive substance, a Geiger counter, and convert its crackling sounds into digital signals – you’ll get a little advanced. The advantages of any of the above methods are the guarantee of complete unpredictability. The disadvantage is the calculation time and some cumbersomeness.
The Dungeon & Dragons board games use twenty-sided dice because six is too few for the many possibilities that can happen.
———————————————————————————————————————–
In the second case (exactly our case), specially written algorithms are used that generate a sequence of numbers. They are called pseudorandom precisely for this method of obtaining numbers. The algorithms are predetermined and their sequence purely statistically cannot be random. Which, however, is not required, because in practice it is enough for this algorithm to pass at least most of the tests for randomness. And the speed of these algorithms makes them much more profitable than any true generators of random results.
Most people expect that the article will focus on generators that are not working, bringing trouble to players with their bugs. And I hasten to please you – no. Moreover, within the framework of the article we will not talk about them at all. It is too difficult to establish the presence of a bug in a random game, unlike the “bugs” that we will examine. Yes, and errors in the style of the 1956 bug from Medieval: Total War – Viking invasion are the subject of a separate and fun article.
The joke is that if the randomization is made fair, the players will suffer even more. The situation can be easily understood using the example of the long-suffering XCOM 2. Let’s assume you’re playing on the most basic difficulty level. The chance of hitting the enemy is 85%, and you miss. The next fighter is standing nearby, and according to the numbers his chance is still the same 85%. But only in numbers. In fact, due to the difficulty level, a hidden 10% chance of success is added to your chance due to a previous failure. As a result, the next shot goes with 95%. Well, if the gods of Chaos are not on your side again, then the third one will be immediately 100%, due to the accumulating bonus.
The most interesting thing is that in XCOM 2 these hidden bonuses continue until the very last difficulty level. That’s right, only on Legend, the last of the 4 difficulty levels, did the developers make the randomness in the game fair for everyone.
Now remember all these complaints about unfair odds, about how cross-armed soldiers miss 99%, and think how this fits with all these concessions from the developer? And how bitterly Yaroslavna would have cried even if there had not been such a head start!
It’s not just XCOM https://cryptoslotscasino.co.uk/ that has fallen victim to player criticism. The excellent strategy game Battle for Wesnoth has come under a lot of criticism for its similar combat system: the chance to hit depends on the positioning of soldiers and blind luck. The great Korean random on weapon sharpening has become a full-fledged deity of the gaming pantheon due to its unpredictability, and World of Tanks players constantly write about his younger brother – the Belarusian random.
In "Vesnota" the chance of hitting a soldier depends on the type of terrain where he is standing, and his personal hide and seek skills in this terrain. And the damage increases or decreases depending on the time of day.
———————————————————————————————————————–
Where does the hatred of interest come from?? Since we have already agreed to assume that the random number generator in the games under analysis works conditionally correctly, there are several reasons:
1. Psychological — it’s much easier for a player to remember an offensive trick from randomness than the rarest 1% luck.People are used to taking success for granted. And mistakes are like cruel injustice.
2. Mathematical. Yes, 50% to hit does not mean that every second hit will be successful. Even 4 misses in a row can happen, because the chances are 1/16 =! 0. But tell me honestly, you will remember this little thing after four misses, or you will simply hate the game?
As we can see, the whole problem with randomness in games lies in the players. The player is a gentle creature and he wants to win, and the element of chance adds unpredictability to the game. And unpredictability is a great opportunity to lose.
The flight of bullet tracers is just a clever illusion – the chances of a successful hit are clearly visible, and in the game it is easy to notice that bullets fly into the target only when the dice say okay to hit.
———————————————————————————————————————–
It turns out that the main task of the developers is to make the randomness in their games as unfair as possible – in favor of the player, of course. In the old X-com there was no difference in the game where to generate aliens – far from the map, or right next to the ramp. During the capture of large battleships, everything often ended before it even began – a good sectoid with a blaster-launcher stands right next to the ramp and ends this episode of the war with one spit. It would be hard to call this a bug, but they tried to fix the problem in any successors and continuations of the games. It would seem that this is a completely fair chance for the AI to win, but it annoys the player because it leaves him helpless! And it doesn’t matter that the basis of tactics in any X-com is to render the enemy helpless and defeat him – he cannot use the same methods unless the developer wants to kill interest.
Of course, trying to edit the numbers is only suitable for games where the hit percentage is always before our eyes. Of course, hidden rolls of invisible dice don’t just happen in strategy games – in Knight of the Old Republic and D&D games, each hit is also calculated with complex rolls for hits, damage and evasion. In Dungeon Crowlers like Diablo, percentages are present in the form of equipment bonuses. Why are there no complaints against them?? Well, mostly because we don’t see these cubes! Most players barely even think about why and how characters hit, and the dynamic nature of the game in pseudo-real time masks a series of misses: it’s all too easy for you to miss them in the rush of using numerous features.
But sometimes a crappy balance can’t create the illusion of honest randomness. In MechWarrior IV, in one of the missions we are suddenly given an alarm and required to move as quickly as possible to help an ally who has been ambushed. How is this done in normal games?? The timer starts, and as soon as we approach the desired point, both the target and the enemies attacking it appear there. But in this game they did everything honestly – the allies were really ambushed, and they were being shot all the time while we were running towards them. Do you feel what I’m getting at?? Exactly – if you are unlucky, they will die faster than the player even has time to approach. Or they’ll die right when you get there. AI misses and accuracy in most games are calculated using the same dice, and in this case, although they wanted to make a realistic situation, they only got a merciless test of your luck.
Xenonauts also has an example of lousy balance. When ordering recruits, they try to take only those who have good parameters. What to do if all options have only terrible stats? Buy disabled people who can’t shoot! Too large a difference between the minimum and maximum in the generated parameters often leads to such problems.
———————————————————————————————————————–
What happens in the end? Where the developers hid the cubes well, there are no complaints. Where everything lies on the surface there is a lot of screaming. The only thing is that we have something to dump our own failures on? Except in those cases where the dice are actually acting dishonestly and creating a false challenge – yes. XCOM: EU has an excellent random number generator that produces adequate results. But since his work is not covered by anything, the players get their own scapegoat.
The developers understand very well that the player does not want to lose, but wants to enjoy winning. And if this requires disguising the operation of the game systems or making the calculation of probabilities a little unfair in his favor, they are happy to start using a few of their dishonest tricks, for which we will kiss them on both cheeks.
But, before looking at the ways developers struggle with honesty, let’s look at what kind of randomness generally happens in games.
Let’s immediately define the terminology – random is when we cannot influence the final result. The maximum we can do is change its probability upward, but the probability of complete failure remains no matter what, and we will not influence this in any way. When we can lie behind a hillock and aim a rifle for a long time to make a successful shot – this is not randomness. When we order a soldier to lie behind a hillock because it is +10% to accuracy, aim for +15% and wait for the weather for an additional 5% to ultimately knock out 99% – this is random. It is possible to create a local situation of guaranteed success, but most of the time we enjoy random results.
The first and most nerve-wracking type of chance is all or nothing. This is where the notorious Korean rand grows. When sharpening a weapon to a plus level in most MMOs, you are given two options: either the improvement will be established, or all progress will disappear altogether. You can’t influence the process in any way, all there is is a chance on the screen and simple luck.
Next comes the XCOM and Wesnoth model – the positioning of fighters on the map and/or special abilities can influence the success of a hit, but in most cases the chance of a hit will still remain in the hands of Fortune. It may already be possible to influence the enemy, but again in the simplest things like increasing the chance of hitting or damage.
Relatively “fair” random is used in more serious games like Wargame: Red Dragon, Warhammer 40,000: Armageddon or Panzer Corps. Not only does the number of variables that can be influenced increase, but also the difference in the result obtained depends on the size of the squad, the strength of its weapons, morale and support. The Imperial Guard squad in Armageddon consists of 30 people, and each individual soldier has a different chance of hitting, which is why it is easy to get into situations where a successful volley from a crowd of weak soldiers is more effective than the powerful volleys from just three tanks.
We were lucky – each of the 25 flamethrowers found their victim. Large units in Armageddon are bad because strong but small troops will destroy them for too long. 5 tanks will not be able to make more than 5 corpses per salvo.
———————————————————————————————————————–
It is easy to see that the degree of “normality” of randomness is directly proportional to the number of ways in which it can be influenced. But this is just one of the ways developers can “fix” dishonest randomness.
What honest and not so honest methods do developers use to deceive the player and make the game more pliable for him??
The first one has already been described a little higher: hidden bonuses and penalties. In some small hidden variable, +10% is written for each miss, and it turns out that even the most unlucky person can score 100%. Most often, the enemy is given similar debuffs – that is, reducing the chances of success.
Another way is to artificially limit the game and create certain scripted “accidents”. For example, in Heroes of might and magic V, the ghosts of the Undead castle have an “immaterial” property, thanks to which they can evade all physical attacks with a 50% chance. In the first versions of the game, these were just percentages, and players happily reported about a terrible bug: how is it possible for a ghost to dodge and four times in a row! The developers got tired of listening to this, and soon the “bug” was fixed: now only one of the two attacks was strictly avoided. It seems that everything remained in place, but the element of randomness was replaced by predictability.
But any developer’s tricks will always come up against the fact that there will still be a chance of a mistake or failure. What should we, the players, do, especially if good Uncle Sid also blocks “savescam” – the ability to load and re-roll the dice??
Although this is only slightly related to the topic, it’s worth mentioning since we started: more often than not, AI is specifically limited. This is very common in various serious wargames: the enemy follows a given scenario and you just have to cope with this situation. In XCOM2, the sectoid will most often use useless psionics even when there is no one to support with it, and the Lancers will run into close combat, spitting on any instinct of self-preservation. This is the removal of the element of randomness already at the global level. Most often, we do not have the opportunity to read the enemy’s documents and find out what he is up to, and such behavior patterns allow an experienced player to immediately understand, without unnecessary problems, how the enemy will move and how to respond to it. Just like in chess, we must know how other people’s pieces will move.
But no matter how hard the developers try for us, it’s still up to us to deal with what they did. What can a player do when the fate of the entire battle depends on the gods of Chaos??
First of all, randomness teaches us to always leave room for a backup option. The first one misses, the second one hits. The second one misses, the third one hits. Statistics can fight on our side, let every failure increase the chance of success
The second follows from the first – leave room for luck! Yes, 95% misses happen – with the same frequency as 25% hits. As an avid fan of iron-man in XCOM, I can confirm that a successful critical shot on a mouton in full cover has saved the team more than once or twice. If there is an opportunity, without much risk, to make an almost hopeless blow – do it. It will hit well, it won’t hit, no big deal.
The chances of a hit were low, but a successful critical shot dramatically changed the situation on the battlefield in our favor.
———————————————————————————————————————–
And the third relates exclusively to the psychological side. Stop complaining. Yes, failures happen. No, don’t memorize them. Take it into account, but don’t treat it as something particularly scary. The unpredictability of interest rates is designed to ruin the best plans.
In conclusion, I can readily say that the best random is the one that is missing. Yes, it is impossible to get rid of it completely (even the result of the flight of an arrow in Mount and Blade is determined by the roll of many invisible dice, even though we aim as usual), but the better it is camouflaged and the more we can influence it, the better. Dice rolls in the style of board games from XCOM – but this is still primarily laziness of the developers. Cubes can do the simplest mechanics, but for a computer it is too ridiculous. Modern technologies allow us, using procedural generation, to create entire galaxies in Spore and worlds in Minecraft, which can produce quite beautiful landscapes. Why are developers embarrassed to work on beautiful and relatively fair ballistics for games?? There are many reasons.
In any case, my opinion is that the capabilities of the computer should be used to their full potential. Misses at point-blank range are 85% offensive not because they happen, but because for some reason they exist, despite the laws of logic and common sense.