If you've been playing XCOM (either the originals or the reboot series) for any length of time, you've undoubtedly come across things that make you want to hurl your computer into the sun. Missing point blank shotgun blasts, seeing multiple 85%+ shots fail in a row, taking one extra step and activating another pod, watching as your flanking shots fail to connect while the aliens crit you in high cover, the list goes on and on.

But what if I were to tell you that not only is the game fair, it's actually skewed in YOUR favor. Hard to believe, right?

Enemy Unknown/Within had hidden modifiers in place to help you along, and XCOM 2 is no different. I'm going to warn you ahead of time; some of what you are about to read is going to make all those misses seem extra impossible, especially on the lower difficulties. Here's what I've been able to dig up in the ini files and the code (using the XCOM 2 SDK):

(Some of this isn't cheating, really, at least not in the "behind the scenes" sense, but just the differences between difficulty levels like you would expect. I got carried away and listed everything I could find, so consider it a comprehensive look at what each option actually means.)

Rookie

Not to disparage anyone's skill level, but if you honestly are having trouble or think the game is tough at this point, then you either have monumental bad luck or are just missing something fundamental. There's no way around it; on Rookie, XCOM 2 is cheating for you, and it's cheating hard:

• Rookies start with +2 HP and +10 will.
• XCOM's chances to hit are all multiplied by 1.2 times. The game doesn't show this, but internally all of your 50% shots are really 60%.
• When you miss, you are given a flat +10 aim bonus on all shots over 50% until you land a hit for each miss. This means your 50% shot is now 70% (50 base * 1.2 +10) after one miss, then 80% after two, and so on.
• If you get hit by an enemy, they get a stacking global -10 aim reduction on that turn, as long as you have less than five squad members. So, each landed shot makes it less likely that they will land another.
• If one of your squad is killed (leaving you with less than four), or if you bring less than four units into a mission, you will see a +15 aim bonus for any shot over 50% and a flat -10 enemy aim penalty for each dead/missing unit below four. This means, if you deliberately bring only one trooper into a mission, you are looking at a +45 hit bonus and a -30 enemy aim penalty. Coupled with your flat 1.2 hit bonus (which is applied first), a 50% hit is now turned into a 105% auto-hit (except it isn't, as I'll explain later).
• The game will try to steer inactive enemy pods away from you as long as you are fighting at least four enemies. You can still trigger nearby pods on your own if you are careless, but they shouldn't wander into you on their own.
• Can get easier missions with less/weaker enemies on them.
• Your troops take less XP to rank up, and various activities and timers such as building or wound recovery are faster.
• Various other small tweaks, such as item costs and reward increases.

The hard cap for aim assisted shots is 95%. It's still XCOM, after all, so it remains possible for you to flub your shots unless it actually says 100% on it, and even then the enemy can dodge for reduced damage (which is HOT BULLSHIT I might add, but whatever). This just makes it much more likely that you will avoid multiple misses in a row as long as you are taking shots with a reasonable chance to hit.

It is possible that the AI is shackled at this difficulty level; not throwing grenades when they can, and reducing panic to just running and hiding versus friendly fire and suicide grenades, but I didn't see any explicit references while trawling through the code, so consider that hearsay.

Veteran

Veteran cheats, not as bad as Rookie, but more than you may suspect.

• Rookies start with +1 HP.
• XCOM's chances to hit are all multiplied by 1.1 times. Meaning that your basic 50% shot is now sitting at 55%.
• When you miss, you are given the same +10 aim bonus on shots over 50% for each miss as you do on Rookie. Your 50% shot is now at 65% (50 * 1.1 +10) after one miss, 75% after two, and so on.
• The "less than four soldiers" bonus is now down to +10 hit bonus (from +15) and -10 enemy hit chance (the same as Rookie). It still stacks.
• The game will steer inactive enemy pods away from you as long as you are fighting at least six enemies (up from Rookie's four).
• Can get easier missions (the same as Rookie) with less/weaker enemies on them (more/harder than Rookie but less than Commander).
• Your troops take more XP to rank up than Rookie, but less than Commander. Build/wound timers are also longer than on Rookie.
• Item Costs are higher than Rookie, with less rewards.
• The enemy AI may still be shackled, but I can confirm your panic isn't. Troopers can and will throw grenades at your own guys when they freak out.

So, there's still a lot of background tweaking going on. All of your shots are still better than you think, you still get the miss streak busting bonuses, and you still get some major buffs when you have less than four units. The game will actively try to keep the number of units you are facing at a reasonable level, but not as much as it did before. It also removes the enemy's hit streak penalty, meaning that you can and will see a lot more landed shots on a single turn.

Commander

You've got to be kidding, right? There's cheating here too? Absolutely; not even a Commander level campaign is on the level, although things also start swinging the other way as well.

• Rookies start with the 4 base HP, no bonuses.
• Enemies are now given +1 health. No more surefire grenade kills on those early troopers!
• XCOM's chances to hit have no base level modifications. 50% is 50%.
• When you miss, you are given a +15 (!!) aim bonus on shots over 50% for each miss. Yes, that's actually higher than on either Rookie or Veteran, which I assume is compensation for removing the base shot multiplier. Now your 50% shots are actually 65% after one miss, and 80% after two. This means that it's actually easier to avoid miss streaks on Commander when you are right at the aim assist threshold (versus Veteran), but it falls behind once the unmodified chance to hit increases enough for that multiplier to take over.
• The game will steer inactive enemy pods away from you as long as you are fighting at least six enemies, the same as before.
• Can get easier missions than Legend, with less/weaker enemies on them (more/harder than Veteran, though).
• Your troops continue to take more XP to rank up, and you will see even longer build timers and wound recovery requirements.
• Still greater item costs with lower rewards.
• The AI is unshackled, as far as I can tell.

I'm surprised they have as much behind the scenes cheating as they do for this level. In a way, I'm a bit sad; I figured Commander was basically the same as Legend without the re-balanced strategic layer and some enemy HP modifiers, but that's not the case at all. You still get the miss streak bonuses, although the "less than four soldiers" buff is totally gone, and you still have the game doing its best to keep engagements at six enemies or less.

If you want to do some modifications to basically make a Commander+ difficulty level (no hand-holding like Legend but without the major strategic re-balancing), you can tweak XComGamecore.ini (search for AimAssist) by dropping MissStreakChanceAdjustment=15 to 0, and XComAI can by modified (search for MaxEngagedEnemies, although it's right at the top of the file) by changing MaxEngagedEnemies=6 to -1.

(Those should be in Documents\My Games\XCOM2\XComGame\Config on Windows, although you can also modify DefaultGamecore.ini and DefaultAI.ini in Steam\steamapps\common\XCOM 2\XComGame\Config to the same effect. I like leaving the defaults alone, though.)

Legend

Cheating in Legend? Maybe for them, but certainly not for you.

• Rookies start with 4 base HP as before.
• Enemies are now given anywhere between 1-3 health and an extra 1-2 armor.
• There's no aim modification at all. No miss streak bonuses, no enemy aim penalties, nothing. What you see on the screen is exactly what's going on...except for that bullshit dodge mechanic which isn't surfaced anywhere.
• The game will not steer pods away from you. There is no max engagement size; if you are fighting six aliens, and that pod is set to path into you, it's going to path into you.
• Legend is unable to get the easier mission types and all versions of the missions are as hard as they can be, including extra enemy pods not seen on Commander.
• Harder enemies will show up earlier on Legend than on any lower difficulty.
• The strategic game is totally re-balanced, to the point that it almost winds up twice as long. I won't go into the details as I doubt I've even found every change, but suffice it to say that everything takes longer, costs more, and is generally just harder to do.

Legend is hard, but I found it easier than Impossible in EU/EW. It's the only difficulty out of the four that isn't making modifications to your dice rolls without telling you, and it's the only one that isn't pulling any punches on the tactical layer. The strategic game is extensively modified from what you get in Commander, but that's to be expected.

If you know of any other behind the scenes or balance changes that I've missed, I'd be interested to hear about them.

Helpful! Thanks for this.

Huh. I actually thought this was less modified, based on how many bullshit 75% shots go awry.

This is generally just good game balancing, although it encourages specific playstyles (namely, gang up on an enemy and keep shooting until it goes down) versus others (go for highly secure one shots across multiple enemies). That could have been achieved by giving you a boost to aim the higher your base chance is (so 80+ turns into 100 and 75+ turns into 90 and so on. It would have generated less "FUCK!" moments as you miss easy shots, though, for better or for worse.

In a way I'm a bit sorry I read this. Now I know that playing on anything up to Commander it's actually in your interest to get bad shots done before your killshot to trigger the gang up bonus. Makes Lightning Hands even more relevant, as it's not just giving you a free shot, it's also giving you a Marked-style aim bonus on your actual sniper shot if you miss. I can't un-know this, so now I have an extra cheese tool.

@noelveiga: Yeah, it's a real pandora's box, except with cheese at the end instead of hope. On anything but Legend, it really is in your best interest to start by aiming low, although amusingly enough that's how I always did it in the past, so perhaps that accounts for the variance I've seen in people's perception of the RNG (it never surfaces that on screen, so all my 15% into 50% shots are actually 15% into 65% on Commander and I just never realized).

I'm not sure why they made it so that all shots add to the miss streak bonus (instead of just shots over 50% which makes more sense to me, as I never expect to hit shots under that threshold), but it certainly seems like you should roll with your 15% "snowball in hell" style shots first before taking anything you are expecting to land.

@mirado: I did that too. It's a reflex from board wargames where overwatch is not a thing. You take your worst shots first in case you luck out and can use your good shots on a different target.

The interesting thing is it devalues overwatch. Unless you know what the AI is going to do that boost in aim is more valuable than the AI choosing your target for you. I was already using it only for ambushes in conceal or on incoming reinforcements. Now I know why.

This is the strangest case of self flagellation I've witnessed on the part of the player since the last Dark Souls release.

Dark Souls at least has a compelling world to explore and unique gameplay that hasn't really been done before. In XCOM's case you might as well just be rolling literal dice and punching yourself in the face if you roll below a certain number.

I'd already read a bit here and there about how the game (and the previous one) presents misleading information, but I wasn't aware of the specific numbers until I read this, so thanks for that I guess. But even though all of this is in favour of the player, I am still not really a fan of keeping the relevant info hidden.

There's this part of my brain which feels that if you're making a game based around dice rolls, then those rolls should simply be presented honestly. Using modifiers, multipliers and whatnot is fine in essence, but I guess the purist in me just wants them to show me those actual modified odds.

In fact, I think it'd be better to create various difficulty levels by adjusting everything around the dice, rather than fudging the rolls themselves. There are plenty of things they could scale; the amount of enemies, their health/armour, the damage weapons do, the amount of turns you get to complete objectives, how smart the AI is about choosing cover and choosing who to shoot at in the first place, etc, etc, without ever having to 'lie' to the player. Don't get me wrong, I am actually really enjoying the game; I just don't believe 'variable difficulty' and 'a higher or smaller chance of succes' are the quite same thing.

I guess the idea behind this is partially to create exhilarating 'holy crap I actually made it' moments for the player, leading them to believe they beat the odds, but I can't help but think it's also to mitigate the rather out of whack expectations some people seem to have of dice rolls. I'll admit I'm guilty of this myself at times; treating the higher end chances as "this should be a succes" and then getting annoyed when a 90% shot still fails. But I try and remind myself that even that is not a certainty; I mean, if with every step I took in real life I'd have a 10% chance of falling flat on my face, I don't think I'd dare to cross the street any longer (not without some power-armour in any case).

I've been reading about this as well, I guess I never notice it on Veteran because after I miss 3-4 times in a row most of my soldiers are dead before aim assist can make a difference lol. I don't believe the enemy pod thing, I don't think I've had one mission where a second pod didn't trigger before I finished the first one, or right after and in flanking position most the time. Not to mention reinforcements after that. :-(

@oursin_360: Sorry man, it's all in the code. There's a clear threshold in which the game will say "Ok, give this guy a break" for a period of time, and it differs based on difficulty level. You can still stumble into a pod or take too long for this "Throttling" to be cleared out, but it's all there.

Yeah, I'm not advocating that they should leave it buried; in fact, that's why I called it "cheating." I didn't mean cheating in the sense that the game is too easy and they should eliminate that stuff, but rather the 50% that you desperately needed to land and the subsequent joy that comes from that is actually unearned. It'd be much better if they displayed the real numbers (with a "Difficulty level" modifier or something) on screen, so you can actually feel like you got a bit lucky, versus this somewhat condescending "Oh look at you, you're on a roll!" vibe I get now.

As for the difficulty level adjustment, a lot of that happens on top of the roll modifications. The health is modified for both your troops and the enemies, and while I didn't do a whole lot of deep diving on the AI's priorities, the entirety of the decision making is available for you to browse should you choose to do so. I'm not sure if it contains any difficulty specific shackling, but I wouldn't be surprised and I've heard mention that it may, especially when it comes to things like grenade usage and panic mechanics.

One thing I forgot to mention is that wound classification is also different based on what you are playing; I don't have the exact numbers in front of me, but in general it's easier to wind up "Gravely Wounded" on Legend than on Rookie, for instance.

The extra health and steering pods away from i'm fine with, but i don't like the idea of the game misrepresenting the chance to hit. If you're going to give soldiers an aim boost on lower difficulty, why not just surface that information?

Either way, thanks for the info, I might just muck with the ini files and make that commander+ difficulty you were talking about.

I agree with what @colony024 said, if you're going to make a game where dice rolls are the basic mechanic then its pretty crummy design to fudge that mechanic so much. It would be like if Mario just jumped higher sometimes (obviously that would be much more noticeable than this stats stuff).

Anyway, let me counteract this mostly negative post by saying that i'm powerfully addicted to XCOM2 and I love it to bits.

yep... this just proves something I already knew all along, I have horrible luck... playing on veteran and even with all that help my dudes still miss like crazy.

@mirado: Yeah I'm with you there. Knowing they're modifying the rolls in my favour is actually diminishing the 'whoa I got lucky' feeling. I didn't mean to imply that they should avoid using any sort of modifiers altogether, merely that I'd prefer seeing them presented transparantly.

That said, a lot of this game is all about managing and positioning your squad in such a way that I'm never in a position where a single roll can make or break me. Obviously this doesn't always work out, and sometimes plain bad luck will occur, but I don't quite understand the people I've seen online complaining about how there's no tactics to it and that it's all random luck.

@onarum: Not that I am disputing the amount of luck you have or anything, but isn't it also in part because of which shots you decide to take in the first place? Most people probably won't even risk taking a 60% shot most of the time, and will only fire if the odds are telling them 80% or above (obviously making these numbers up on the spot here). It's likely people are seeing shots with high probabilities failing a lot more than they'd expect, simply by virtue of those being the only shots they take to begin with.

@colony024: well, more often than not I don't have much of a choice really, a poor ranger gets flanked by like 3 mutons from a unseen pod and then you have to rush there with the rest of your crew with no room to flank them yourself but you're desperately trying to thin their numbers a bit so that the ranger will have a chance to live, so you take whatever shots you can.

Though I have missed my share of 80%+ shots lemme tell ya...

@onarum:Haha, yeah like I said I wasn't trying to dispute your luck. Sometimes this game is just plain harsh, but for what it's worth at least that .gif made me chuckle.

Hmm, seeing this makes me wish they did difficulty like the Silent Hunter games: here's a list of all the difficulty options (unlimited fuel, unlimited air, no dud torpedoes, etc.), turn them on/off as you wish and as you do the 'realism' score changes.

I'd like to mix and match options spread across their difficulty levels.

• no bonus health or will (commander)
• no enemy bonus health or armor (veteran)
• no aim modifications (legendary)
• pod steering with six enemies (veteran)
• can get easier missions (rookie)
• unshackled panic (veteran)
• unshackled AI (commander)
• strategic layer (veteran)

I'd be pretty shocked if any game outside little indie projects in the last decade used pure rolls without invisible modification. I suppose this is a reveal for people who don't know how/why games are made like this/haven't modded games before/read modders notes. Very, very few games operate around true randomness. It is known, to some extent, that such a thing is a fallacy anyways. But anyways..

Edit: note to self: isn't there a PoE blogpost about this?

I feel I have to point out, at some risk pf appearing lame, that the use of the word "Cheat" here is wrong vs Design. They are part and parcel of THE BALANCE SYSTEM. Only when you circumvent the system are you cheating. (Also "Cheating" when a system is involved, has nothing to do with whether you cheat for ease or challenge.) Your blogpost is interesting OP, but for me the deductions are all wrong (well, as regards the "cheating for you" frame of the post). When we see a value of 0 for example, we should not read that as the "true value" because it is not modified. We need to respect the intelligence of the Firaxis designers and realize the value of 0 (or whatever it is) exists *to be* modified. This is a really complex thing to do! But it has a hell of a result.

Looking at all your difficulty notes, it's obvious there is no pure game buried within there. They all have "quirks". Every difficulty functions as it should, offering an (attempt) at different accesibility/experience levels. Because no one can simultaneously experience all at the same time/with benefit of being at the apoproate skill level, it's way too early to say how well balanced it is overall.

My opinion out of the way, a thanks for posting! I think it's really fascinating to get a look at what the designers thought was fair. It says so much about how they assume frustration works in players and the "Swings" the devs want you to feel. It also reveals the system is not binary usually, which is very brave, and probably explains some of the people reporting back to back squad wipes even on rookie. Seems to me it's because the modifications are not meant to guarantee in most cases, only increase the odds. The "No 100%" rule is pretty neat.

@geraltitude: I meant "cheating" as in "the number you see isn't actually the number that's calculated behind the scenes," not in the sense that they shouldn't be messing with the numbers at all. I have no problem with them doing what they are doing, I just want that surfaced so that the 50% I felt lucky to land shows up as the 75% or 80% that it really is. The way it's set up currently, you get a rush for landing a shot that you didn't expect, but internally you are actually more likely to hit than not, and that reframes it from "I can't believe I landed that." to "I should have landed that."

It cheapens the experience a bit. You're being "cheated" out of the gambler's thrill once you realize someone's got a magnet under the roulette wheel, making it more likely that the ball lands on black.

I'd never presume to know more about game balance than a designer, only that it makes your triumphant moments a bit weaker once you know that brave rookie, who dodged those shots to get into cover and blast that alien on a 50/50 flip, actually had a very high defensive bonus from his dead friends and a 95% shot on that alien, and you just didn't know it.

Even with all of this, there are still those that aren't playing on Legend who claim to "miss 7 or 8 70-80% shots in a row" multiple times or "Miss more 70-80% shots than they hit." I always found that a little hard to swallow, and even more so that the mechanics and math simply don't support things like that happening.

I hated this in Enemy Unknown and I hate it in XCOM 2. While I understand they had design reasons to do things this way, I'm definitely going to have to work out how to mod all of these hidden bonuses out before I play XCOM 2 again. Knowing that this is going on behind the scenes will kill my enjoyment in this game otherwise.

Good to know the game designers cared more about manufacturing contrived emotional states than creating a mechanically interesting strategy game. Bugs the crap out of me when game designers obfuscate mechanics like this in service of a certain sentiment, in this case the "that's xcom baby!" crap.

I would care less if they just incorporated this stuff into the actual game mechanics and didn't just lie to you. Get rid of the displayed percentages and just list something like "poor" "average" "great" odds for hitting. Make the comeback mechanics a "momentum" or "desperation" meter or something.

I remember hearing about these things for Enemy Unknown and I think they are great additions. Even while playing Vetern the game is a mother fucker; I can only imagine my frustration if they didn't tilt slightly in the players favor.

I've seen, but have no played, X-COM: UFO Defense and I don't think the newer XComs would be nearly as popular if they were as difficult as that game.

Hm, I dunno.. That rookie could still die or fail to make that shot and leave his team dead. And I feel so far as the internet has spoken, that's just more common. I've seen this occur myself anyways, so the fact that when the good version happens it was secretly more likely to happen than I expected doesn't change things that much for me. Frankly, on any difficulty, good things just don't seem to happen that much!

If the shot says 50%, and you feel good about making it, I'm not sure how much it matters what the truth is.. Especially considering less than 1% of players will ever know "The Truth". So while I can't say that you, OP, feel worse about the game now that you've seen it guts, I don't feel like that's universal. Playing on Rookie and Veteran I've landed and missed a number of real low percentage shots, as well as real high percentage shots. So, I dunno. In game the rolls do feel to me like they swing intensely. It feels random, which is more important than the truth of the design I think. If the game really didn't want me to lose on Rookie, it wouldn't have killed a squad of captains and majors in 2 back-to-back turns. I knew it was time to move up from Rookie to Veteran when I had two missions back to back with 100% shot accuracy. Just felt weird! As for commander.. definitely I got better at the game as I played it more but it more or less feels similar to Veteran for me. The strategic layer is where I bungled that run up.

Anyways, I guess I just feel there's no such thing as pure XCOM buried in this code. There's no fiddling with any ini values that will somehow leave you "without your hand held". That's just a crazy simplification of balance in my eyes! The purpose of each enemy and soldier class. Every value of every ability, every cost of every item was balanced against these modifiers. On top of that you can't touch any combat values without touching the strategic layer values. So, while I like the idea of a "what you see is what you get" %, attaining that doesn't mean I'm playing either a more balanced or more fun game.

All that said, I'm definitely interested in playing and making my own XCOM rebalances because I think there are a world of possibilities. Flavors, more accurately.

It's because people complain when their 95% to hit at point blank range misses, but don't talk about those clutch moments when a 25% shot hit and saved everyone as often. It's just the nature of the internet and people in general. Of course it's going to appear as though the former is happening more than the latter, but there is no hard data to support that is actually what is occurring.

@ssully: The original X-COM UFO Defense is actually less difficult than people give it credit for. It can be a hell of a meat grinder early on, even when you know what you're doing, but by the end when you have an army of psychic-powered flying soldiers with blaster launchers capable of precision guidance straight into a UFO it becomes something of a joke. In that way, it's not much different from the new ones, with their similarly limp end game difficulty curve.

Fascinating! Thanks for breaking this down.

That's kind of the point though isn't it? To feel like you are near the brink of destruction, only to adapt to your enemy and finally fuck them up?

Yeah, certainly that's part of the appeal of XCOM both new and old. That said... I do wish the difficulty curve was a little more consistent in those games. I can't speak to XCOM 2, but there's a point in the mid-late game of Enemy Unknown and Enemy Within where your soldiers basically become a bunch of bulldozers capable of demolishing anything weaker than a sectopod or ethereal and it turns into a bit of a cakewalk to the end.

I think I am about midway through 2 and it hasn't hit that point yet. It's more of an ebb and flow; I get better shit and bulldoze for a bit, and then they introduce a bunch of new units that force me to rethink my strategies. Seeing some of the research I have ahead of me though I wouldn't be surprised if the games end on full bulldoze mode.

@mike: yeah I 100% agree with that! It's what is so interesting about all difficulty debates. The same could almost be said about performance. Basically we are relying on some kind of consensus/popular opinion to create Truth. Hard, complete, data is almost never at hand.

@geraltitude: I wouldn't go so far as to equate this with the reports of bad performance. It is absolutely clear that this game has some major technical problems, that isn't hyperbole.

The most difficult thing for me in XCOM 2 is having multiple objectives to accomplish (research, base building, etc.) and a doomsday clock and setting up relays and picking up supply drops. I'm not having a problem in combat on Veteran at all, but oh my god does the top level stuff give me a migraine.

@geraltitude said:

Hm, I dunno.. That rookie could still die or fail to make that shot and leave his team dead. And I feel so far as the internet has spoken, that's just more common. I've seen this occur myself anyways, so the fact that when the good version happens it was secretly more likely to happen than I expected doesn't change things that much for me. Frankly, on any difficulty, good things just don't seem to happen that much!

I can only counter your anecdotal experience with my own (which has been the opposite), and mention that people seem FAR more likely to complain about a miss then to remember all of the hits they landed.

@geraltitude said:

I've never really mentioned a desire to find a "pure" XCOM or even implied what difficulty I considered to be "true" (I don't agree with people that look at it that way), but I do think that the game's fluffing of your hit percentages without telling you is condescending in a way, even if they didn't intend it to be. Like I said, I feel that it's akin to holding a cartoon style magnet under a roulette wheel; it robs a bit of the thrill for me, but it's cool if it doesn't bug you. However, the idea that the classes are balanced against those modifiers is not only untrue but backwards: they're built the same for every difficulty level (minus a small HP change), but the hidden modifiers are tweaked to make each step down easier. The other parts of the game are fairly disconnected from the hit rolls (beyond the wound timer/classification changes), but in the end I'm not so worried about finding a game with better balance as I am with just understanding what's going on under the hood.

I personally would have more fun if I knew that shot I blew was actually as tough as they said it is, but I can understand why that doesn't bother you, or why people don't care. With that being said, it seems as though I'm not alone if the mix of responses to this blog post are any indication; in fact, the almost 50/50 (ha) split between caring and not is skewing far harder in the "care" direction than I originally anticipated.

At the end of the day, I'm just happy people are having a good time with XCOM 2, at least when it isn't on fire.

@nickhead said:

I'm glad you found it interesting! :D

you hiding data somewhere :P ? yeah definitely agree, but never talked about hyperbole anywhere.. basic point is no hard data: I don't have access to any numbers about how many copies were sold, and how much trouble every user is having, or what configurations are problematic. duh, it's kinda busted (so many hard crashes on my machine D: ). But how busted exactly? Anyways not the point of this thread! Don't wanna hijack an interesting mechanics post with this.

Hey people are just bad at understanding random outcomes, hidden modifications or not, we're just irrational (incidentally I overheard people disagreeing over the solution to the Monty Hall problem the other day). If people are frustrated when 70% shots miss 3/10 times imagine how little fun they would have playing the game if that 70% wasn't actually 95%. Frankly it is a credit to their devotion to the purity of the strategy that they include a difficulty mode that doesn't fudge the numbers to account for human perception.

This is great stuff. There is an XCOM 2 SDK. At some point I want to get into modding, but there is only so much time.

Just to chime in, I'm definitely one of those who doesn't care what they do as far as modifiers for difficulty, but don't lie to me dammit! I personally would have an option for "base + modifiers = total" display like on every tabletop character sheet I've seen. The dodge mechanic is an invisible "eff you" number, and it's not fun when a codex splits two extra times because it dodged and got grazed by flanking attacks forcing me to proc all three of my Run and Guns in the same turn to chase them down. And then two faceless appear one after the other because I skulljacked on a retailiation mission. That's the "It's Xcom, baby!" mentality for you, I guess. Doesn't make me feel like a master tactician, but a man out of control. That enemies killed per turn stat can be quite telling if accurate.

That said I enjoy the games in the new series a lot, but then I'm not above light save scumming in order to maintain personal enjoyment. Knowing the next number is garbage lets me plan around it. It's a habit I picked up playing old Fire Emblem on emulators.

So where in the config files can I change that 1.2 chance to hit to a 10?

Also I have found that you can't really save scum certain things. I was shooting at a dude (with a 65% chance to hit) and kept missing regardless of how many times I reloaded the save. Another time I was trying to hack an alarm thing so I saved before I did it and kept failing that too even when I kept reloading the save.

The fact that veteran silently boosts accuracy by 1.1 times the stated value makes those sword misses more unbelievable since one my rangers missed a 90% chance to hit with a sword, when in reality he had a 99% chance. Holy cow.

This was talked about in another thread, but like the first game XCOM2 has a seed. Someone in the other thread described it as being that the rolls are done in advance and saved, so if you reload your next roll will always be, say, a 65 (and fail an action that needs a 66 to win). If you reload and do something different, then go back to the hack you might get a different result.

In the first game there was a "second wave" option to have a random seed so that reloading would give different results.

Personally, if I'm not playing a ironman game I like the seed, as it feels like you can be more tactical if that makes sense. Like, if you know 3 steps leading up to that roll work, you can try small variations on a tactic rather than leaving it up to fate.

First off, a quick thank you to @rorie for linking to this blog, as seeing that in my Twitter feed was kinda cool.

I actually used the SDK to pull of some of these statistics and formulas; the code is very well commented (if a bit outdated with the current balance changes, as it talks about a 20% miss streak bonus which doesn't exist at any level anymore) and uses fairly clear variable names, which is something I'm notoriously bad at (I was once told that I wrote great code, but everything I make feels like it already ran through obfuscation). It's a HUGE download, though; something like 50GB so keep that in mind.

Well, if you are primarily interested in just playing with the hit chances, open XComGamecore.ini or DefaultGamecore.ini (I usually choose the former, which is normally in Documents\My Games\XCOM2\XComGame\Config) and Ctrl-F for AimAssist. That will pull up the section that contains every variable for the hidden dice modifiers, and you can change everything from what threshold the aim assist kicks in (default is 50%), to how big the bonuses is for losing a trooper.

Just keep in mind that the "1.2" actually means "multiply by 1.2 aka take 20% of the base number and add it on top" whereas every other variable (all the 10s and 20s, etc) actually means "just add it to the base".

If you are interested in changing the actual formula, let me know and I'll try to find the file it's in, but you'll probably need the SDK to make that change.

Remember, aim assist caps at 95%. Not saying that's a hell of a lot more believable, but it's still a bit better. Also, enemy dodges (when "grazed" pops up and you do less damage) are a separate calculation that can kick in even on shots that read 100%, but I didn't include that as I think it isn't affected by difficulty.

I'm pretty near the end I think, and I've definitely started to feel that "wheels spinning" feeling a bit more. I certainly haven't necessarily played optimally in the geoscape, but I feel that the problem isn't the drop off in difficulty per se, it's just that when you reach the point where you should be able to steamroll into the endgame, you are still being held back a little bit on the overworld from doing so.

You spend a bunch of time waiting out research timers and stuff, and the more of that you have left to do once you hit the "steamroller" moment the more minor missions you have to do that feel like busywork.

This was way way worse in the first game though, and I think XCOM2 does a few things better to stop that happening:

• There has been a slightly elongated spooling out of harder enemies. I feel like it's easier to get into a groove, and then your first sectopod shows up and you have to be on your game to deal with it. It's not necessarily a moment that destroys you, but you can get lulled into a false sense of security.
• The strategy layer is slightly more engaging, with making choices of where to go, and having a few different resources to manage. You're not just sitting their scanning on the geoscape waiting for whatever to go through.
• I can't say for certain but I feel like you just get slightly fewer "busywork" missions. This is partly because the mssions are often necessary to advance your strategic goals. I think it's also because UFOs are fewer and further between.
• Speaking of UFOs, by the end of EU I hated doing them, because the maps just became too big. Part of what was boring about them was just having to move across them, only to flatten any resistance. So far I haven't had any mission that couldn't be done in an hour(ish).
• Finally, I think if I'd braver in tackling objectives I could be pretty close to wrapping everything up by now.

On the other hand, I feel like it could be improved by having the drop off be less present in the strategic layer. I feel like the Avatar meter is barely moving for me now. If the pressure was the same it was in the early game I think it would be more engaging - it would certainly push me to tackle objectives quicker!

Someone should mod in the ability to see the chance to Dodge.

So what you're saying is, there isn't a difficultly level where all the "base" stats are actually the base stats? I don't know a lot about game design, but isn't that odd?

I mean, if I were designing a game, I'd work on balancing until I had base stats that felt "fair" and then call that mode normal or whatever equivalent you want to call it. Then every difficulty below that you'd get a bonus and every difficulty above it the AI would get a bonus. That's probably not the way it actually happens, but it seems the most logical. What XCOM2 has got going on implies they just threw numbers on things initially and then worked the multipliers until everything felt kinda OK.

I think the part where this entire discussion completely breaks down is because two groups of people are arguing for two different things that has to do with the same concept: randomness. I can't speak for anyone else, and I shouldn't, but I kind of will: when someone is upset that they miss 70% chances three times in a row, they are not oblivious to statistics and probability. They just wish the part where you are playing the game would accommodate for these scenario's. But, the other person comes in and goes on to start explaining how math works. It's far less irrationality, than it is not having fun. Sid Meier realized this himself as he said in a keynote speech where he was blown away how the vast majority of his playtesters hated losing to Barbarians over and over in Civ when the game told them they had 90% chance to win. It's not that they don't understand math, it's just that it doesn't make for a fun game. That's not to say shit should always hit, but in any - I'll say again - ANY person's game, if the game tells you you have 90% chance to win, it should round out to 90% wins. The fact you can draw the short end of the stick and round out at 50% wins, makes it incredibly frustrating, even if statistically possible. If you're going to make a game that is solely based on probability and nothing else, then statistically, some people will have an easy time with your game because they routinely get the desired outcome, and also statistically, some people will have no fun because they keep losing no matter what they try.

If you're telling me you're making a game where it depends on statistics whether I will have a fun time or not, then your game is flawed. This is what Sid Meier realized. Some players, when recorded, only had like 40% battles won against Barbarians during their playtime. There was nothing wrong with the math, it was 90% chance to win each time. They just got unlucky rolls. That makes for a frustrating experience. So the game cheats for the player to even out higher as you get further away from that 90%.

And it is apparent in the game. I've been tracking my hits and misses in my most recent finished game on Commander difficulty (which uses cheats for the player), and they were spot fucking on when I entered the final mission. I recorded it in increments of 5% (so 4-9%, 10-14%, 15-19% and so on) and the value always ended up in there. For example, my 80-84% column has 82% success rate. So even if the game cheats, it pans out in my own game that I play.

I think the dissonance in this discussion comes from people (in my opinion wrongfully) suggesting that proper probability that doesn't cheat for the player makes for a strictly better game. I'm not saying it's bad if there are no cheats, but cheats prevent a player from being screwed over purely because of RNG purposes. I will play XCOM 2 roughly 20 more times. I might never have a bad RNG run. But, someone somewhere will only play it once, and without cheats might have his only run be totally fucked. I honestly believe with all my heart I can call that poor game design. Because if you're smarter, you can account for RNG outliers and correct them to create a more fun and challenging game.

For one last example, I'm fairly sure some of the abilities do not get the cheats applied to them. Bladestorm is one such. I have seen it proc more than when I started counting, but when I did, I reached 38. That's 38 procs of what is supposedly 85+% chance to hit and it didn't hit once. I am on my fourth game now and got the choice for Bladestorm or Implacable again. I thought to myself, why would I pick Bladestorm? It's a dead perk. Then, I convinced myself I had bad luck and should keep trying. 9 procs later, still all misses. It is literally a dead perk in my game. I honestly do not think that's good. That amount of bad luck is totally statistically possible yes. And get that shit out of my game. It's not good, fun or challenging. It's just dumb. It led to me installing a mod far quicker than I thought I would. One that increases the chance to hit on swords to be more in line with shotguns (and actually still lower than those) guess what? 5 out of 7 Bladestorms hit now. Far more reasonable.

By the way, I'm doing my 4th run on Legend difficulty, and how much easier this run (on a higher difficulty setting!) is than my first two runs on Commander is just actually hilarious. Although the type of RNG responsible for that is not so much in chances to hit, but moreso map generation and pod encounters. The objective in Guerrilla Ops are generally 3 dashes away from my spawn, whereas they were 6 dashes away in my first two games. Especially 'protect the device' missions were sometimes literally impossible, whereas here I'm already there after they shoot it once. Pretty much every time. I've been watching some Let's Plays too and all I can think was: 'you guys have no idea how much worse your luck can get'.

To be clear, I like randomness. But not unfiltered randomness. This is still a videogame, not a showcase for the concept of probability. If shit gets too ridiculous to the point of not standing a chance just because of random numbers being generated, than the game should totally make up for that.

I really do not like this practice. As a player who takes joy in really delving deep on a game, and progressing through its many difficulty levels, lying to me about something as fundamentally true as a percentage is not good for me moving forward. It builds preconceptions that actively defy both the highest difficulty and also any other game with "surfaced" percentages, and also, you know, reality.

Does this apply to overwatch as well?

I'm playing on commander, second go around, and after a full game never using over-watch ambushes (because why would you?), I decided to do it on a whim with squaddies. They all had a similar chance to hit the pod, 75~%, so I set three to overwatch and had the fourth shoot. All 4 shots miss and I threw my hands in the air and thought "that's why you don't do that". I mean, missing four 75%'s in a row is not particularly special in xcom terms, but after reading this post, it seems like it must have been rather unusual. I would have had to miss a 75%, a 90%, a 95% and another 95%. Now, that can totally happen. But it still seems more likely that I am missing some piece of the puzzle here than I actually just rolled that bad.

@foolishchaos: Overwatch outside of concealment applies a penalty to accuracy that amounts to a 30% decrease in accuracy against a normal move and a 40% reduction if the target is dashing. Using your example, a 75% hit chance against a stationary target is reduced to 52.5% when on overwatch and 45% when on overwatch against a dashing target.

I do not believe any aim assist applies to overwatch but I don't have the code in front of me to confirm that. It's also possible that cover applies to overwatch shots, making it even less useful as a method for actually killing targets (versus locking them in place).

Again, concealment negates the aim penalties, but I still do not believe that it applies any aim assist.

@thomasnash: My only comment/question related to your response is with the avatar project. Idk about you, but I was very close to filling up the bar in the early game(2 tiles away, to be exact). I simply didn't build out comms facilities or communicate with other areas as fast as I should have, so the enemy went uncontested. Once I poured 100% of my focus into expanding and taking out black sites/other objects, I haven't had the avatar project progress at all. The only reason I think that is so is because I am taking out a blacksite or objective every 3-5 missions now. I don't know how it works behind the scenes, but everytime I take out a black site it reduces the bar, but also says progress is slowed/halted. So I wonder if the "timer" in the background for the avatar project is reset everytime you take out a high value objective? If so that might account for the avatar project not being as intimidating as the early game.

@ssully: It seems like that to me too. Whenever I took out a facility, it would typically be several weeks or even much longer before the aliens made any new progress on the Avatar project. Also, if you notice on your map when you look at Alien Facilities, they will have 1-3 red dots next to the icon. That's the number of bars that will be taken off of the meter when you take it out. Handy for prioritizing targets if you're running out of time.

@foolishchaos: Something seems off about Overwatch in general. My guys would pretty reliably miss multiple Overwatch shots in a row during concealment ambushes and this was only on Veteran. Even whether it would trigger at all felt like a lottery at times. I never really felt confident in using it again until I had soldiers with a bunch of Superior Aim boost equipment / PCS.

Am I the only one who thinks its kind of unethical to straight-up lie to the player about how the game is working? I get that's it's 'just a video game,' but it really does feel manipulative and like the player is just trapped inside a Skinner box when the designers are going out of their way to trick you into being excited about epic unlikely moments. Am omission is one thing (and wouldn't the observant player notice that they NEVER get patrolled into once there are 6 enemies on screen?), but a straight up lie crosses a line.

For example, if a developer were doing this same kind of stuff in a free to play game to try and induce the player to purchase microtransactions, play longer, share with friends, etc., then we'd be pretty aghast. Is doing it just so that the player has a better experience any better? And I don't think Firaxis added this stuff to try and make more money; I think they added it to make a better game just for its own sake. But still, it's not like this lie doesn't contribute to their finances at all.

I get that they did this stuff to try and curate a particular experience and to avoid issues with variance, but it seems to me like there are many many ways to go about this that don't involve intentionally presenting incorrect information to the player. To me, it suggests a lack of respect for the player as a person, like the player is a psychological black box to be manipulated because we would react negatively to true variance. And it also suggests a lack of faith in your game when you have to tweak things behind the scenes so that the player constructs a false narrative about what just occurred.