This explains the yo-yo nature of the combat. This also appears to be a case study of how not to do game design. I always felt the game was bullshitting me with the on screen stats.

Also makes me think that I would of been fine playing on Legend instead. But three dodges in a row might of made me break the keyboard in half.

@foolishchaos: I think the aim boosting happens on the next action or turn if I read that correctly and not on the same occurrence since at that point the RNG seed would have already been determined. Personally, that's why I don't put everyone on overwatch during an ambush and would rather leave one or two that still have two action points to be able to adapt to the situation if needed.

Also you finally made this into a thing, @mirado!

@anytus2007: To be fair, I wouldn't necessarily call them unethical as if they are just doing something very insidious in order to come up with a better end result and this is some sort of ethical quandary (like charging for aim boosters to have better shot percentages in-game for example). However, I'm curious as to what kind of solution you could come up with considering you made a lot of deep points.

@gaspower said:

I did indeed! And quite a thing, apparently; it's the most viewed post in the XCOM 2 forum and my most commented blog post by a mile (although I very rarely write blogs to begin with, but still). It even snagged Rorie's attention, which was nice. It helps that the title is a bit clickbait-y, but I like to think that people are actually interested in learning what goes on under the hood of some games, and that's a real credit to the XCOM 2 dev team; there aren't many games that make information like this so easy to view, understand, and modify.

@mike: @ssully: Yeah, this was what I meant, sorry if I wasn't clear. I've taken out 3 facilities, including the first one, and the Avatar project just isn't moving anymore. I think I've had it go up one bar in the past 3 months or something like that. I wonder if it's a little harsher with that stuff on higher difficulties though, I'm only playing on veteran.

Yes it is. I don't think they should do it either, but doing it with the intention of making a better game is obviously 'less bad' than doing it with the sole intention of tricking people out of more money.

Having a balancing device like this in a game which uses luck as the core mechanic is probably a good idea, but I'm not sure why they felt they couldn't expose it to the player via a 'Concentrate!' status being applied to your team or something.

Seems to me that ultimately the issue here is more about presentation, rather than what's actually happening to the rolls. @geraltitude was probably right about this sort of thing happening under the hood all the time, in lots and lots of games. If done subtly enough, most of us will be none the wiser.

But by presenting itself as a 'serious' tactics game; showing you exact percentages all the time and having tabs showing you other modifiers (cover, weapon range and height advantages, etc.) the game very much suggests that all of this info is perfectly accurate. Combine that with the extremely easy access to the .ini files where it quickly becomes clear that this is actually not the case... I just think that creates some dissonance in a lot of folks, making them feel they're being misled.

I personally don't mind them implementing this (well ok it bugs me a wee bit, though I'd never call it straight up bad design), but I wonder what the response would have been if instead the game told you something like "Be aware this is not the actual chance but rather what your soldier thinks his or her chance is" or perhaps if it simply went with descriptions such as "No chance in hell you'll make that" or "You could be blind and still connect this shot".

Because let's face it, for practical purposes the difference between 80% and 81% isn't going to matter, but by giving me the number you're telling me one is better than the other.

Also I don't mean to attack anyone, but people (and I don't mean just on these forums) calling this 'unethical' are taking it a bit too far in my opinion though ;-)

Variance. There is no code that dictates overwatch shots outside of the reaction fire penalty which doesn't apply to concealment.

REACTION_FINALMOD=0.3f

REACTION_DASHING_FINALMOD=0.4f

So I have to rescind some of my hubris, because I've been getting my ass kicked around a bit since every mission had sectopods and gatekeepers. Even had the doomsday clock start on me.

Usually they won't have both sectopod and gatekeeper in a given mission. The shadow chamber briefing when you select a mission will tell you what enemy types are available and you can prepare accordingly (e.g. bring EMP bombs and bluescreen rounds if there's a sectopod, bring more flashbangs if there are gatekeepers.) Semi-spoiler: the only mission that had both enemy types in it if I remember correctly is the very last mission.

This happens on Legend and Commander, not sure about the lower difficulties. On Legend, usually it won't be just one of each, either.

@mike: I finished it on Commander. Could probably be a case of how its seeded every time the world is made.

@mike: @gaspower: Pods are assigned via a system that takes the month, "threat level," and whether or not the unit is consider a boss. Some units are also restricted to certain mission types, hence why something like a Berserker seems to be so rare; the right combination of month (not to early but not to late as to be overridden), threat level, and mission doesn't always click into place. The later you go, the more willing the game seems to put the big boys together in one map.

You can check the system out in DefaultMissions.ini if you are interested.

Bring Specialists for Sectopods and Psi Operatives for Gatekeepers. As a backup plan, put AP rounds on your Snipers. I used to think they were the most boring ammo type, but they are quite possibly the most powerful. Shredding things is a fools errand at some point. The game tells you 'damage is applied after shredding', but that just doesn't seem to be true. Magnetic cannons do 6-8 damage and I routinely shred 2 armor and apply only 4 damage. Also, Sectopods and Gatekeepers have like 5-6 armor so even if it did, AP rounds is still the way to go.

@zevvion: I usually opt for Bluescreen rounds vs Sectopods since they ignore armor and do +5 damage. I typically bring a mix of ammo depending on what enemies are in the mission, though.

@mike: @zevvion: Yeah I often have a mix of AP (usually on snipers), Bluescreen (if there is a sectopod), and Dragon Rounds. I put Tracers on a guy who needs the aim boost.

Bluescreen adds 5 damage to robotic enemies. It doesn't ignore armor. It's why I found them somewhat pointless. At least with my playstyle. I tend to ignore armor entirely, so for me, AP rounds are almost strictly better than Bluescreen rounds since they basically 'add' 5 damage to any armored enemy type, not just robotic ones. Especially since Gatekeepers and Andromedons can get a lot of armor and aren't robotic, I like it. I tend to hack robotic units anyway, so Bluescreen almost always is just a waste for me.

@zevvion: Yeah sorry about that, I meant to say shields, not armor.

They do also lower hack resistance, which can be useful for Sectopods, where I tend to have pretty bad luck taking control of them (although I don't often fail to shut them down).

@thomasnash: True enough. I like controlling Sectopods. I have a pretty decent 46% chance to control one at this point. Which apparently is high, because I have a Colonel Specialist with the upgraded Skulljack and her hacking stat has been improved by 20 twice.

This is fascinating. Thanks for breaking it all down and laying it out like this!

I'm playing on Veteran and regret nothing...

On my second playthrough (after failing/quiting the first one) the game is most definitely doing some bullshit in my favor that it wasn't before. Last time i got a fully armored soldier take 1 damage and get gravely wounded, this time i've had a ranger get half their life gone (wounded) then actually get put into bleed out (lightly wounded) lol. I find it crazy how different the game is this time around, although i didn't do the tutorial mission this time which is the only discernible difference besides a few cosmetic mods and voice packs. I wonder if doing the tutorial actually changes the numbers any (I highly doubt it)?

@oursin_360: Wound classification is weird, as it is only partially determined by how much health your solider has actually lost. When damaged, you roll a PointsToHeal value which is a range that is determined by the percentage of health lost. The issue is that these ranges overlap:

For a wounded solider with more than 75% health: WoundSeverities=(MinHealthPercent=76, MaxHealthPercent=10000, MinPointsToHeal=1000, MaxPointsToHeal=10000, Difficulty=1)

For a wounded solider with 20% or less HP: WoundSeverities=(MinHealthPercent=-10000, MaxHealthPercent=20, MinPointsToHeal=10000, MaxPointsToHeal=40000, Difficulty=1)

As you can see, it's actually possible to roll the same value in both cases: you'd need a shitty roll, but you absolutely can hit 10,000 PointsToHeal on a one damage tick, the same as the minimum roll for a trooper that gets completely blasted down to 1HP.

These PointsToHeal are then converted into days depending on a number of factors, including difficulty. There are mods out there which eliminate this overlap, which reintroduces a bit of sanity back into this aspect of the game. It isn't so much that your trooper gets "Gravely Wounded" or "Lightly Wounded" so much as those labels are assigned to dice rolls which are messed up.

@mirado: WOW, yeah on my next play through i'm definitely using those mods. That IMO shouldn't be relegated to a dice roll, or it should be a more indepth calculation. When that kept happening it just kinda dragged the fun out of the game knowing that minus 1 health could put my best dude out for a month. This time around though since i know more about the early enemies I've been able to avoid most damage and have even gone through most matches without taking a single hit.

Another thing that isn't said but I've tested is that on Veteran and Rookie when you load a save. You get random seed again. I didn't test if it does it on Commander or not yet.

On Legendary, you don't. If you missed a shot you still missed it. I found that out by reloading a save after a crash and for the same attack, the advanced repeater (10% chance) killed the target twice in a row while on Veteran it wasn't doing that. I verified if it was indeed working that way by doing the same move a total of 10 times and every time, my repeater would execute the target. Odds of this happening is 1 in 10 billions so I believe that it's a difference between the modes.

Also, while experimenting with the difference on the save/load seed. I found out the even if the hit or miss is fixed with the save on Legendary. The damage output of the attack isn't. So if you hit for 3 damage the first time, you can hit for 5 damage the next reload of the save.