The RNG in this game...

[cgRIPPER]

Am I the only one that feels like whatever random number generator this game uses is flawed?

I've been playing FE since blazing sword came out in America (more than half my life) and I have never had such a ridiculous string of bad luck, and it is still coming! For example, I went into an arena match and got a match that should have easily been my win. We both dealt a similar amount of damage to each other, but the enemy had only 44% hit rate while I had a 92%... I missed 2 attacks while the enemy hit all 7 out of 7! That's just 1 specific example as well, I've had to reset during the actual chapters due to attacks with 30%, 40%, and even 15% hitting back, to back, to back, to back. It feels like they intentionally nerfed EVERYONE'S ability to dodge for the sake of making tanky characters more prominent.

Am I going crazy or are other people experiencing similar things?

[shadowofchaos]

You might want to review "True Hit".

Because there is a chance that this game is back to single RNG for hits.

[HF Makalov Fanboy Kai]

i've had the feeling it uses a single RNG for this game just because i'm replaying FE6 for my boyfriend, and its hitrates and this game's hitrates are pretty similar.

[Water Mage]

It's not flawed, rather it's "real" in a manner of speaking.

Previous Fire Emblem games had a second, hidden RNG that was geared in the player's favor.

It's what shadowofchaos called "True Hit".

[Griffonite]

You're definitely not the only one that feels that way. I've seen quite a few posts on here saying as much. For me though, it seems to be having the opposite effect. I've had 23% hits and 96% dodges. I also once had quite a string of 4~6% crits courtesy of Kaze.

I don't know how the RNG changed in this one (if it did), but I personally hope that it stays that way. This is the luckiest I've ever been in a game.

[Topazd]

i've had the feeling it uses a single RNG for this game just because i'm replaying FE6 for my boyfriend, and its hitrates and this game's hitrates are pretty similar.

FE6 uses 2 RNs though, unless that wasn't what you were getting at.

[XeKr]

Anecdotes that confirm or deny true hit are a wonderful study in confirmation bias. Complaining about RNG is particularly biased because there’s nothing to complain about if things go as “expected”.

I think it’s still in though, from general impressions and reseting C2 to figure out the 3 turn strat in Lunatic (like mostly 90% disp hits).

http://serenesforest.net/forums/index.php?showtopic=55887&p=3898379

Here is a post where sylvanllewelyn pointed out that 2rn even underestimates the hit rate for 65 data points (edit: the poster claimed >100 but the 2 datasets seem identical to me). This is stronger evidence because we can observe clearly all the misses and hits, rather just un/lucky ones people choose to report. I threw together a script for the averages and got…

1RN     Actual Hit%   2RN
72.91   87.69         81.79

So 1RN/Displayed hit is way off, while 2RN/True hit is closer but again is lower that reality (for this data set, which isn't that large, but more objective than "DAE get screwed by RNG"). if I really get bored I'll run a CI.

Also FE6 has true hit, it just has low hit rates. I do think Fates has a less extreme version of that, more 80%s than 90%s, so it just seems like there are more misses overall, and then people start looking for/being surprised by particularly “rare” outcomes (as a single event, but not really across the game).

edit2: Okay it seems trivial but my stats is rusty so eh. I got [0.794884939055563,0.958961214790591], so it's quite a bit outside the 95% confidence.

Hmm yeah maybe I just suck at stats (and I'm not sure it applies in this context with 0/1s) but I'm getting [0.757519973720077,0.996326180126077] to be the 99.5% CI (actually can go to like 99.9% but I don't like upper bounds >1)... Regardless, data seems to suggest 2rn and I'll numbers/math later.

Edited February 24, 2016 by XeKr

[Shadow Knight]

It's like Xcom. You have to put yourself in situtations where you can maximize your chances. Low percentage chance of hitting? Position a soldier next to your attacker to either whittle down the enemy or get a second roll. If you position your forces correctly you can end up with plenty of attacks per turn.

But the one thing that does bother me about the RNG is stat growths. I've started to get better luck and my Avatar has been solid but my other characters have been getting hammered during the level ups. You can only position so many times before you're simply outclassed or worn down by debuffs. Gods help you if you anger the RNG level up god.

[Loki Laufeyson]

I think its True Hit, just not rounding up

[Anacybele]

Well, this explains why I've been having more bad luck in this game than any other. And why Ryoma took several hits in a row that had a pretty low chance of striking and almost got killed. He lost his helmet tho, which was funny. I like how the units lose pieces of armor and have their clothes rip, it just makes me chuckle for some reason. XD

[Sieghart]

The RNG can be a really troll In this game yesterday , was going to finish a boss In Soleil paralogue the enemy hit % was 33 and my aim was 87 when i trigger the battle i miss and the boss hit me and worse he do a critical, killing my unit. And for that i have to do all paralogue again.

[cgRIPPER]

This has been quite informative, I had noticed throughout my years playing FE games that it seemed like hit percentages were less true at values that were further away from 50/50, but I didn't know the specifics of why that happened.

So maybe, it's a combination of the 2RN system not being there to hold my hand and happening to get a string of bad luck at the same time that has resulted in such absurd things happening all of a sudden in this game.

[LuxSpes]

I've been wondering about the RNG myself since I've missed way more 80%+ hits and been hit by 10% or less hits way more than I can remember in any FE I played.

I'm still salty about the time I had to restart Birthright Chapter 25 after an hour of careful play because one of the last enemy on the map got a 1% crit.

[SatsumaFSoysoy]

No one really had an answer on whether true hit is used in this game back when the Japanese version was released, for some reason. Guess the dataminers didn't bother looking? Anyway, I'm inclined to believe true hit is gone, but that's just me.

[cgRIPPER]

Right after I made my last post I got hit by a 4% crit... BACK TO BACK twice in a row... smh

[GoXDS]

just last night I was having the worst luck. getting hit by 29% Hit about half the time, getting hit by 16% as well. one time I get killed by a 33% so I restart. then a freakin Berserker oneshots Ryoma at 16% Hit and 4% Crit. not to mention all the single pt lvl ups.

because of that, when I was doing endgame today I was terrified of even 1% crits and played so slow >_>

[shadowofchaos]

Right after I made my last post I got hit by a 4% crit... BACK TO BACK twice in a row... smh

...Crits have always been single RNG though.

[Sentacotus]

Yeah I've been having similar experiences myself. Meanwhile, I gave a killer ax to Charlotte with the worst hit rate known to man and just get my hits off crits. So the lesson here is Fates+Killer Weapons=Profit?

Edited February 25, 2016 by LordTaco42

[HF Makalov Fanboy Kai]

FE6 uses 2 RNs though, unless that wasn't what you were getting at.

whoops i feel retarded

[Hunter Nightblood]

Oh hey, two berserkers with an 18% hit rate I should be fi... why did they both hit?

Yeah...

Of course, it' not like we've had strange RNG situations before:

Edited February 25, 2016 by Hunter Nightblood

[XeKr]

Anecdotes that confirm or deny true hit are a wonderful study in confirmation bias. Complaining about RNG is particularly biased because there’s nothing to complain about if things go as “expected”.

I think it’s still in though, from general impressions and reseting C2 to figure out the 3 turn strat in Lunatic (like mostly 90% disp hits).

http://serenesforest.net/forums/index.php?showtopic=55887&p=3898379

Here is a post where sylvanllewelyn pointed out that 2rn even underestimates the hit rate for 65 data points (edit: the poster claimed >100 but the 2 datasets seem identical to me). This is stronger evidence because we can observe clearly all the misses and hits, rather just un/lucky ones people choose to report. I threw together a script for the averages and got…

1RN     Actual Hit%   2RN
72.91   87.69         81.79

So 1RN/Displayed hit is way off, while 2RN/True hit is closer but again is lower that reality (for this data set, which isn't that large, but more objective than "DAE get screwed by RNG"). if I really get bored I'll run a CI.

Also FE6 has true hit, it just has low hit rates. I do think Fates has a less extreme version of that, more 80%s than 90%s, so it just seems like there are more misses overall, and then people start looking for/being surprised by particularly “rare” outcomes (as a single event, but not really across the game).

edit2: Okay it seems trivial but my stats is rusty so eh. I got [0.794884939055563,0.958961214790591], so it's quite a bit outside the 95% confidence.

Hmm yeah maybe I just suck at stats (and I'm not sure it applies in this context with 0/1s) but I'm getting [0.757519973720077,0.996326180126077] to be the 99.5% CI (actually can go to like 99.9% but I don't like upper bounds >1)... Regardless, data seems to suggest 2rn and I'll numbers/math later.

Okay after sleeping on it (I think) I realize the obvious/stupid error with this line of thought. The idea of a population mean here doesn’t make much sense at all, given that the disp hit rates are changing.

A better experiment would be to test the same hit rate (something around 75% disp hit) repeatedly, and then you could observe the percentage of hits. Then you could say adopt the 1RN system as the null hypothesis, and reject it to a certain statistical significance in favor of the 2rn alternative. We would be assuming that it's the same system that accounts for 75% hit as the other hit rates, but that seems fairly safe as it would be silly to program otherwise. This should require far less than thousands of data points, as the 1 rn and 2 rn systems differ by a lot, around 12.75%.

Here's a link to BlenD's topic also, who's proposing a large scale data collection.

Edited February 25, 2016 by XeKr