What is the argument for 50% being a bad growth?

[Chiki]

>_>

I legitimately can’t tell if you’re being serious.

Answer this, suppose you have a 50% chance ticket to win the 1 billion dollar lottery or could trade it for any % below 50%. Would you take anything below 50% because of less “risk” or because 50% is a "coinflip"?

Your interpretation of probability makes no sense. Please clarify your logic.

This post is a little immature. I'm pretty sure Kai was just presenting an argument he thought was someone else's.

The argument is hilariously bad, though. It's basically saying that 50% is worse because of a psychological bias against coinflips.

this is the argument for 50% being bad, though for me its not bad but a gamble.

Edited May 24, 2014 by Chiki

[Zaprong]

Basically what Kai said was in other words:

30% growth: Hey, that stat will not go up that easily, so you better watch out for that stat!

70% growth: This stat will most likely cap, which is always a good thing.

50% growth: Oh ummm, it should be 1 stat up for each 2 levels, right? In theory that would be correct, but... what do you expect from 50%? Balanced stats?, it's too hard that every 2 level ups you get 1+ in that stat... maybe you will not get any stat in those 2 levels... or might as well get 2+ in that stat.... I can't tell, the probability for succeed is 50% but for failure it is too...

So basically it is a gamble because you are trying to get something without being able to expect what is gonna happen. If the growth rates were like 30% it's not so much of a gamble anymore, because you know that it has a low probablity to succeed and thus you know what will probably happen-something that people need in order to analize a character on any fire emblem game.

I hope this makes sense somehow

[eclipse]

The only time I'd consider 50% to be bad is if I need a precise stat at a given level (literally cannot be higher or lower than X), and I don't have any of the nifty RNG tools available. The odds of me running into this situation is really low.

Edited May 24, 2014 by eclipse

[Redwall]

Basically what Kai [tried to say]:

I think what some of you are trying to say is that 50% growth has the highest information entropy of the possible FE growths. But as has been noted, regardless of information entropy or variance, having a 50% growth is strictly better than having a growth rate lower than 50%, unless you (say) want a low Def stat to bait enemies.

Edited May 24, 2014 by Miikaya

[Zaprong]

I never said it was worse than a 30%. But it's just difficult analizing a coinflip.

As in: 50% isn't a bad growth but it's really difficult to make predictions on how a character will turn out (See FailWood and Eliwin, there is never a normal Eliwood)

[Redwall]

I never said it was worse than a 30%.

I never said that you said it was worse than a 30% growth.

But it's just difficult analizing a coinflip.

As in: 50% isn't a bad growth but it's really difficult to make predictions on how a character will turn out (See FailWood and Eliwin, there is never a normal Eliwood)

Without being too pedantic, this is more or less what I said. No need to repeat ourselves here.

Edited May 24, 2014 by Miikaya

[Zaprong]

I was just using my words if there were people that still were confused and wow now I notce that my post actually came really late.

That was meant to be in the first page whooops.

Seems like I got carried away thinking I was on page 1, somehow without noticing the other posts, my bad.

[Nicolas]

stuff

Listen this guy!

He know what is math and probablity!

[XeKr]

@Thor: That’s why it’s preferable to calculate binomials instead of averages, as possible/convinient.

@Chiki: <_<

@Eclipse: Fair enough, though I don’t recall this ever happening myself (across the game, wanting a lower growth).

@Miikaya: Whoa, let’s not bring actual math into this.

Edited May 24, 2014 by XeKr

[Baldrick]

The only time I've had a situation eclipse describes is in FE4 drafts. Sigurd having too much defence means some enemies won't attack him and instead block his way.

[eclipse]

I. . .just found my theoretical situation.

Both Priscilla and Fiora have a 50% Res growth, and they both start at 7 Res. I'm going to grab Priscilla for my purposes, since she's at a lower level. Let's say that I need to put Pent in range of a Berserk staff, but I don't want that staff to target him. As long as Priscilla's RES is less than Pent's (16 at base), then she should be able to wander into range and draw the "blow". This means that, at most, she can be blessed up to +1 in Res, unpromoted at level 19/20 - any more, and she'll match Pent's Res, and I don't know what the AI will do.

Odds of this happening? REALLY LOW. Completely improbable? Nope.

Edited May 24, 2014 by eclipse

[HF Makalov Fanboy Kai]

to clear up the confusion, chiki nails it on the head, i was just reciting the anwser to what the OP's question was.

its clearly not a bad growth, but to me its a gamble, thats it really.

sheeh this is worse when i got attacked and flamed for my theory that soren was suppose to be a girl, cause at least that was something i actually believed in.

[Jave]

All growths are a gamble (except 0% or 100%), it's just that the odds are different depending on growth.

50% is a coin flip, sure, but 17% is close to a dice throw.

Edited May 24, 2014 by Jave

[dondon151]

I. . .just found my theoretical situation.

if FE7 staff AI is anything like FE6 staff AI, it targets units in reverse deployment order with a certain res threshold, so this situation largely doesn't even apply.

[MisterIceTeaPeach]

Growth rates are just an arithmetic average like hitrates. If an unit has a 50% growth in a stat, it is not guaranteed to get the stat on each second level up.

[PKL]

Evidence.zip of it being a good growth:

75 50 50 45 25 20 5 FE6's own Miledy, must be a Kappa in str and skl right?

90 50 45 45 25 40 30 FE8's Seth. Man, his str must be so lame right?

80 45 25 60 50 45 40 45 Titania has a 50 spd growth. Damn, her spd must suck.

This is obviously a joke post.

[Nicolas]

All growths are a gamble (except 0% or 100%), it's just that the odds are different depending on growth.

50% is a coin flip, sure, but 17% is close to a dice throw.

That's why every FE player loves units with good base.

Growth are always shaccky, no matter if that 's 35, 50 or 60%. I once have Joshua with more Str than Garcia on same level. I once have Franz with 11 str at 10/1. However I've never have dissapointing Duessel.

[eclipse]

if FE7 staff AI is anything like FE6 staff AI, it targets units in reverse deployment order with a certain res threshold, so this situation largely doesn't even apply.

Probably would work better with Silence, then; IIRC, they don't bother with non-mage units. Still an extreme corner case.

That's why every FE player loves units with good base.

Yeah let's not lump the entire fandom into one sentence, kthx.

[XeKr]

okay I lied, it’s a little interesting.

I. . .just found my theoretical situation.

Both Priscilla and Fiora have a 50% Res growth, and they both start at 7 Res. I'm going to grab Priscilla for my purposes, since she's at a lower level. Let's say that I need to put Pent in range of a Berserk staff, but I don't want that staff to target him. As long as Priscilla's RES is less than Pent's (16 at base), then she should be able to wander into range and draw the "blow". This means that, at most, she can be blessed up to +1 in Res, unpromoted at level 19/20 - any more, and she'll match Pent's Res, and I don't know what the AI will do.

Odds of this happening? REALLY LOW. Completely improbable? Nope.

Okay maybe I’m totally confused, but…Miikaya or someone tell me if I’m butching statistics

1. Priscilla has 6 base res according to Serenes. Fiora has 7 base Res.

2. Priscilla’s average Res at level 19 is 14. Fiora’s is 13. (okay true averages b/c of caps, etc, w/e)

3. How I would try to describe things is: At 50% growth, the chance of Priscilla getting 10 or more Res procs in 16 levels (to have 16 or more Res or we can say “too much” Res at level 19) is around 22.7%. The chance of Fiora getting 9 or more Res in 12 levels, for the same benchmark, is around 7.3%. These are low but not completely improbable, true (not even that low).

4. If the growth was 60% instead, the chance of them having “too much” res is 52.7% and 22.5% respectively. Which are higher than for 50% growth and “worse”.

5. If the growth was 100% instead, there’s a100% chance, with 0 variance, they would have more Res than Pent and couldn’t Staff bait. I guess that’s bad?

6. If the growth was 0% instead, there’s a 0% chance, with 0 variance, to have more Res than Pent. They can always staff bait. Yay?

7. There is nothing special about 50% here (besides the aforementioned notes about variance and information entropy)

In this specific case, lower growths might be better in a sense, but because they’re lower, not because they have less variance. The flip side, higher growths being worse (for this specific case) is also true, despite them having lower variance. Could pick growths like 49% and 51% and see the same trends.

To clarify an earlier statement of higher growths being “strictly better”: you will more consistently have higher stats with some growth (including 50%) than all growths lower (independent of if higher stats are strictly preferred over lower stats, or not, for uses such as bait. Even then, that’s a specific case and there may be many other cases, across the game, where the higher growth and stat is preferred)

to clear up the confusion, chiki nails it on the head, i was just reciting the anwser to what the OP's question was.

its clearly not a bad growth, but to me its a gamble, thats it really.

sheeh this is worse when i got attacked and flamed for my theory that soren was suppose to be a girl, cause at least that was something i actually believed in.

If we’re being a bit more clear about things, the implication was I disbelieved someone would seriously state/take a position that could be considered (in the succinct words of others) “garbage” or “hilariously bad”. The fact remains however, that I’m actually seen it around, propagated by others. So apologizes if you felt singled out, but my posts were intended to challenge that position.

Still, (imo) that’s quite mild of an attack compared to the content of some other posts we might see elsewhere in SF.

Also, I admit I still don’t fully understand your interpretation of probability (the one you refer to with “I”, whether you believe it or not I dunno), including why 50% is more a “gamble” or “risk” and statements like “i can guranteed that Gatrie in FE9 is gonna cap his defense.” (just 60% growth and apparently 87% chance to cap, more because of his base than anything else).

EDIT: btw +Def/-Luck Avatar is obvs the most unreliable character ever, with Str/Skl/Spd/Luck/Def all at 50% growth, and Mag at 45%. dat risk.

EDIT2: formatting...

Edited May 25, 2014 by XeKr

[Redwall]

Also, I admit I still don’t fully understand [Kai's] interpretation of probability (the one you refer to with “I”, whether you believe it or not I dunno), including why 50% is more a “gamble” or “risk”

I think the use of words like "gamble" and "risk" is an attempt at summarizing the intuition behind the information-entropy maximum. Looking at the outcome of a single level-up, for a known growth rate p with p =/= 0.5 and 0<= p <= 1, the information entropy is less than one full bit, since you can predict the outcome more often than not. For p = 0.5, it is because you cannot predict the outcome more often than not that the information entropy is maximized wrt p.

Edited May 25, 2014 by Miikaya

[XeKr]

A bit meta here, but I ninja'd a glance at your post pre-edit which is somewhat relevant. >_>

Anyways, fair enough, I think my difficulty/confusion stems from conceptually reconciling the idea of information entropy with certain supplementary and clarifying statements such as those re: Gatrie's Def.

Edited May 25, 2014 by XeKr

[Redwall]

Yeah, the edit was mostly to avoid suggesting that information entropy had anything to do with the Gatrie thing, which I thought out of place in a topic about 50% growths.

[The Geek]

If 50% is considered a bad growth, then 75% of Fire Emblem characters are terrible, because I don't think growths that high are all that common.

[PKL]

If 50% is considered a bad growth, then 75% of Fire Emblem characters are terrible, because I don't think growths that high are all that common.

And Avatar FE13 would be literally the worst.

[Jotari]

That does make me wonder, who in Fire Emblem has the honor of having the absolute worst growths?