Wanted: people to test Blossom

[Tables]

Nfl champ, it appears Ike's bases have suddenly dropped. Why?

You misinterpreted his post, is why, he made 21 levelups and got that many growths in each stat.

[grandjackal]

Soren Data

-Magic Skill Speed Luck Def Res

-Mag Speed Res

-HP Str Mag Skill Res

-HP Skill Speed Luck Def Res

-HP Mag Skill Speed Luck Def Res

-Str Mag Lck Def Res

-Hp Str Skill Speed Luck Def Res

-Hp Str Skill Speed Luck Res

-Hp Mag Skill Speed Lck Res

-HP Skl Res

-Hp Mag Skl Def Res

-Str Mag Skl Speed Def Res

-Hp Mag Skl Luck Def Res

-Hp Str Mag Skl Res

-Str Mag Skl Spd Res

-Hp Str Mag Skl Lck Def Res

-Hp Mag Skl Spd Lck

-Str Mag Skl Spd Lck Def Res

-Str Mag Skl Lck

-Str Skl Spd Def Res

12/20 levels HP

8/20 levels Str

15/20 levels Mag

18/20 levels Skill

12/20 levels Speed

12/20 levels Luck

11/20 levels Defense

18/20 levels Res

Results: I had to have gotten severely unlucky with the Magic, as everything else basically certainly seems boosted.

All I know is, I certainly noticed a higher average in points per level, and a LOT higher than just being an additional .1 for each growth.

Figures, only with my luck could I not proc an 80% growth fully even with Blossom...

Regardless, seems my theory that growths get a 1.5x boost is wrong, as otherwise Soren would have never missed a magic level, and he would perhaps have even gotten a +2 in there.

Edited April 8, 2010 by Cait Sith

[Tables]

Hmm... Standard Deviation is np(1-p), right? In which case

Growth = Grw
Blossom Growth = Blossm
Standard Deviation  = SD
Grw | Blossm | SD
5% |  9.75% |  8.799375%
10% | 19.00% | 15.390000%
15% | 27.75% | 20.049375%
20% | 36.00% | 23.040000%
25% | 43.75% | 24.609375%
30% | 51.00% | 24.990000%
35% | 57.75% | 24.399375%
40% | 64.00% | 23.040000%
45% | 69.75% | 21.099375%
50% | 75.00% | 18.750000%
55% | 79.75% | 16.149375%
60% | 84.00% | 13.440000%
65% | 87.75% | 10.749375%
70% | 91.00% |  8.190000%
75% | 93.75% |  5.859375%
80% | 96.00% |  3.840000%
85% | 97.75% |  2.199375%
90% | 99.00% |  0.990000%
95% | 99.75% |  0.249375%

[grandjackal]

Ugh, I really do have shit luck with magic if that's the case.

If this is true, this is what I had to work with.

HP: 64%

Str: 43.75%

Mag: 96%

Skill: 84%

Speed: 57.75%

Luck: 57.75%

Def: 43.75%

Res: 91%

Which case I was probably slightly screwed in a lot of stats, blessed in defense, did decently with Res, and got uber-screwed with magic.

...Regardless, that is a pretty epic growth spread for Soren.

[Narga_Rocks]

Well, random anecdotal evidence supports the theory. Lots of people say they get ~5 stats per level for Astrid while using blossom.

.7975	.64	.36	.6975	.64	.91	.51	.75

If the 2 RN theory is correct then those are her growths, and the total is 5.305 which is just over 5 per level on average. (And she caps everything but mag and def on average during tier 2. Yay Astrid.)

[grandjackal]

Would mean that she also caps HP and Res 4 levels sooner, at level 11. Let's see what her other stats would be by then.

18.75 Str, 11.24 Mag, 22.2775 Skill, 20.76 Speed, 25.19 Luck, 14.59 Def

Problem is, she only has access to Blossom from 3-11 onward...So this isn't even a likely scenario.

Edited April 8, 2010 by Cait Sith

[Tables]

However, a thought occured. As for it's use and who would enjoy having it the most...who exactly would it be? Any ideas?

It's hard to say, due to caps and everything, but...

[spoiler=3 Comparisons]

Total growths with Blossom:

1: Sanaki (586.5%)

2: Vika (574%)

3: Nolan (570.5%)

4: Elincia (565.25%)

5: Leonardo (552.5%)

...

69: Leanne (382%)

70: Rafiel (381%)

71: Reyson (376.5%)

Proportional Increase in growths with Blossom:

1: Giffca (1.65 times higher, 396%)

2: Caineghis (1.633 times higher, 392%)

3: Nailah (1.626 times higher, 406.5%)

4: Ulki (1.61 times higher, 442.75%)

5: Naesala (1.598 times higher, 447.5%)

...

69: Mordecai (1.399 times higher, 440.75%)

70: Micaiah (1.351 times higher, 540.5%)

71: Volug (1.35 times higher, 438.75%)

Direct increase in growths with Blossom:

1: Ilyana (189.5% increase, 549.5%)

=: Tormod (189.5% increase, 539.5%)

3: Sanaki (186.5% increase, 586.5%)

4: Pelleas (184.75% increase, 519.75%)

5: Fiona (184% increase, 544%)

...

69: Skrimir (123.75% increase, 388.75%)

70: Leanne (122% increase, 382%)

71: Volug (113.75% increase, 438.75%)

I included worst as well. Herons and beast Laguz look like the least benefiters in general, while Mages seem like some of the best bets.

[Light Lord]

I can give it a try, started once more RD.

I tested it in the cats (Lyre and Ranulf) but with really few results and not worthy enough

May test it in a beorc.

Name the class and I'll see what I can do.

[Silith]

Units with many Middle growthsa seem to get the most benefit out of it. Idealy a 52% growth would gain the most but 50 and 55 are the winners here.

Ilyana has 55-50-50-45-40-60-30-30 as growths (from most Blossom optimal to least blossom optinal) and therefor comes out as biggest gainer.

The units who have a few monster growths coupled with a few crappy ones gain a lot less in comparison. At best Blossom is going to make an average growth into a good growth but it won't bump up low growths that much and it won't push High growths above 100.

Still checking Vika. I will stop at 300 level-ups.

Did some extra and got some results.

Here you see the last levels she got. a 1 means it went up and a 0 means it didn't.

Then in line 353 I placed the total amount of 1's in that colom.

Line 355 has the % Growth with Blossom.

Line 356 has growths without blossom according to Serenes Forest.

Edited April 9, 2010 by Silith

[Interceptor]

If that's true, it explains why I always get excellent results out of Mist with Blossom when I raise her. Her growths are pretty solid other than STR/DEF, and even those two are high enough to get a significant benefit if the theory about Blossom is correct.

Edited April 9, 2010 by Interceptor

[Tables]

Slith, where did you get 52% from? I'm fairly sure 50% is the best, it's easy enough to prove it's better than 52% (52% goes to 76.96%, so a 24.96% increase, 50% goes to 75%, so a 25% increase), and calculus shows 50% is the optimal value, with other growths distributed evenly around it (48% goes to 72.96%, a 24.96% increase, like 52%).

[grandjackal]

Well that's the thing, how does one determine what a character wants out of Blossom? I think it depends more on the character than anything. I think the closer it comes to basically doubling the character's growth, the more impact it has.

That in mind, I'd say the best cantidate would be Soren, as it only ensures that the growths he's close to capping cap sooner in levels while also substantially boosting his other stats (namely, he has about 50% speed now), which would overall make him cheaper+less dependent on BEXP.

Only problem is simply that it kills leveling speed, thus why I suggested that the closer it is to nearly doubling people's growths. Then again, how about we test the leveling speed to see just how much Blossom slows it down?

[VincentASM]

Thanks for all the help, I now have some nice, large numbers for a bunch of growth rates. I'm still in need for more numbers, so please keep sending them in ^__^

Also (and wrong board), are there any volunteers who would like to test Blossom in FE9? I can do this myself, but the more numbers, the better.

Current tallies:

Original growth -> Effective growth with Blossom (# of tests) Theoretical growth with 2 RNs
0  -> 0%     (25)  0%
5  -> ---    (0)   9.75%
10 -> 25.35% (71)  19%
15 -> 33.33% (21)  27.75%
20 -> 40%    (45)  36%
25 -> 43%    (100) 43.75%
30 -> 54.10% (61)  51%
35 -> 50.50% (101) 57.75%
40 -> 55.67% (106) 64%
45 -> 80%    (20)  69.75%
50 -> 86.67% (45)  75%
55 -> 76.19% (21)  79.75%
60 -> 86.49% (111) 84%
65 -> 85.24% (61)  87.75%
70 -> 92.5%  (40)  91%
75 -> 80%    (20)  93.75%
80 -> ---    (0)   96%
85 -> ---    (0)   97.75%
90 -> ---    (0)   99%
95 -> ---    (0)   99.75%

(I made a stupid mathematical assumption before, with adding averages, which I've fixed.)

EDIT

Missed Silith's numbers. How on earth did you continue for that many number of Level Ups? o__o

Also, is that 350 Level Ups or 351 (in case you skipped a row or used it for something)?

Edited April 9, 2010 by VincentASM

[Silith]

on Row 1 I have HP Str Mag Skl SPd DEf and REs.

Just to make sure I list the correct stat. After a while it goes correct automatically. It's 350 level-ups.

Just redid the first turn on 4-E-1 enough times to get a heapload of data. and the 52% thing was a quick false calculation. You are correct. 50% gets the most benefit. 55 and 45 are still pretty impressive (almost as impressive as Danved).

EDIT: Forgot to mention. I rechecked my list twice and it's still correct.

Edited April 9, 2010 by Silith

[VincentASM]

Alright, updated data:

Original growth -> Effective growth with Blossom (# of tests) Theoretical growth with 2 RNs
0  -> 0%     (25)   0%
5  -> ---    (0)    9.75%
10 -> 25.35% (71)   19%
15 -> 27.23% (371)  27.75%
20 -> 40%    (45)   36%
25 -> 43.78% (450)  43.75%
30 -> 54.10% (61)   51%
35 -> 50.50% (101)  57.75%
40 -> 55.67% (106)  64%
45 -> 80%    (20)   69.75%
50 -> 76.45% (395)  75%
55 -> 76.19% (21)   79.75%
60 -> 84.15% (1161) 84%
65 -> 87.38% (761)  87.75%
70 -> 92.5%  (40)   91%
75 -> 80%    (20)   93.75%
80 -> ---    (0)    96%
85 -> ---    (0)    97.75%
90 -> ---    (0)    99%
95 -> ---    (0)    99.75%

Silith's data by itself practically tells us that we're on the right path.

[Anouleth]

If that's true, it explains why I always get excellent results out of Mist with Blossom when I raise her. Her growths are pretty solid other than STR/DEF, and even those two are high enough to get a significant benefit if the theory about Blossom is correct.

I remember trying Blossom on Mist to up her crappy strength and it utterly failing, but all her other stats were good. Maybe I was just unlucky.

And it would be more worthwhile to try and get Mist to promotion as quickly as possible since she has h4x promotion bonuses. I'd rather use Blossom to finish off a level on someone who levels really slowly either way, like Titania (75% chance of proccing speed is nice).

[Tables]

This is probably going to make little sense to most people, but using Blossom seems to end up, in general, with a prisoners dilemma type situation: Your levels are never increased by more than double, and that's only for exceptionally low growths, and without Blossom you level twice as quickly. Therefore, regardless of your level, it's in your best interest to not use Blossom. However, if you use Blossom non-stop on one person, said person will eventually be so far behind levels-wise that they level at the same speed as everyone else in the party, but gaining about 50% more stats every level, clearly giving you an advantage once you catch up the lost levels (by my estimate, you'd tend towards something like 10 levels behind, after which you gain exp at about the same rate as everyone else). If you want to, as you get towards the end of the game, you can take it off and take advantage of double-speed levelling for a while, too.

So taking Blossom off at any time (akin to giving evidence in PD) is advantageous in every way, but actually keeping it on (akin to staying silent in PD) would be better in the long run.

[Interceptor]

And it would be more worthwhile to try and get Mist to promotion as quickly as possible since she has h4x promotion bonuses. I'd rather use Blossom to finish off a level on someone who levels really slowly either way, like Titania (75% chance of proccing speed is nice).

This is true, but when I train Mist, I do it without any pretense of efficiency, since she is not an optimal character and I am only doing it because I love Mist more than jellybeans. In other words, I bring her to 3-8 with Blossom and a bazillion staves, and just have her chain-heal people who get hit by lava rocks, until she naturally hits promotion. It takes a while, but she's completely badass as a Valkyrie.

[grandjackal]

This is probably going to make little sense to most people, but using Blossom seems to end up, in general, with a prisoners dilemma type situation: Your levels are never increased by more than double, and that's only for exceptionally low growths, and without Blossom you level twice as quickly. Therefore, regardless of your level, it's in your best interest to not use Blossom. However, if you use Blossom non-stop on one person, said person will eventually be so far behind levels-wise that they level at the same speed as everyone else in the party, but gaining about 50% more stats every level, clearly giving you an advantage once you catch up the lost levels (by my estimate, you'd tend towards something like 10 levels behind, after which you gain exp at about the same rate as everyone else). If you want to, as you get towards the end of the game, you can take it off and take advantage of double-speed levelling for a while, too.

So taking Blossom off at any time (akin to giving evidence in PD) is advantageous in every way, but actually keeping it on (akin to staying silent in PD) would be better in the long run.

10 levels behind you say? So, would it then not be best to give it to low level schmucks, like Soren or the CRK, guys who already start essentially behind like that?

Granted that the CRK have two things of Paragon, but...

[VincentASM]

Results for FE9 Blossom:

Original growth -> Effective growth with Blossom (# of tests) Theoretical growth with 2 RNs
0  -> ---    (0)    0%
5  -> ---    (0)    9.75%
10 -> 17.5%  (200)  19%
15 -> ---    (0)    27.75%
20 -> ---    (0)    36%
25 -> ---    (0)    43.75%
30 -> 61%    (200)  51%
35 -> 59.5%  (200)  57.75%
40 -> ---    (0)    64%
45 -> ---    (0)    69.75%
50 -> ---    (0)    75%
55 -> 83.75% (400)  79.75%
60 -> 80%    (200)  84%
65 -> 83%    (200)  87.75%
70 -> 93%    (200)  91%
75 -> ---    (0)    93.75%
80 -> ---    (0)    96%
85 -> ---    (0)    97.75%
90 -> ---    (0)    99%
95 -> ---    (0)    99.75%

Assuming Wonder Egg didn't make any mistakes in their RNG abusing, I think it's safe to say Blossom rolls a 2nd RN if the 1st roll failed. I'm not sure if they tried RNG abusing in FE10 though.

I rarely play FE9 Random mode, so I never realised how much larger of an effect Random!Blossom has compared to Fixed!Blossom, which divides the growth rates by 90. Confusingly, Blossom also has the internal name FRAC90.

EDIT

Updated FE9 figures.

Edited April 9, 2010 by VincentASM

[Tables]

10 levels behind you say? So, would it then not be best to give it to low level schmucks, like Soren or the CRK, guys who already start essentially behind like that?

Granted that the CRK have two things of Paragon, but...

It's just an estimate, really, making lots of assumptions, many of which were false. And no, low levels won't make it worse, in fact, they're the best people to use it on, since their levelling speed will be somewhat constant rather than starting very slow and speeding up.

[Nitrodon]

In case you wanted to know, a 95% confidence interval for the true growth rate for 60% (the growth rate with the most data collected) with Blossom in FE10 is (81.922%, 86.207%).

Edited April 9, 2010 by Nitrodon

[Zanarkin]

BTW i was wondering there is no problem by giving the 99 EXP points and then attacking/writing down the level ups/and restarting right?

Edited April 10, 2010 by Mordecai

[grandjackal]

Well, that's what I did.

[Zanarkin]

Oh all right, i thought it could affect the results in a way. Anyways i'll get working on them.