Digimon Stats

[1Meguru]

Hi Hinaody, great work on this, very accurate and handy as a reference.

I was just wondering if it is possible to determine the relationship between the Base stats, particularly AT and CT, with the enchantment level?

As the enchantment bonus to stats depends on the base stat itself, is it possible for the bonus to be scalling or stable along with the base stat (determined by size at lvl 90)?

The enchantment scales with your stats, varies on how much though, dependds if you get good % increases when enchanting(For example, a lv6 Critical hit rate enchant seems to range between 95 and 105% bonus, so if you got 0.1 increased when leveling and had a 105% bonus your enchant would increase by 0.105)

[RajaXP]

The enchantment scales with your stats, varies on how much though, dependds if you get good % increases when enchanting(For example, a lv6 Critical hit rate enchant seems to range between 95 and 105% bonus, so if you got 0.1 increased when leveling and had a 105% bonus your enchant would increase by 0.105)

Very interesting! This would mean that it might eventually be better and worth investing on getting the higher rate of the enchantment, instead of increasing size after all! Thank you for your response .

[gameisme]

Sure. Basically, the assumption I make is that some stats are a linear function of size.

For example, let's take the HP stat. The two variables I am dealing with are called X and Y. X is the dependent variable (size here) and Y is the independent variable (HP).

X Y
100.000% 2877
119.580% 3287
124.091% 3383
122.502% 3350
125.109% 3405
125.370% 3410
120.301% 3305
127.900% 3461
125.089% 3404
121.724% 3335

For reason I will explain later, I took only the SA BM here and left out the QA BM.

With the data above, we can make a linear regression: Y=a*X+b. This can be done in Excel and the results we obtain is:

a=2098.8797959922
b=778.559245020504

This means that if you have the size, we can calculate the HP: HP=2098.8797959922*size+778.559245020504. As a sanity check, we can see that with MirageGaogamon BM (size=100%), we recover the actual size 2877.439041 (rounded down to 2877).

We can do linear regression with any data but it can be non-relevant if there is no linear relationship between two variables. That is why we need to look at some quality diagnostics. In this case, we have:

R-squared=99.99% --> the higher the better
p-values for a and b: 1.2789E-18 and 1.68007E-14 --> the lower the better

With the results we have here, we can conclude, we have perfect linear relationship (or almost nearly, depending on how picky you are). If you draw the points on a graph, you would see a straight line.

The reason why I left out QA is because, they have their own behaviour. If added to the sample, it would have caused distortion to the results and give lower quality diagnostics.

Other stats are a bit more difficult to deal with because they are more "discrete" while HP is more continuous. But the idea is the same.

More info at the usual place: http://en.wikipedia.org/wiki/Linear_regression

Hope it helps

why do i thnk that HP should be the Y, which is dependent variable..I dont really get it on that part. But, basically i understand what u have done so far with the linear regression.
One thing for sure is I did not expect someone could use regression on digimon data. That is awesome

[Hinaody]

You are right, I made a mistake in my explanation. X is the independent variable and Y is the dependent variable. I went a bit fast and didn't proof-read my post =/

[gameisme]

Well, its ok though.
btw, I keen to know the correlation between size and AT (possibly CT as well). Well, obviously its positively correlated but to what certain level. How does AT and CT changes when 1% of size increases? I assuming the rate of changes is based on SA and QA type, rather than base damage of a particular digimon.
Just curious how much is the differences in AT and CT for normal 4/5 to 5/5 at highest evolution, since the gap should be widened compare to previous evolution.
If u busy, just give me the data though

[Hinaody]

For CT, I think I have a better method than just linear regression (check first post where I explained in detail about CT). Using linear regression we would have found similar results. I think CT is pretty accurate.

AT is more difficult. First thing to notice is that AT is dependent of AS and AT/AS is actually the variable we should look at. This is fairly intuitive. Then we can go about using linear regression again.

I can look at my regression result again when I get back home but if you want to play with the maths, the original data are on the first post.

[Omanty]

why is it that i hear that MirageGaogamons BM is the worst?

[Hinaody]

why is it that i hear that MirageGaogamons BM is the worst?

Probably skill-wise.

[Raidenmon]

Basically...

Exemple:
Craniummon AT
100% = 875
130% = X

130x875 = 113.750
100X = 113.750
X = 113.750/100
X = 1137,5

A Craniummon 130% LV90 have 1137,5AT Basic. Easy.

[asdftt77gx]

Or,
130/100=1.3
1.3x875=1137.5