Talk:Attribute

Early thoughts
I'm attempting in finding out the formulas for the derived attributes. So far I know that for each point in Body raises HP by 79/14.

Experiment done at 36 body, 30 mind, and 30 spirit; raising body by 1 and rounding the result yields hp + 79/14. where hp at those stats = 269hp. At this point hp was 269 before rounding. Rcollins779x 18:41, 29 September 2007 (UTC)

Attribute Formula
8 + 2 x Level = Base Attributes

Example: Level 11 Base Attributes are... 8 + 22 = 30 (Body; Mind; Spirit)

HP Formula
Body and Spirit affect the base HP where Spirit attributes equal half of Body attributes. The rate of increase is 37/14 (2.642857...) for Spirit and 2x that for body.

((2 x Body Attributes + Spirit Attributes) x 37/14)-4/14 = HP Note: the -4/14 seems to be required to conpensate for rounding.

Current HP Formula tested at lvl 11 and does NOT hold true for levels 1-10. Further Testing required.

Rcollins779x 00:57, 1 October 2007 (UTC)

Nonlinear and free != fixed
I did a simple test with a level 1 toon as well as my "main" (23 commando).

If we look at the automatic attribute increases for Body. Level 1->2 gave us +2 Body, which gave +9 Health. Level 2->3 gave us +2 Body, which gave +9 Health. This would indicate you get the same amount of Health for each increase in stats.

However, manually putting a point in Body at level 2 gave us +6 Health. This indicates that "fixed" points and free points count differently for the derived attributes. The case is even clearer for Mind/Power. 1->2 gave +3 power, as did 2->3. But putting a single point in Mind at either level gives at least +4 power.

The increases are also nonlinear. For my 23 Commando, 1 point of Spirit gave +2 regen and +4 Health (vs +5/+3 at level 3), 1 point of Mind gave +2 power (vs +5 or so at level 3), and 1 point of Body gave +9 Health (vs +6 at level 2).

This all would seem to indicate that fixed stat effects are much less than those of the free points. It would also seem to indicate that how much you get for each point increase differ depending on level (and class?).

This also opens up an interesting speculation; does it matter WHEN you apply the points? IE, if you have three rangers at level 19 (18 levelups):
 * Character A spent all free points in Body from 2->10, then all free points in Mind from 11->19.
 * Character B spent all free points in Mind from 2->10, then all free points in Body from 11->19.
 * Character C saved all points, then spread them equally into Body and Mind at level 19.

Would the three characters have equal derived stats?

I did some data mining the other night; the results are in a table below. Leveling two characters up, one applying the points each level and one saving them until 5, there was no discernable difference between the stats of the two characters. Parts of the table below conflict with information in the table above, this may be due to a patch, but for future comparisons these stats were gathered with a character that choose soldier at level 5. Using the data below I was able to come up with a preliminary “character builder” which appeared to provide correct values for levels 1-5, but failed when tested against my 11 specialist and my 24 ranger. Going forward I’m going to work under the assumption that not only does each tier have different (or transformed) formulas, but also that, potentially, so does each class.

Another thing requiring some consideration is caps. I leveled my 11 to 13 and noticed that from 11 to 12 and 12 to 13 there was no boost to his power pool from the automatic stat increases. Putting points in mind did provide an increase in power despite no automatic boost.

--TaeKahn 15:40, 29 November 2007 (UTC)


 * Do note that just because an attribute didn't go up in the character window does NOT mean that it in fact did not go up. The Character window does not show fractions and many of the increases will be fractions of an actual number.  For a lot of numbers see Here -KillerNoodle 20:54, 29 November 2007 (UTC)


 * It is unlikely that class affects attributes. A test I did in beta showed a level 18 biotech and level 18 commando had the same derived attributes from the same base attributes. Another test showed that raising a stat from 44 to 75 on a level 18 biotech gave almost completely linear gains, but these gains did not match biotechs with lower level. This suggests that level is the mystery factor. - Dashiva 17:22, 29 November 2007 (UTC)


 * Do you still have the data from these tests? I'd like to integrate it with I already have.

--TaeKahn 19:01, 29 November 2007 (UTC)


 * Yeah, but it's not in an accessible format. Anyhows, what we need is basic concepts. Here are a few:


 * A character who doesn't spend any attribute points has power 100+level*3 and regen 100+level*2.
 * Health is less cooperative, but it grows fast (sharper curve than n^2).
 * For -added- points, a primary point is exactly twice a secondary point.
 * Added points are completely different from automatic points:
 * Level 23 character: (672, 169, 146)
 * Level 26 character: (872, 178, 152)
 * Level 23 character with +6 in each: (747, 187, 162)
 * Level 29 character: (1131, 187, 158)
 * Level 23 character with +12 in each: (822, 206, 178)
 * Per before, automatic points give 3 power and 2 regen for every level (+2 in each stat)
 * Manual points here gave ~6.2 power and ~5.4 regen for every +2 to each stat
 * Health, which leaps on auto (+200, then +250) is quite linear on manual (+75, then +75)


 * There a few more things that need checking. Most importantly: Are added points completely static, or based on the level you added them, or completely dynamic? But regardless of that, I think it's quite safe to say that we have a clear separation between the two points types. In fact, I posit that the automatic points might as well be pure eye candy, with base values for health,power,regen being based solely on level. - Dashiva 20:43, 30 November 2007 (UTC)


 * I've based my equations off of points instead of level. The result is the same either way, 1.5 and 1 per point, but doing it based off of the points seems more adaptable to changes to me. This changes the constants from +100 to +88 and +92. Like you said the bases provided by automatic increases are pretty much irrelevant, only having a single equation makes things easier then a table of values. As far as the base HP goes…I do have an equation that reliably predicts the value, but I’m sure it is not the correct equation. I plan on using this to produce data pairs to conduct tests for figuring out the correct equation. I certainly won’t be able to get to it this weekend as I have two finals and a couple of other assignments that need finishing. The, ugly, equation is 64.591e^.0434x where x is the automatic points in body. If someone is skilled in latex, feel free to do that up nicely. With this, you can figure out the health from 1 to 50. While I don’t want to believe this is the correct formula, it must certainly be exponential.--TaeKahn 21:42, 30 November 2007 (UTC)


 * A one shot equation for stats will probably look like this. (Base) + (Added) + (Modifications) = Total
 * Going more indepth would start to look like (Base) + (Added Major Stat + Added Minor Stat) + (Modifications) = Total
 * With base stats being (2 * level) + 8
 * The added stats are TotalStats - BaseStats = AddedStats
 * Then the hardest part the added stats:
 * They will look something like MOD * (Added Major * LevelMod) and then most likely 1/2 MOD * (Added Minor * LevelMod)
 * The hard part really is finding out the LevelMod that is used since the totals go down from 10-15 and then at higher levels will increase (the teir levels are the ones to look out for 5, 15, 30) KillerNoodle 21:37, 30 November 2007 (UTC)

Give this a shot. BaseAttributes = (2*Level) + 8

BaseHP = A1*(B1)^{BaseAttributes}

AddedBody  = [A2*(Level)^3 + B2*(Level)^2 + C2*(Level) + D2] * [TotalBody - BaseAttributes]

AddedSpirit = [A2*(Level)^3 + B2*(Level)^2 + C2*(Level) + D2] * 0.5 * [TotalSpirit - BaseAttributes]

HP = BaseHP + AddedBody + AddedSpirit

Where

A1 = 64.634047

B1 = 1.0443241

A2 = 0

B2 = 0.0156

C2 = -0.2703

D2 = 6.4003

This will give you a very close value to the HP you have. The reason I separated the numbers is because they give only an estimate. Though a very close estimate it still needs some work....and I can only do that with more numbers. So as I get more I'll update it with more exact figures if others see this as an accurate function. HAVE FUN!....The main places where this equation breaks down is when there are a great deal of added points. I'll continue to work on getting data points. The main thing that is needed more information from 15 - 30. If you are able to provide data with a single addition to body this will help the most. This will tune the equation so that it fits the data more. Thanks too.

- KillerNoodle 19:08, 3 December 2007 (UTC)


 * I must say that I am deeply distrustful of any formula involving third-grade polynomials, but that doesn't mean we shouldn't keep trying. I've put the most useful attribute data I have at User:Dashiva/Attributes/Data. - Dashiva 21:52, 3 December 2007 (UTC)


 * Following up on that comment....I removed the A and found that it is an even better equation with just the 2 polynomials....Updated the numbers. Thanks. (Max Error 0.8% at level 3) - KillerNoodle 22:21, 3 December 2007 (UTC)

Power
Known: Base power: (100 + Level * 3)

Unknown: Added power by attribute points(It is affected by level) [ Decreases as the level goes up ]

Unknown: Added power by modifications

Regeneration
Known: Base Regen: (100 + Level * 2)

Unknown: Added Regen by attribute points(It is affected by level) [ Decreases as the level goes up ]

Unknown: Added Regen by modifications

Perfect Attributes?
Hey, I'm just new to TR and I'd like to know if there are any perfect attributes for a Spy and/or Sniper... But I couldn't find any guides, and lurking around a bit got me here. Could anyone tell me those attributes/direct me to a guide? If there are 'perfect attributes' at all, that is. Thanks in advance. Also, keep up the good work with the wiki! :] -77.248.147.210 15:23, 6 December 2007 (UTC)


 * Attributes are being changed big time in v1.4, according to test server notes. So until then, the perfect attributes are no points spent at all. - Dashiva 17:08, 6 December 2007 (UTC)
 * Okay, thanks! I hate wasted attribute points :p -77.248.147.210 17:19, 6 December 2007 (UTC)


 * Man All that leveling with saving points and stuff...tossed out the window. Eh time to ask the server what their stats are.  -KillerNoodle 21:13, 6 December 2007 (UTC)
 * Where on the test server patch notes do they say this?...all I see is them reseting everyone's so that they can fix the rage situation (More points for body since the Soldiers wont be doing as much damage) and the Skill Points (Those who feel that the change to rage makes it not work the points. -KillerNoodle 21:29, 6 December 2007 (UTC)


 * Not actually in the patch notes, but in Paul Sage's commentary about the changes. The interesting statements are the following:


 * Attribute changes are going in a future patch (Ver 1.4 which is NOT on test right now) which will make Health more meaningful, and this will greatly aid the medics. (...) There will be a reduction in Bio-Augmentation efficacy. However, the attributes will mean much more in the 1.4 update, making this change a necessity. (...) Disease will debuff the attributes of the attacker, seriously lowering them. This change will be complimented by the 1.4 update changes to attributes (...)


 * Like so. - Dashiva 22:42, 6 December 2007 (UTC)


 * Thanks. It is good to see the evolution of classes (Read as The Game). Seeing as how armor used to have the Bleed Through thus making Healing needed....they took that out and now they are looking at fixing the Class(es) to reflect this.  I honestly can't wait but hate to see all the work done on the skill equations go bye bye.  Taking a look at the numbers for power and regen they might be trying to fix the fact that the added stats go down as you level up.  Though this might have been planned it does seem weird that +1 at level 2 gives more Power or Regen then at level 10. But we can only wait and see...... -KillerNoodle 03:22, 7 December 2007 (UTC)

Formulas
It's been a while since I dabbled in numbers, but as far as I recall the primary attribute effects are relative to the base value, with the proportion being actual attribute value / default attribute value. We should be very clear that they are not static increases. - Dashiva (talk) 21:17, 20 October 2008 (UTC)