Forums → Games → Gemcraft CS stronger Gems with Pool&Blood

[IMG]http://i.imgur.com/zxq5Gbi.png[/IMG]

Uhmn i kinda don´t know how to put images in my post.... sry guys...

Instead of [IMG] use [img]. and make sure it links to actual image not page with it.

thx=D

Not to discount your work, but...

My understanding of the base gems is that the modifiers for Bloodbound and for Poolbound are not calculated together. Instead, you take the base, multiply them out separately, and then add them together to get the final result. That tidbit was why no one ever seriously looked at combining both bloodbound and poolbound into one gem.

At higher grades, Bloodbound significantly outpaces Poolbound in terms of effectiveness, as well. I forget the exact mix on the base gem that I used, but I compared two mana gems once by making them almost exactly alike, with the only difference being the use of Poolbound instead of Bloodbound for one of the gems (the Mana Leech and Chain Hit put into those starting gems were the same for both gems). I then used psorek's 16 gem upgrade plan and found that at grade 28, the Poolbound gem was 12% more mana leech than the Bloodbound gem. At grade 34, the Bloodbound gem was 42% more mana leech than the Poolbound gem. Unfortunately, I did not stop at Grade 31 to compare the two. That gem upgrade plan gives a cost increase of 4 grades, but a displayed increase of 3.

At lower grades, the generally accepted wisdom was that Bloodbound would diminish the impact of the Poolbound in your gem. At higher grades, the generally accepted wisdom was that the Poolbound in the gem diminishes the impact of the Bloodbound.

Your own images illustrate this : look not at the damage, but at the mana leech and chain hit of the gems you created. I'm surprised that the trap with amp gems are the way they are though. Can you run another such comparison, showing the tri-color using Bloodbound, to show side-by-side with your Poolbound and your quadcolor?

As for the Armor Tearing and Suppression gems being given very little notice, well, there is a reason for that. The way that Armor works in this game is to reduce the total damage done on any hit by the amount of Armor the monster has. A wave of monsters that has 240 HP and 101 Armor, for example, needs to be hit for at least 103 damage to do more than 1 actual damage to the monster. A hit of 160 would only do 59 damage, and would thus take several hits to kill that monster. The damage gem most commonly used at the high end is the Critical Hit gem, and it is used for a reason. If the gem being brought to bear on that monster is a grade 2 Yellow gem that does 101-274 damage, with 80% chance to deal 5.6 times the damage, that gem has an 80% chance to one-shot those 240 HP, 101 Armor monsters. This is because that critical multiplication is done before the armor is reduced from the hit. At 101-274 damage, if it crits the range is effectively 565-1,534, being used against a monster that has a combined armor+HP of only 341. The sheer power of the critical hit gem, especially with the way it gets boosted by Bloodbound at higher grades, makes monster Armor a moot point. They do not have enough Armor after a certain point for their armor to matter any. And when you are killing the monster in about one second each, they cannot regenerate enough HP to matter, either, rendering critical hit vastly superior to Suppression, not just Armor Tearing.

All that aside, though, Armor Tearing has another weakness : it can only remove as much Armor as the monsters have. If the monster has 101 Armor, it can pretty much laugh at a gem that does 45-76 damage. Once that gem has reduced the Armor down a bit, the gem starts to do more than 1 damage per hit, but it can not do more than the 76 damage it is rated for, no matter how low the monster's armor gets.

BTW - if you plan on running such a comparison again, remove all Bloodbound and Poolbound from your Amplifiers. They do not work in Amplifiers. Your better bet is to use gems of your other colors; if you are looking at a gem with Red and Orange, then use Orange, or a dual Orange/Red gem for the amps.

You wrote everything I didn't have lust to write, thanks

Dear Astroshak,

Im glad that someone notice my post and take me serious.

Yes thats right this is the stat calculation and thats what i try to mention before.

I dont wanted to but if u insist i will show u the exact calculation:

For just the Gem:
Basevalue*Poolbound+Basevalue*Bloodbound=final result

*Note:
x=Basevalue
Poolbound=pb
Bloodbound=bb
x*pb+x*bb=x*(pb+bb)=resultingstat

if u want to get the basevalue of the gem just divide
Resultingstat/(pb+bb)=x
*
Firingrate:not effected by Boundbonus

For Gems in traps without Amps:
Dmg: -80% -->0,2 multiplier
nonbound specials --> +310% ---> 3,1 multiplier
Enhancement Barrage --> +30% -->1,3 multiplier
Firing speed: 165%-->1,65 multiplier *Note: There is no addition here*

Final stats:
Dmg=x* 0,2*(pb+bb)
Specials=(x*100%+x*310%+x*130%)*(pb+bb)
* Note: ->x(100%+310%+130%)=x*540%=x*5,4 *

Firingrate= x*1,65

For Gems in traps with Amps:
Dmg: -80% -->0,2 multiplier
nonbound specials --> +310% ---> 3,1 multiplier
Enhancement Barrage --> +30% -->1,3 multiplier
Dmgbonus from Amps=Dmgb
Specialbonus from Amps=Specb
Firing Speed: 165% -->1,65 multiplier
Firingspeedbonus from Amps=Spdb

Final stats:
Dmg=(x* 0,2+Dmgb)*(pb+bb)
Chainhit=x*5,4*(pb+bb) *Note: no Amplifierbonus here*
ManaLeech=(x+Specb)*5,4*(pb+bb)
Firing Speed=(x+spdb)*1,65

Awww.... i don´t mentioned the triple and Quadgem modifier.... well u can do it by urself =D this multiplier is just relevant if u compare triple vs quad but not with triple vs triple or quad vs quad.

U can check the maths by urself if u want to. I checked it with some excel tables:link
*Note: the Results are not exactly identical due the game does not display all stats and is constantly rounding values up and down between calculations but its still in the correct range of values*

So the reason why the quad gem is superior with Amps is that the addition of the bonus stats of the Amps is done before any multiplication, as u can see from the firing rate stats.

The reson why most people dont use quad gems is because adding more color decreases the special abilites. As most gamers want the highest special abilities of a specific color it is better to use as less colors as possible.
Thats why the poolbound multiplier of tricolor is 92,27 and of quad is 89,06. This difference probably grows with higher grades.

The Special ability loss of Tricolor to Quadcolor is 25%. But this can be easily surpass by Bloodbounds own multiplier which is added to the final result.

So probably the Quadcolors grows of special abilities and Dmg is much higher then Tricolors with the usage of Amps.



Mhmm yeah i thought that would be true but i couldnt confirmed that the bloodbound is lategame essentially better. Well i have to accept that.

If u insist i will start a bloodbound run in comparison on the same bloodbound lvl tomorrow.

Yeah its true that they are absolutly not effective to kill but they should just support the killgem and not kill the monsters =D and btw the most important thing is a support tower based on slow and chain to support the killgem. A slowing tower is far more effective then a slowing trap because of the greater range as it hits the monsters more often before their slowduration is over and btw it should shred the healing and armor capacities. To destroy the armor or healing is not primary target.
*Note: Target Priority of any support Tower is always on random =D*

Yeah the Bloodbound in the amplifiers is just something i want to try out because i read in some other forum that i still has any effect on the amped gems.

My point about the dilution was this : if you have a given gem, say 1 red, 7 orange, 4 black, 4 white, and U that to upgrade it, you will always maintain the same proportion of black and white. At first, White is a lot better than Black, so the gem is weaker than a 1 red, 7 orange, 8 white would be.

Later on, after around Grade 30, the Black portion's strength grows to be stronger than the White's portion, resulting in a flip : the 1 red 7 orange 8 black will be stronger than the 1 red 7 orange 4 white 4 black.

The following Grade 5 gems were made with one G1 red, seven G1 orange, with the remaining eight G1's being all black, all white, or four of each, all combined in the simplest manner possible (shift-G + click until all are combined) :

BOR
531-1498 damage, 7.6 range, 1.14 shots/second
4.71 chain, 5.19 mana leech, x1.19 bloodbound

WOR
3107-8284 damage, 7.4 range, 1.21 shots/second
15.24 chain, 30.16 mana leech, x6.89 poolbound

WBOR
2807-7611 damage, 7.6 range, 1.19 shots/second
14.23 chain, 21.65 mana leech, x1.11 bloodbound, x6.07 poolbound

That's with a hit level of 1, and a MP level of 13. Obviously, as the hits accrue, the Black portion will level up (and provide better specials) while the white portion will remain constant until the mana pool levels up. However, note that the quad gem's damage and specials are not as good as the white triple.

If I put those gems into a tower and surround it with eight G5 orange gems :
BOR
1548-3691 damage, 27.9 range, 3.99 shots/sec
4.71 chain, 34.59 mana leech, x1.19 bloodbound

WOR
9036-20972 damage, 27.2 range, 4.06 shots/sec
15.24 chain, 200.91 mana leech, x6.89 poolbound

WBOR
8088-19074 damage, 27.8 range, 4.04 shots/sec
14.23 chain, 174.86 mana leech, x1.11 bloodbound, x6.07 poolbound

Replace those 8 amp gems with dual WB gems increases the damage done but the specials are not boosted at all :
BOR
1704-5327 damage, 28.3 range, 4.5 shots/sec
4.71 chain, 5.19 mana leech, x1.19 bloodbound

WOR
9938-30473 damage, 28.1 range, 4.57 shots/sec
15.24 chain, 30.16 mana leech, x6.89 poolbound

WBOR
8897-27599 damage, 28.2 range, 4.55 shots/sec
14.23 chain, 21.65 mana leech, x1.11 bloodbound, x6.07 poolbound

Since, for manafarming, you want to emphasize the mana leech, it is better to put the pure orange gems in the trap. It would be interesting to see whether yellow, or yellow/black, would be better for the killtrap's amps.

As for the slow gem in a tower idea .. you can get as good results, without having a tower shooting things willy-nilly, merely by placing a trap right in front of your manafarm trap. Into that trap place a gem made by :
G2 Black + G1 Red
result + G1 Blue
result + G3 Orange
That resulting gem gives you a little bit of mana back (a little free mana never hurt anyone, right? - though it won't amount to much in the long run) and with patience, or a few U upgrades, hits the 90% slow cap easily. By placing it right before your mana trap, you dramatically lengthen the time that any given monster stays on the trap, allowing for more mana to be stolen from it. By placing one in front of your killtrap (recommended if not doing the permafreeze manafarming) and upgrading it to the 90% slow, you can basically take your target monster HP and increase it by a factor of ten. If your wave bombing HP target was 140,000 HP, bomb to 1,400,000. If your target was 3.4e27, then bomb until you are a little below 3.4e28. And so on.

The trap also has the benefit of not slowing a monster that is not about to reach the manafarm anyway. You want mosnters to get to the farm quickly, and to linger there. You don't really want monsters to dawdle on their way to your manafarm, after all. For that matter, if you are not doing the permafreeze, you want them to run to your killtrap, and not to dawdle on their way there, too.

daniil sizov has explained everything thoroughly.
reddit

Say, you had B/W/O/R with x1000 bloodbound bonus and x100 poolbound, so you have x1100 bound bonus. Without that poolbound component you would have (1000 * (0.5 / 0.4) / 0.79 ) * (0.5 / 0.4) / 0.89 = 2222x (!!) mana leech per second instead of x1100!

And that does not even take into account that the bloodbound bonus would be even higher if white would be left out (blacks percentage would be higher then).

You should have tested the tricolor with Black instead of White before posting this.

Thesis is still confirmed for now:





My game crashed before i could make the tricolor poolbound which is probably much weaker because i didn´t got to the pool lvl 39 for now. This Gems are in a Tower and have 8 Amps with a pure 29 orange gems.

This time i also just fooled around and not enraged any waves. Gonna make a serious run and get the gems to 40 to get a real comparison.

It is not reasonable to think that in a quadcolor Gem with blood and pool, the multiplier of blood is 1000 while the poolmultiplier is just 100.
Because as the bloodmultiplier raises the manaleech special is increased to. With the Manaleech is increased the pool lvl and poolbound multiplier increases too.

The reason for the diffrences between quad and Tricolor is that the ampbonus is added before the multiplication. And even if the Multipliers of the bounds are added before multiplication and the penalty loss due the quad modifier, no guide has shown the results with the amp bonus.

But im gonna make a serious run to get a real comparison 4 u guys.

It is clear that your method has an advantage - as there is an accepted advantage of poolbound over bloodbound for lower grades. By doing what you do, you just push the point where tricolor with black is superior a little further.

My estimation would be that black gets better for the latest somewhere in the mid 40ies. If one plays a game where gems with higher G's are neither wanted nor needed - you're right - quadcolor with black and white is the best choice.

This actually makes perfect sense mathematically. I will run through it quickly below.

The author is posting that combined black/white is better than pure white or black in mixed gems. The antithesis is that people are saying pure black is better than pure white at higher grades.

So the claim from the super-high-level skeptics is that when you combine white and black in one gem the end addition to the gem is not (White * Black, * rest of gems) but (White * rest of gems) + (Black * rest of gems). Using the numerous examples posted above, it is clear how combined white and black would be superior.

Using the latest image, pure black gave x72.93. In the combination, white gave x56.21 and black gave 42.18. Mathematically, it is clear that 56.21+42.18 > 72.93. The same follows for all of the previous examples. So unless progression is not consistently linear, geometric, or exponential, this will always hole true.

So again, following how everyone has already defined how the combinations work, it is very clear why in the cases given white and black combined are better than either alone. However, as progression may not be consistent, this could be different at various level ranges.

All this begs the question : how should those 32 gem starting recipes of psorek's be updated to include white?

If, of course, quad winds up better at all reasonably achievable levels. I would not expect anyone to check for grade 100 gems, but grade 70 would be a good indicator for any reasonable-length manafarm.

According to my theoretical rough calculations its about like this:

G35 > Tricolor (with pure B) would need about Hitlevel 30 (60.000.000)
G40 > Tricolor would need Hitlevel 25 (5.000.000)
G45 > Tricolor would need Hitlevel 19 (246.000)
... to reach break even with Quadcolour (if the Manapoolsize isn't much bigger than needed to build the corresponding gems; maxed skills where 'necessary'.

In practice it should be somewhere in between G40 and G45 that Tricolor surpasses Quad. After the advantage of Tri will increase very quickly.

1) Mine recipes are optimal end-game, let's say past grade 40.
2) Even if not, I don't know how I would generate 4-color formulas, it's way more complex mathematical problem than 3 color ones.