|
TossFloss
Canada606 Posts
2010-05-08 Updated for patch 11. Added carrier, void ray(uncharged), void ray(charged). Now properly accounts for air damage vs colossus.
% Advantage Analysis
+ Show Spoiler +
Number of Hits
+ Show Spoiler +
How-To
The charts show the unit vs unit upgrade advantages for each matchup. The advantages were computed by calculating the number of hits needed to kill a unit, NOT the dps.
The data show the % advantage of upgrading from the previous upgrade level.
The following example illustrates how to interprate the % advantage table data.
+ Show Spoiler +
ATTACK ANALYSIS lv1 lv2 lv3
Roach vs (-/+0)Zealot 11.11% 12.50% 0.00%
Roach vs (-/+1)Zealot 10.00% 11.11% 12.50%
Roach vs (-/+2)Zealot 20.00% 11.11% 12.50%
Roach vs (-/+3)Zealot 18.18% 10.00% 11.11%
Explanation:
Against zealots with no armor upgrades (-/+0), +1 roaches kill 11.11% faster than +0 roaches. +2 roaches kill 12.5% faster than +1 roaches. +3 roaches kill 0% faster than +2 roaches.
Against zealots with +1 armor upgrades (-/+1), +1 roaches kill 10% faster than +0 roaches. +2 roaches kill 11% faster than +1 roaches. +3 roaches kill 12% faster than +2 roaches.
And so on for (-/+2) and (-/+3) zealots.
ARMOR ANALYSIS lv1 lv2 lv3
+0 Zealot vs Roach: 18.18% 15.38% 26.67%
+1 Zealot vs Roach: 10.00% 18.18% 15.38%
+2 Zealot vs Roach: 11.11% 10.00% 18.18%
+3 Zealot vs Roach: 12.50% 11.11% 10.00%
Explanation:
Against zealots with +0 attack upgrade. Roaches with +1 armor upgrade tank 18.18% tougher than roaches with no armor upgrade. Roaches with +2 armor tank 15.38% tougher than roaches with just the +1 armor upgrade. Roaches with +3 armor tank 26.67% tougher than with +2 armor.
Against zealots with +1 attack upgrade. Roaches with +1 armor are 10.00% tougher than roaches with +0 armor. And so on...
Assumptions
The data was computed with the following considerations:
- zerg units seem to regenerate 1hp almost instantly when taking damage from full health
- Units take at least 1 damage from every attack
Misc.
Click here to view the unit stats database used to generate the charts
Click here to download the python script used to generate the charts.
|
would be good to see how many hits u need as a number instead of %
|
TossFloss
Canada606 Posts
On March 22 2010 05:05 MorroW wrote:
would be good to see how many hits u need as a number instead of %
Sure I can put that info up as well. Will be up in a bit.
|
awesome
btw i was making a program in c++ exactly like this, but i suppose that will be useless now. will still keep working on it since i mostly do it for fun ^^
|
Yeah defintiely the number of hits would be easier. This % system can be confusing.
|
Incredible amount of useful information here, thank you very much for this!
|
the chart is really confusing, I don't understand your explanation either.
so something like this:
Roach vs (-/+0)Zealot 11.11% 12.50% 0.00%
is broken down into this:
Roach vs (no R armor upg/no Z attack upg) Zealot. level 1 Z attack= 11% better, lvl 2 Z attack= 12% better, lvl 3 Z attack = worthless.
Or is it levels for armor upgrades? Or do I have them backwards all together (Z armor vs R attack)?
need # of hits breakdown, yea. It's a lot easier to understand and more applicable to a game than unit efficiency.
|
TossFloss
Canada606 Posts
On March 22 2010 05:12 CharlieMurphy wrote:
the chart is really confusing, I don't understand your explanation either.
so something like this:
Roach vs (-/+0)Zealot 11.11% 12.50% 0.00%
is broken down into this:
Roach vs (no R armor upg/no Z attack upg) Zealot. level 1 Z attack= 11% better, lvl 2 Z attack= 12% better, lvl 3 Z attack = worthless.
Or is it levels for armor upgrades? Or do I have them backwards all together (Z armor vs R attack)?
need # of hits breakdown, yea. It's a lot easier to understand and more applicable to a game than unit efficiency.
Hi. Sorry for the poor explanation.
For the attack analysis table:
- Attacker's attack upgrade increases moving across rows from left to right.
- Defender's armor upgrade increases moving down columns from top to bottom.
For the armor analysis table:
- Defender's armor upgrade increases moving across rows from left to right.
- Attacker's attack upgrade increases moving down columns from top to bottom.
Will put up # of hits breakdown soon.
|
Interesting, hits would also be appreciated!
|
TossFloss
Canada606 Posts
Added # number of hits charts for each matchup.
|
Awesome, this data will affect my play style for sure.
Battlecuiser stats are missing tho, i would love to see the armor/damage upgrade difference.
|
It shouldn't be
+ Show Spoiler +ARMOR ANALYSIS lv1 lv2 lv3
+0 Zealot vs Roach: 18.18% 26.67% 31.58%
+1 Zealot vs Roach: 15.38% 18.18% 26.67%
+2 Zealot vs Roach: 11.11% 15.38% 18.18%
+3 Zealot vs Roach: 10.00% 11.11% 15.38%
instead of
+ Show Spoiler +ARMOR ANALYSIS lv1 lv2 lv3
+0 Zealot vs Roach: 15.38% 26.67% 31.58%
+1 Zealot vs Roach: 18.18% 15.38% 26.67%
+2 Zealot vs Roach: 10.00% 18.18% 15.38%
+3 Zealot vs Roach: 11.11% 10.00% 18.18%
?
|
TossFloss
Canada606 Posts
On March 22 2010 06:53 Golondrin wrote:It shouldn't be + Show Spoiler +ARMOR ANALYSIS lv1 lv2 lv3
+0 Zealot vs Roach: 18.18% 26.67% 31.58%
+1 Zealot vs Roach: 15.38% 18.18% 26.67%
+2 Zealot vs Roach: 11.11% 15.38% 18.18%
+3 Zealot vs Roach: 10.00% 11.11% 15.38% instead of + Show Spoiler +ARMOR ANALYSIS lv1 lv2 lv3
+0 Zealot vs Roach: 15.38% 26.67% 31.58%
+1 Zealot vs Roach: 18.18% 15.38% 26.67%
+2 Zealot vs Roach: 10.00% 18.18% 15.38%
+3 Zealot vs Roach: 11.11% 10.00% 18.18% ?
The original is correct.
To understand why, have a look at the equivalent raw # of hits chart.
################ lvl0 lvl1 lvl2 lvl3
+0 Zealot vs Roach: 13 15 19 25
+1 Zealot vs Roach: 11 13 15 19
+2 Zealot vs Roach: 10 11 13 15
+3 Zealot vs Roach: 9 10 11 13
In the first row a +0 zealot kills a -/+0 roach in 13 hits. Getting the +1 armor upgrade increase that to 15 hits. So it takes 15/13 = 15.38% more hits to kill the roach.
Looking at the second row: a +1 zealot kills a -/+0 roach in 11 hits. Getting the +1 armor upgrade increases that to 13 hits. So now your +1 roaches are 13/11 = 18.18% more effective than without the +1 armor upgrade.
|
|
|
interesting, blings are one of the only (maybe the only) unit to end up doing more damage when both sides are fulled upped (most just equal out).
spoke too soon, roaches & ultras too!
edit- is this because there is no shield armor taken into consideration?
edit- I found a problem, it says that corruptors are n/a vs collosus, but collosus can be attacked by both ground and air units. Dang, this was one of the important ones too imo.
Air Attack: 12 (+1)
+10 (+1) Damage against Massive.
Armor: 2 (+1)
200/150
Armor: 1 (+1)
double damn, this thing http://sc2calc.net23.net/ doesn't have all the zerg units (missing corruptor/broodlords/broodlings/banelings and maybe some other?)
quick math on corruptors vs no up collosus
upgrades: 0 1 2 3
hits: 16 15 14 13
|
I think there might be something wrong with the marine vs mutalisk in ZvT hit chart too.
and marine vs broodlords.
yea and then the marine(shield) is all N/A too vs mutas and imbalords.
interesting that mutalisk are better AA over Corruptors vs banshee and viking.
|
Is there any way to account for splash in these?
IE, a Siege Mode tank kills a Marine in 1 hit regardless of anyone's upgrades, but maybe tank attack reduces hits for nearby Marines from 3 to 2, or something. Do we even have a way of knowing what the Tank's damage is at what range from primary target?
|
On March 22 2010 08:15 pure.Wasted wrote:
Is there any way to account for splash in these?
IE, a Siege Mode tank kills a Marine in 1 hit regardless of anyone's upgrades, but maybe tank attack reduces hits for nearby Marines from 3 to 2, or something. Do we even have a way of knowing what the Tank's damage is at what range from primary target?
That kinda stuff requires micro/positioning and can't be accurately tested, and isn't really applicable data because it's very specific/situational. I mean you could even take it another step and say well in most ZvT they have marine/marauder and a tank doing splash. Each marine is 2 less hits and marauder is 1 less or whatever. It gets too complicated and messy trying to figure it all out.
Imagine trying to figure out broodlords and broodlings damage output considering normal factors.
For these things, you just need to playtest and get a feel for it.
BoT, Wow upgrades for roaches vs hydras are very important. They both kill each other around 8~9 seconds and an attack upgrade from the hydra beats a roach, but an armor or attack beats a hydra again.
|
Useful, a thing like that should be added to Liquipedia II.
|
TossFloss
Canada606 Posts
On March 22 2010 08:08 CharlieMurphy wrote:
I think there might be something wrong with the marine vs mutalisk in ZvT hit chart too.
and marine vs broodlords.
yea and then the marine(shield) is all N/A too vs mutas and imbalords.
interesting that mutalisk are better AA over Corruptors vs banshee and viking.
Thanks a lot.
There was a mistake in the unit stats database. The marine's was doing x6 attacks to air units o_O. Fixing the charts now.
|
|
|
I did something similar to these charts except you add the numbers yourself.
http://spreadsheets.google.com/ccc?key=0AhDmnMwVh3wMdEFLYzcwaXpBRGN2cTNleUpCRFJNQ0E&hl=en
Here's a preview of it.
You just copy paste the 2 gray lines for Attacker/Defender from the compilation at the bottom and you get all the attack/defense info calculated underneath.
You can also modify the shield defense stats and make your own in the gray box on the right. I didn't add shield data because that would be 4x more information to cram into an already pretty convoluted spreadsheet.
I'll add dps if people think it's useful. I made this more for myself to see what's useful for microing than anything else but sharing is caring.
edit: oops forgot to share, should work now
edit2: I'm looking at the phoenix bonus +5... Does anyone know if that +5 is also multiplied by number of attacks?
|
Thanks a lot, I am playing with a heavy zealot push and this analysis is gold as the upgrades make all the difference when there are large # of troops.
|
So, correct me if I'm wrong. Upgrading armor for Protoss give HUGE advantage compared to upgrading their Attack. As for Terrans upgrading Attack is mroe effective than upgrading Armor (for bio at least). Which means it is more effective for Protoss player to focus no Armor and more effective for a Terran player to focus on Attack (for bio).
Also
For those who like Thor upgrading either Armor or Attack doesn't give any good benefit.
Nice chart!
|
Thanks a bunch, this is so helpful!
|
|
|
Excellent was looking for something like this to min max as much as possible.
nvm should have scrolled down further
|
Thanks for taking the time to do all this man, very cool.
Only thing i see wrong so far is Hellion vs. Banelings. Banelings are just bio, not light armor, so pre-igniter shouldn't have any effect on them.
|
Thanks for the post  this helps clear up a lot of the "issues" of when to upgrade I was having.
|
Hate to necro a 4 day old thread, but I've found this one to be very useful, any chance you'll continue to update the tables, especially the hits to kill for patch 6 and perhaps beyond? If not I'm sure someone (if not myself) can continue your work.
|
Hate to necro a 4 day old thread, but I've found this one to be very useful, any chance you'll continue to update the tables, especially the hits to kill for patch 6 and perhaps beyond? If not I'm sure someone (if not myself) can continue your work.
+1,that's what we are expecting as well
|
TossFloss
Canada606 Posts
2010-03-27 EDIT
- Charts updated for patch 6
- Added Battlecruiser
- Fixed Marine vs Muta charts
- Published chart builder program
- Published unit stats database
|
TossFloss
Canada606 Posts
On March 22 2010 23:32 Bosko wrote:
Thanks for taking the time to do all this man, very cool.
Only thing i see wrong so far is Hellion vs. Banelings. Banelings are just bio, not light armor, so pre-igniter shouldn't have any effect on them.
Can anyone else verify this?
According to the SC2 armory, Banelings are both bio and light armor.
|
can you distinguish lings and cracklings in the same manner as hellions w/ and w/o igniter?
EDIT: sc2armory is incorrect, banelings are just bio
|
have stalkers been updates on this?
|
TossFloss
Canada606 Posts
On March 28 2010 05:51 ccou wrote:
can you distinguish lings and cracklings in the same manner as hellions w/ and w/o igniter?
EDIT: sc2armory is incorrect, banelings are just bio
Because the adrenal gland upgrade increases attack speed, not damage, the upgrade stats between cracklings and zerglings are the same
Thank you for the info regarding banelings. I am currently away from my home. I will update and post new charts when I return tomorrow.
On March 28 2010 06:07 FeeL_ThE_RusH wrote:
have stalkers been updates on this?
Yes.
|
Few things from the Protoss perspective...
PvP
Immortals are heavily favored against stalkers regardless of upgrades
Collosus are very good against zealots and stalkers regardless of upgrades, typically only taking 4 to 5 hits to kill
Zealots, with upgraded attack reduce the number of hits to kill versus Stalkers/Zealots by 1 per upgrade
Stalkers receive the most benefit from +1 attack upgrade versus zealots, reducing hits to kill from 17 to 15
PvZ
Zealots with +1 attack take 2 hits to kill Zerglings as opposed to 3 (Like SC1). Attack upgrades for zealots not very noticeable versus hydras/roaches.
Stalkers receive little benefit from attack upgrades versus hydras/roaches/zerglings.
Colossus do mostly(barring +3 armor situations) 1 hit kills versus zergling/baneling regardless of upgrades.
PvT
Attack upgrades for Zealots/Stalkers/Immortals do not change hits to kill(very much) versus marines/reapers and negligible difference is seen versus marauders.
|
TossFloss
Canada606 Posts
|
You're awesome, TossFloss.
|
can someone clarify something for me?
so the queen has 4 damage for 2 attacks and the zealot has 8 damage for 2 attacks and many other units have X damage for X attacks
so if a queen or zealot attacked a roach who has 2 armor, would the queen and zealot attack be reduced by two both times? so for the queen (4-2)*2=4 damage and for the zealot (8-2)*2=12 damage
or will the queen do (4*2)-2=6 damage and the zealot (8*2)-2=14 damage
|
On April 01 2010 01:11 FiercE wrote:
can someone clarify something for me?
so the queen has 4 damage for 2 attacks and the zealot has 8 damage for 2 attacks and many other units have X damage for X attacks
so if a queen or zealot attacked a roach who has 2 armor, would the queen and zealot attack be reduced by two both times? so for the queen (4-2)*2=4 damage and for the zealot (8-2)*2=12 damage
or will the queen do (4*2)-2=6 damage and the zealot (8*2)-2=14 damage
Armor is applied to each attack. The 2 armor Roach would reduce each of the Zealot 8 attack to 6 attack, so it would be 6*2 instead of 8*2.
|
holy shit it looks like getting +1 armor first vs terran is soooo much better than +1 attack (as Protoss).
|
How could I use the python script to generate the charts? I google it and download python 2.6, but it seems not working?
I know the database is for EXCEL, but which software should I use on python script when I found python 2.6 is not working.
any guide for the whole generating process?
thx a lot...
|
TossFloss
Canada606 Posts
2010-04-15
* Update for Patch 8
* Added Marine (Stimmed)
* Added Marine (Shield + Stimmed)
* Added Marauder (Stimmed)
|
TossFloss
Canada606 Posts
On April 16 2010 11:05 TossFloss wrote:Show nested quote +On April 09 2010 04:42 PeterDoe wrote:
How could I use the python script to generate the charts? I google it and download python 2.6, but it seems not working?
I know the database is for EXCEL, but which software should I use on python script when I found python 2.6 is not working.
any guide for the whole generating process?
thx a lot... Hi PeterDoe The python script downloads all the unit stats from a MySQL database. You'll need to have MYSQL installed. Create a database with a table called 'units'. Edit the python script to include the login, password and database name for the database. You'll need to upload unit data. You can do so by opening this link And saving it as a CSV file. To upload the file into the database I strongly recommend you use PhpMyAdmin (this will require an Apache + PHP installation) After you have the database setup, run the python script and it should generate the charts.
|
Has anyone tried getting super fast +1 armor as toss and pushing out with zealot sentry? With guardian shield the zealots and sentries would take a whopping 2 damage per marine attack on their health and a crushing 5 damage from marauders.
|
Thank you soooooooo much Floss! This chart is an extremely helpful reference, I have been looking since day 1 to find a chart like this!
|
this is really great information..
im especially intrested in the immortal results
|
TossFloss
Canada606 Posts
On April 20 2010 09:04 Percutio wrote:
Has anyone tried getting super fast +1 armor as toss and pushing out with zealot sentry? With guardian shield the zealots and sentries would take a whopping 2 damage per marine attack on their health and a crushing 5 damage from marauders.
It's counter-intuitive, but the marginal advantage from each armor upgrade increases the more armor you have.
Would be great if someone could try this out.
The charts are great for coming up with ideas, but those ideas need to be proved or disproved in real games.
|
TossFloss
Canada606 Posts
2010-05-08 Updated for patch 11. Added carrier, void ray(uncharged), void ray(charged). Now properly accounts for air damage vs colossus (e.g. corruptor vs colossus)
|
Any chance we can get an update to this?
Specifically, the number of hits spreadsheet seems like the most useful. A 15% damage upgrade might seem useful to know, but unless we know how that translates as a marginal differences in a battle, its useless. The number of hits tells us exactly what that marginal difference is.
|
TossFloss
Canada606 Posts
Sure I'll produce an update which will be up by tomorrow
I'm not 100% happy with the format & display. Any suggestions here would be welcome.
|
I may be misunderstanding a simple dynamic here but this is what I assume to be the case:
In regards to protoss:
+1 armor does not reduce damage until the shields are gone and it's attacking health.
That being said, are your calculations accounting for the transition from shield to health?
Like, looking at the TvP spreadsheet; it has a marine requiring 60 additional hits to kill an archon once it has +3. I'm not sure if this +3 is armor or shields but from my (possibly flawed) understanding, I imagine that +3 armor would only tack on about 2 extra hits required to kill it.
Although I may be approaching both the inner workings of the game AND the spreadsheets wrong so forgive me if I am; I just woke up T_T
|
The formats nice. I'll probably play around with it to make it easier to compare and move things around tomorrow.
@Noggin good question about toss armor. I think that his spreadsheets use both types (ie., a +1 armor Zealot would have +1 for both types of armor...but thats just a guess
|
TossFloss
Canada606 Posts
On July 06 2010 04:05 taarna wrote:
The formats nice. I'll probably play around with it to make it easier to compare and move things around tomorrow.
@Noggin good question about toss armor. I think that his spreadsheets use both types (ie., a +1 armor Zealot would have +1 for both types of armor...but thats just a guess
Yeah this is correct.
I should probably only upgrade the armor and leave the shields at zero; or somehow make additional columns.
Sorry for not yet releasing the update. I'm pre-occupied with real-life responsibilities. But I will get this done as soon as I can.
|
On July 07 2010 02:00 TossFloss wrote:Show nested quote +On July 06 2010 04:05 taarna wrote:
The formats nice. I'll probably play around with it to make it easier to compare and move things around tomorrow.
@Noggin good question about toss armor. I think that his spreadsheets use both types (ie., a +1 armor Zealot would have +1 for both types of armor...but thats just a guess Yeah this is correct. I should probably only upgrade the armor and leave the shields at zero; or somehow make additional columns. Sorry for not yet releasing the update. I'm pre-occupied with real-life responsibilities. But I will get this done as soon as I can.
Thanks for this database ! I like it !
|
I have another idea for a separate sheet that would be very useful. Include attack rate and also include time to kill. I realize I'm asking a lot, but maybe you're interested :-).
|
On March 22 2010 05:05 TossFloss wrote:Show nested quote +On March 22 2010 05:05 MorroW wrote:
would be good to see how many hits u need as a number instead of % Sure I can put that info up as well. Will be up in a bit.
Just wondering if this was ever done, the how many hits to kill part that is. I don't see it.
|
This is so awesome. There should be more than 3 pages of comments (considering whiners make threads many times this long about whining about how weak their race is).
Gonna bump as well as say @sensenmann: yes, it's right under the % advantage and its bolded and says Number of Hits.
But thank you tossfloss. Was this ever in the spotlight? I hope it was.
|
This is some useful stuff if you take the time to look at it. It can be intimidating with the amount of data though.
|
Wow... Thanks a lot for your data really. I just realized that i underestimated the banelings a little too much... I still fail to see the usefulness of roaches since blings seems to perform way better unless there are a lot of zealots (because blings are useless after 1 hit obv). For everything else in zvp it seems that glings/blings are better...
|
Just wanted to say thank you for making this and keeping it updated. I was looking for exactly this, and know it will help me improve my decision-making in game.
edit: hmm, misread the "last edit for patch 11" as "1.1"
In that case... perhaps someone knows of a more updated version? This will do very nicely until then though.
|
|
|
Yeah, it would be great to update this.
It also shows that +range att for zerg is laughable, its absolutely thrash. Only zerglings really benefit from the upgrades, the others upgrades are quite small.
|
|
|
|
|
|
EPT
