In short, the aliens win a disproportionate number of games while they aren't any more, or less, skilled than humans are. Simply put, for killcounts the aliens and humans are perfectly matched; for won games, however, the aliens are winning 66% of all games and the humans, 33%. This is because of 3 reasons: First, the maps are not built right, secondly the aliens gain major advantages on large maps due to eggspamming and ammo limits, and finally, the inequality of human and alien superunits at stage 3 in melee combat.
(note, some stats are old, but the percentages only differ by a few tenths at the most. The latest stats reside at
www.tremulous.net/balance and are compiled daily.)
1: Both teams are nearly-perfect balanced, kill-for-kill:[/size]
Say I were to gather statistics for how many times each side gained a stage then multiplied S1 stages by 0, S2 by 4 and S3 by 8 to reflect the number of kills per player required to gain that stage, without counting S3 as gaining S2 as well. What would the overall scores be?
Stage 1:
39056 Games ended with Aliens as Stage 1.
47121 games ended with Humans as Stage 1.
Stage 2:
36058 Games ended with Aliens as Stage 2.
30890 Games ended with Humans as Stage 2.
Stage 3:
65261 Games ended with the Aliens as Stage 3.
62364 Games ended with the Humans as Stage 3.
In total:
1288952 points total.
666480 points for the Aliens. (51.170%)
622472 points for the Humans. (48.829%)
Stages Gained:
As a control, what are the percentages for stages gained?
322198 stages gained total:
166580 stages gained by aliens (51.17%)
155618 stages gained by humans (48.829%)
Meaning, if you removed the scorecount due to differences in kills required to gain a stage on differing servers, the difference in percentages is less than 1000th of a percent.
I may not counting the kills gained at stage 1 or after stage 3, which are obviously skewed due to poor map starting positions for the humans and overwhelming winratio's for the alien team at stage 3, but the point remains that every game inbetween points to a very even 52/48 split. If the alien team were truely superior and truely unbalanced, they would be denying the human team far more kills.
In total, I'd call that a very slight balancing discrepancy, not one team being nubbish. The human team are grossing as many kills as the alien team, but for some reason, the alien team is winning a disproportionate number of games.
2: Aliens are winning 66/33% against the humans; how do we know this and where are they winning?[/size]
In games ending as a
S3 team against a S3 team, a total of
32253 games (26% of the total number of games played),
Aliens won 25852 games (80%) and the
humans won 6401 games (20%).
In games ending as a
S3 team against a S2 team, a
total of 27541 games, (22% of the total number of games played), the
aliens won 15360 games (55%), and the humans won
12181 games (44%).
In games ending as a
S3 team against a S1 team, a total of
20767 games (16% of the total number of games played), The
aliens won 11305 games (55%), and the humans won
9462 games (45%).
In games ending as a
S2 team against a S1 team, a total of
19386 games (15% of the total games played), the
aliens won 12342 games (63%), and the
humans won 7044 games (37%).
In games ending as a
S1 team against a S1 team, a total of
17652 games (14% of the total number of games played), the
aliens won 12431 games (70%) and the
humans won 5221 games (30%).
In games ending as a
S2 team against a S2 team, a total of
6164 games (4% of the total numbers of games played), the
aliens won 3408 games (56%), and the
humans won 2756 games (44%).
Gain from it what you will, but overwhelmingly the aliens are winning games more consistantly in S3VS3 and S1VS1 scenario's.
Why?[/size]
1: Compair the S1VS1 win ratio's of map such as ACTS, who's human base has 2 exits and the base is covered by turret fire adaquetly and the same goes for the alien base, to transit, which has 3 telenodes and a reactor, in the open, covered by nothing (same goes for alien base; 1 om in the open and 3 eggs).
Transit: (Victory ratio's at a particluar stage.)
O: Overall, (percentage of total games played on this map end in this way)
A: Aliens, (out of those, this many % aliens won)
H: Humans. (out those, this many % humans won)
S1VS1
O:3111-24%
A:2136-68%
H:975-32%
ACTS:
S1VS1
O:3562-11%
A:2321-65%
H:1241-35%
In other words, 24% of all games played on transit end with both teams at stage 1 and ACTS ends with that being at 11% (The overall average is 14%). Out of those games, 68% end with an alien victory at stage 1 compaired to ACTS's 65% (70% alien win average during S1VS1 overall). What causes this difference? Why is it so likely that the game ends that way and more often yet retains a proportionate win ratio to the overall average?
Because neither base starts with defenses but the aliens move out more readily and are better able to have a presance on such a large map.
You will find that the reason the alien team wins more often at stage 1 is because they are able to more readily able to take advantage of the human teams vulnerability at stage 1 when they start out with a crappy base setup. This skews statistics heavily in my opinion, with the aliens getting a free 70/30 win ratio over the humans. More importantly, a sure-fire sign of the imbalance is when games are won more often below 6 players than otherwise such as on this graph:
Compair this graph to uncreation,
S1VS1
O:879-11%
A:529-61%
H:350-39%
All graphs spike at lower player numbers naturally due to players leaving when a game is going badly before it is over. However, when they spike a great deal near the bottom, it is an issue with balance at S1VS1 (also notice it's 3% less likely for the game to end in a S1VS1 ratio for uncreation AND it's the least played map). This is because 1 rambo is better able to destory an enemy base with fewer players around than when there are more and therefor, it is more likely the game will end early.
2: Aliens gain advantages on larger maps and humans find hinderances on larger maps.
Stage 3 eggspam is a major problem for the human team. In a 10-on-10 game, if the aliens get their base smashed, 3 or 4 grangers may spawn and begin rapidly spamming both the overmind and eggs. When the human team begins an easter bunny and egg hunt, they have a major problem; they MUST split up to effectivly counter this tactic by continuously saturating the map. They have to check every nook and cranny enough that should an egg spawn, it is killed before it gets itself up. At stage 3, a human in a battlesuit with a lucifer cannon is at a disadvantage against a tyrant in certain areas (hallways and open areas, mainly. Closed in spaces that allow the human to jump and avoid the tyrant makes the tyrant toast)). Therefor, that egg hunt is often stopped before it can be completed by the human team getting picked off 1 by 1 and eventually closed in.
Simply put, there is no counter to eggspam on larger maps.
Additionally, the human team is less able to attack the alien base due to the inability to get there with ammo and is less able, even in groups, to kill larger game with their weapons because each fight they get into wastes more and more ammo and the longer it is between bases, the more likely it is the humans will come into conflicts with the alien team and waste ammo.
3: The game starts, for the most part, even at stage 1 with the humans having advantages in hallways and closed in spaces and the aliens having advantages in wide open areas. In wide open areas, the aliens are better able to sneak up and attack the humans by wallclimbing and may attack from more degree's; as the possibilities of attack diminish to 1 direction, the aliens become easier to kill. On arachnid2, for example, the slow-opening door to the alien base is a perfect example of this issue. A dretch will approach the door, wait dead center and the moment it opens an enterprising human will open fire, killing the dretch. In a hallway, the dretch has a couple of movement opportunities; they can attack from around a corner or on the celieng, they can jump from wall to wall but for the most part they are advancing on the human from 1 direction and retreating in one direction. Same goes for dragoons, balskalisks and marauders.
At stage 2 the advantages humans have in closed in spaces increases with the plasma rifle/helmet/battery pack combo with grenades. In open areas, only their defensive capability is increased and this deals with threats such as zap and gas attacks as well as headchomps. Large aliens such as the dragoon fare very poorly in hallways against humans with the right setups who can aim simply because they are unable to move quickly enough and are big targets. This is best shown on smaller maps such as ACTS and uncreation where the humans will outpace a S1 alien team at S2 massivly and get S3 before the aliens can even get to S2 due to the killcounts. Uncreation, for example, has a S3VS1 and S2VS1 winratio for the humans. The humans win 62% of S2VS1 and 64% of S3VS1 games on uncreation, for example. ACTS shows a similar deviation.
Uncreation:
S2VS1
O:1072-14%
A:47+350=397-38%
H:629+46=675-62%
S3VS1
O:1362-17%
A:21+470=491-36%
H:852+19=871-64%
S3VS2
O:2025-26%
A:147+713=860-42%
H:1083+82=1165-58%
ACTS:
S2VS1
O:1857+2300=4157-13%
A:222+2098=2314-55%
H:1635+202=1837-45%
S3VS1
O:3005+2566=5571-17%
A:114+2459=2573-46%
H:2891+107=2998-54%
S3VS2
O:4213+3454=7667-24%
A:376+3270=3646-47%
H:3837+184=4021-53%
At Stage 3, the humans lose their hallway advantage to the tyrant and maintain very few advantages in open areas. This is best shown by uncreation; due to the closed in hallways, humans must take on tyrants 1-on-1 and they will lose as they cannot charge up quickly enough with a lucifer cannon to get the 2 shots off that are necissary to toast a tyrant at point blank range.
Human bases are most often ajacent to hallways, and therefor getting pinned in at stage 3 is common because they simply cannot concentrate their fire collectivly down a hallway to toast an alien unit. Ranged units which are aligned liniarily do not fire well toward their obliques.
====================
*
RRRRR XXXXXXXXXXXXX
*
====================
The 5
Rangers can fire fine at the * (dretches) but in the X-zone, they block eachothers fire.
More importantly, when overtime sets in, a human team that is unable to move out will be crushed by snipe and tyrant rushes. 1 by 1, turrets dissapear and eventually vital equipment is destroyed. The aliens have a major advantage at stage 3 in that if their overmind is destroyed they can rebuild it and therefor evolve; if the human armory goes down, the human team cannot rebuild it.
What can be done to rebalance tremulous?1: Change the starting locations and configurations of the alien and human teams. Minimally a human base should have enough to stop a dretch rush and the aliens should be notified well in advance that their base is under attack before the first eggs go down.
2: The Overmind takes the combined time of a reactor, medistation and armory to build (basically, 4 times what it takes right now) and gets a HP boost to 1500HP (now being much more important than it was before). In OT the buildime of the overmind is doubled again, to 8 times what it is now (to properly equate it to the loss of a medistation and armory during OT). Meaning, when it goes down, you had better have a plan for defending the remaining sections of your base until it gets back up. This stops the spamming of the overmind, and requires the alien team to actually defend their base but also gives them more time to get to the overmind to defend it.
3: Reactor and Overmind spawn with 25% health already established, and all other buildings spawn with 10% of their health. This reduces the chances of a single dretch or spawn human destroying a newly establishing base and killing a game before it has started.
4: The battlesuits cost is increased to 450, the battlesuit gains radar and certain weapons gain 1.25 times the amount of carrying capacity. For clip based ammo weapons, this goes into the clip size; for energy-based clip-fed weapons this goes into the clip count. For belt-fed ammo based weapons this increases the belt size; for energy-resivour based weapons this does nothing:
Pain Saw: Nothing.
Blaster: Nothing.
Rifle: 40 round mag, 6 mags.
Shotgun: 10 round mag, 3 mags.
Las Driver: 200 round feed.
Mass Driver: 5 Round Mag, 5 mags.
Chaingun: 375 Round feed.
Flame Thrower: 190 round feed.
Pulse Rifle: 40 round magazine, 5 mags.
Luci Cannon: 90 round feed.
This should end most of the stage 3 ammo issues on larger maps and properly equate the tyrant to the battlesuit. Tyrants and advanced dragoons have the advantage of retaining their radar and therefor, a battlesuit has a major disadvantage if a tyrant can both hear and see the direction of the incoming human and counter the implied direction of their gun. Additionally, during end-game egg hunts, tyrants must now be afraid of locked and loaded battlesuits that can see them, as supposed to helmet/light armor toting search parties or guys with jetpacks.
5: The range of the tyrant is reduced 15%. A tyrant has a rediculous arm range when it is fighting; long enough to dodge fire from lucifer cannons and pulse rifles while landing hits. A 15% reduction would require the tyrant to dance much more tightly.
