Posts Tagged ‘Theory’

Starcraft II Single player: Zerg Research Tree

August 29, 2010 1 comment

After discussing the Protoss research tree last week, it shall be no surprise that the Zerg tree is the topic of this post. In my experience, making a choice in the Protoss tree was easier than in the Zerg tree, because the Zerg tree seems to be offering more specialized benefits.

5 points: Shrike Turret [BEST] vs Fortified Bunker

Adding life to a bunker with Fortified Bunker is nice, but adding damage regardless of load is better in many circumstances, especially since Shrike Turrets can also shoot air. You can break the food cap with Shrike Turret Bunkers.

10 points: Planetary Fortress [BEST] vs Perdition Turret

Even if Planetary Fortress was not great, Perdition Turret would still be redundant. As Terran player you already have access to Firebats (in bunkers) and Hellions to burn down melee weak melee units. The Planetary Fortress is the single most fortified structure you can build and a great stationary tank even against Kerrigan herself with repairs.

15 points: Predator versus Hercules [BEST?]

To be honest, I like neither choice. With a factory I prefer to be making Siege Tanks and if I get a Fusion Core up I might as well be making Battlecruisers. The Hercules does add in a level of mobility for large forces, but so will a group of Medivacs and those will heal organic units. The Predator does fair damage, but with greater risk to itself compared to the Siege Tank. Neither unit is necessary at all.

20 points: Cellular reactor [BEST] via Regenerative Bio-steel

Easy choice, the extra energy on specialist units (especially on Science Vessels) is much more beneficial than slow regeneration. Only take Bio-steel if you never ever make a specialist unit at all (which is silly).

25 points:  Psi Disruptor [BEST] vs Hive Mind Emulator [Specific use]

The key signature of Zerg is that they attack with many expendable units. That makes the Psi Disruptor seem more attractive on first sight. There is only one unit I would really like to take over and that is the Brood Lord. Even Ultralisks can be easily dealt with as Terran and a mind controlled Ultralisk would last only seconds against all the other Zerg it is attacking with. Slowing down a whole horde in a fairly large area on the other hand is much more useful in most of the missions.

However, if you face air in the final mission, having Brood Lords on your side makes dealing with all ground forces much easier.

Categories: Starcraft II Tags: , ,

Starcraft II Single player: Protoss Research Tree

August 23, 2010 1 comment

Gathering Protoss and Zerg parts during the single player part of Starcraft II: Wings of Liberty gives some pretty interesting choices. Let’s take a look at what the Protoss have to offer:

5 points: Ultra-Capacitators [BEST] / Vanadium plating

Speed and Damage increase versus Health and Armor increase. 5% is not that much, however, it adds up over many units. The more units you have, the better the damage increase relatively gets compared to health and armor which only matter on a single unit with a large health pool like a Thor.

10 points: Orbital Depots / Micro Filtering [BEST]

Supply depots can be build instantly or refineries produce gas faster. With a bit of planning you should never get supply capped and there are several missions where you start with a large supply anyway. On the other hand, getting gas faster on timed missions is a great boon

15 points: Automated Refineries [BEST] / Command Center Reactor

Getting Automated refineries means you no longer need SCV’s on gas, but you still get the same amount of SCV’s to start with. At food cap, you save three food per refinery to use on armies. It also allows you to build refineries without a Command Center nearby. Command Center Reactor on the other hand is just nice at the start of the game to get some SCV’s up quicker, but it becomes mostly worthless quickly.

20 points: Raven / Science Vessel [BEST]

I love the Raven, but the Science Vessel is borderline overpowered. It can repair mechanical units with no cost except energy and that regenerates, which is a great on levels with little resources. Suddenly it is possible field a large air army that needs no repairs or backup. Banshees, normally quite fragile units, become a truly scary force if a few Science Vessels join them.

25 points: Teach Reactor [BEST] / Orbital Strike

Tech Reactor works both like a Tech lab and a Reactor and the double build ability also applies to high-tech buildings. So a Starport with Tech Reactor can build two Banshees at the same time. This gives both versatility and saves up resources that would otherwise be spent on additional buildings. Dropping barracks units anywhere on the field (with sight) by the Orbital Strike is fun, but units still have to build, so to drop a sizeable force anywhere, you will need many barracks and time the building of units. Putting a force in a Medivac is often more useful. Orbital Strike does shine on a single mission near the end though, but Tech Reactor outshines it everywhere else easily.

Starcraft II Basics: Scouting

May 30, 2010 1 comment

Scouting is one of the most important things to do during a Starcraft II map. If you do not know what your opponent is making, it will inevitably leading to nasty surprises. This article will cover several of the basics, including timing. Protection against scouting will be dealt with in another post.

Scouting is sending one of your units over to the opponent’s base / point of interest (like expansion spots) to find out what kind of units and buildings your opponents currently has and to get an idea of what he is planning. If you do not know where your opponent’s main base is, then that is the first priority of scouting. The distance between bases both on the ground and through the air is often an important factor of what will be built.


When to scout remains one of the most difficult things to learn for beginning players. Looking at your opponent’s base all the time may seem like a good idea, but by doing that, you are actually wasting resources. E.g. a Terran player using all energy of the Orbital Command to scout is not gaining minerals through M.U.L.Es.

The question to ask yourself is: Will information about my opponent impact the decision I am about to make?

For example: If your plan is to always build two Barracks and get one Extractor first whatever you opponent is doing, then you do not need any information from him up to that point.

For example: If your plan is to go all-in then you will probably not have to resources to do extensive scouting (you do need to know where your opponent is located), so scouting at this point will actually delay your attack.

For example: Can I safely expand now? depends on if your opponent has strike force ready to attack you.

For example: (As Protoss vs Terran) should I sacrifice economy to build a fast Stalker? depends on if your opponent is making fast Reapers.

There are many, many decisions to be made during an SCII match, so it is important to get as much information during a single scouting run to cover more than one decision. Learning that is a matter of experience.

Race specific

Every race has its own scouting methods. Basic workers are a good choice in many cases, although specialized units can be much more effective at it. However, scouting with those will automatically give your opponent a clue about some of your tech: e.g. do not scout with a Banshee if you are planning on using it to do a mineral line attack.


SCV’s make excellent scouts during the early game. Once the Orbital Command is up, it can be used for 50 energy. Note, however, that 50 energy could also be used on gaining minerals through M.U.L.Es. Fast air units like the Viking and Banshees (with and without) cloak are excellent choices as well. Personally, I like scouting with a Raven, because it can do some pretty safe harass with autoturrets as well. Reapers make fast and mobile scouts that can do massive damage to workers as well.


Probes are good scouts, but the Observer is the true scout of the Protoss. It is difficult to stop as it is permanently cloaked. Phoenixes work as well. If you are investing in Sentries, the hallucination of a Phoenix can also be used as a cost-free scout.


Zerg drones are good scouts, but they are soon overtaken by the cheaper Zerglings Speedlings can run through the smallest openings to get a total view of a base. They can also be buried at expansions to see if your opponent is building there. The Overlords make good early game scouts, but they are very slow without speed upgrade and thus vulnerable to early ranged units. A single Marine can take out a badly positioned Overlord and that will invariably lead to a supply problems. Later in the game speedy Overlords can do suicide flights. Overseers make excellent scouts as well. They can also drop the Changeling, but that will often be spotted.

Categories: Starcraft II Tags: , , ,

Global Agenda: PvP Tanking

February 28, 2010 Leave a comment

Always keep an eye on the healers.

Global Agenda is a combination of MMO and Shooter and it shows in the survivability of the classes. The Assault is the only class capable of tanking, but its ability to withstand fire relative to other classes is a lot less than, say, the survivability of a protection warrior compared to a mage in World of Warcraft.


Tanks have to do two things: ensure that the opponent attacks them and be able to survive those attacks. Computer controlled opponents typically attack the person that is doing the most perceived damage to them. Of course, if tanks did the most damage and were able to survive best, then no one would play a pure dps class. Therefore tank attacks do more perceived damage (called ‘threat’ or ‘hate’) than actual damage. Human opponents are, of course, not fooled by this trick. It takes more effort to convince them that they should attack the tank and not the easier target behind him.


A great planning tool for skill point allocation can be found at hexagenda. Changing skill points costs nothing in GA (contrary to most MMO’s), which invites experimentation. Although it is tempting to start with the Tank tree, the first points should really be spent in the Balanced tree. It takes a decent amount of skill points to become tanky at all and the levels to get there should be enjoyable als well. Jet Pack Power and Power Pool Increase greatly increase mobility and sustained fire rate, so I would always take those before going into the tank tree. Super Tank is the final skill in the Tank tree and it is something to strive to get as well, since it removes the movement penalty on shield use. Where points are to be spent, is mostly a matter of personal preference. However, I would always go for passive protections above shields. A decent example build is this.

Which weapons to take depends on the map, but in case of doubt: take the mini-gun. It can be fired for a long while, it pushes people back and it does additional damage to Robotic constructs. The inferno rocket is a good choice as well: while it is power hungry, it does massive AOE damage and it leaves a fire DOT on anyone hit.

Off-hands are matter of personal choice and the map. I always take the EMP-grenade to stun mechanical constructs and the Ranged shield, because these are useful in about any match. Perfect Target can be quite powerful, or a complete waste of time (see tactics, below). The knock-away grenade is pretty effective on some maps, especially near deep drops.


All maps in GA are about getting control of a circle. Being in the circle makes one a target and even if a player is not directly targeted, AOE damage will still hit him. Therefore standing in the circle is by far the easiest way to build threat in PvP. If the opponent wants to take control of it, they have to go through you first!

Protecting healers is another and it has the bonus effect of getting healed (most of the time). As said before, the Assault’s survivability is not that much higher than for other classes. However, healing  is potent in GA. An Assault that is getting healed by two medics is nearly unstoppable as long as the medics are alive. The Assault can protect Medics and other squishies fairly well from direct fire by standing in front of them, or pushing opponents away with the Mini-Gun. However, snipers and long range turrets (e.g. Rocket Turrets) should be dealt with by Recons.

It's OK to find a safe spot while the point is under your team's control, but be ready to jump back on it at any moment.

While most opponents in PvP are human, the Robotics class creates turrets and drones that are computer controlled. The Perfect Target off-hand ability forces these to attack the Assault, while the Assault is invulnerable. The Assault cannot attack during PT, but the team is free to deal with the turrets. Note that someone under the effect of Perfect Target is counted for point control. It’s possible to hit Regeneration before PT to ensure that you’re at full health when PT ends. While PT is a strong ability, it is solo merely a break in the action and thus it is most effective while in a team.

A team of Assaults pushing the point together is a force to be reckoned with, but it is also a prime target for Recons dropping bombs and other AOE attacks. Be sure to spread out.

Standing still allows one to fire the Mini- and Inferno-gun longer, but stationary targets are easy prey for snipers. It is also not necessary to stand on a point that is under control at all times. bit be ready to jump back on it at any moment!

Categories: Global Agenda Tags: , ,

All hail the Tank Mage

December 13, 2009 Leave a comment

It's a mage! It's a tank! It's a TANK MAGE!

The concept of the Tank Mage is as old as Ultima Online. For those not familiar with the term, check this article on gamasutra (it’s an interesting article about hybrids in which the Tank Mage is explained on the first page). Basically, the Tank Mage has the survivability of a tank and the power of a mage. What a surprise. If the mage has healing spells as well, the Tank Mage is a holy trinity all by itself. Designers try to avoid Tank Mages for PvE because it makes other players redundant and for PvP because it is by definition more powerful than any other concept.

Tank Mages in an MMO with classes are bad. They can solo more than anyone else and if they have to group they can stick well under the intended group size. It’s even worse in a PvP game. The constantly  shifting of class balance can suddenly allow one class to become a Tank Mage and dominate, while the other classes can do nothing about it. Therefore, Tank Mages should be avoided by the designers to keep their game healthy.

Tank Mages in an MMO without classes are unavoidable. If players have completely free choice then they will gravitate to playing the sturdiest character with the best damage output. The question is if this is a bad thing. In terms of PvP balance it seems no problem: if everyone is overpowered, then no one is. However, this is only true if damage output and survivability are balanced. Let’s use Champions Online as the example. While Mindful Reinforcement (ranks 2,3) was bugged, people had incredible survivability, but not enough damage output to defeat each other, which led to very long or infinite battles. On the other hand the current situation sees very high damage as done by STC (for rants about that, check the forums, in short: it removes 55% of someone’s health in 10 seconds while ignoring all mitigation) and (de)buff stacking with force cascade and power armor which leads to oneshots. In the first situation the Tank Mages can kill anyone but other Tank Mages, while in the second everyone can be killed, it just depends on who gets the first shot off. Neither situation is desirable.

PvE balance is just as difficult. Can the designers assume everyone is a Tank Mage? Probably not, which means some teams will struggle while others will breeze through the content purely based on power selection without any regard to individual skill in playing. Also, if everyone can do everything then the roles (tank, healer, damage) players can take up are gone as well. Nothing makes a WoW dungeon fresher than going through it in another role. The lack of roles will actually cause burnout sooner.

The worst situation that can occur is if it is very difficult to make a Tank Mage; only one or two builds qualify as one. Because all information about builds is open, many people will jump on the flavor-of-the-month turning the beautiful open choice system in practice in a one class game. This has already happened in CO to some extent: powerful PvP builds have scared many people away in tier4, while forcing everyone to pick up STC in lower tiers or die.

Since an open skill system cannot avoid the Tank Mage, it should be relatively easy to build one and there should be many viable paths. This will lead to a spread in skills and choices that will make the overall game easier to balance, while individual players keep a sense of power and versatility (especially while soloing). Also, because of the many paths to the Tank Mage, there will be some role diversification anyone, e.g. some will have more tank than mage and vice versa.

So, while the Tank Mage is something that should be avoided in class-based MMO’s, it is a concept that classless MMO’s will have to build around to provide a fun and equally challenging game to everyone.

The low cost of freedom

November 8, 2009 Leave a comment

buildshop01Whatever Cryptic does, there will always be Flavor-of-the-Month builds in Champions Online. Even if all powers would be perfectly balanced against each other, there would still be combinations of powers that would have a more than average synergy. That said, the FotM builds need not be that much more powerful than all other ones. To achieve this, three things have to be done: broken mechanics have to be fixed, broken powers have to be fixed and the cost of cherry picking the best powers from every framework needs to be higher. Cost is the main topic of this post, so I’ll keep it brief on the first two points.

Broken Mechanics and Powers

Some mechanics that are said to be implemented simply do not work (well). A good example is the knock up/back/towards mechanic. With enough endurance I can easily uppercut someone, haymaker them away, while throwing some force blasts after them that makes them bounce around as well. The resistance that should be applied is not there or too weak. An often abused power is from the Shadow Framework: Shadow Embrace with Fatal Allure, which can juggle someone for the full duration of the power.

Another is buff/debuff stacking. Types of (de)buffs stack with each other which results in the oneshots by Chest Beam and Force Cascade after some minor setup.

Some powers are broken into themselves. Swallowtail Cut remains the best example. It is the only ability doing percentage based health damage, so it may even be a fundamental problem in the engine. Imbue may be too powerful as well, but it may be brought in line if stacking is fixed.

Freedom should not be free

When building a powerset in Champions, we’re nearly completely free to pick whatever powers we want. That is one of the things that sold me on this game. However, I had not expected that we could do this would any significant cost. If someone wants to have weird combination of powers, say Uppercut, Invulnerability, Freezing Breath and Eye Beams, then he’s basically as powerful as someone who’s going full Might and he even has more versatility. Some powersets already have synergies like the Negative Ions in Electricity and the ability to trigger Id blades in Telekinesis, but after taking a few of these fundamental powers, there are plenty left to pick from. Staying in a tree should reward players without taking away the worth of powers outside the main tree.

There are many ways players could be given the incentive to stay within a tree. I will give two examples below: Bonus stats and Scaling with stats.

Bonus stats

I think that there should be passive bonuses for powers taken from a single framework, for example: five powers from Force would give 5% damage increase on all Force powers. Less than five powers would give no bonus at all and more should perhaps give another bonus, like 10% at 8 powers, These are just numbers I came up with. The bonuses do not have to be damage, there can be a healing increase, endurance cost reduction, increased damage reduction et cetera.

So let’s say player one stays within two frameworks. He gets five powers from Framework A(damage) and gets 5% damage bonus. He gets eight powers from Framework B(healing) and gets 10% healing bonus.

Player two takes powers from all kinds of sets. His A and B powers are less powerful than those from player 1, but he’s more versatile.

Player two has paid for his versatility.

Scaling with stats

Superstats define damage done and I think this is a great example of freedom implemented well. However, that means nearly any power will work well with any superstat. Healing depends on Presence, so a heal ability used by a Dex/Ego character is already weaker than one by a Presence character. There could be more powers that explicitly scale with certain stats. Someone who would cherry pick powers would get the benefit of versatility, but he would have to make stat choices that would leave all his powers weaker than someone who sticks to a focused build.


player 1 has powers A1 and B which scale with different stats. The powers are very different

player 2 has powers A1 and A2 and they scale with the same stat. A1 and A2 have about the same purpose with a small difference.

Player 2’s A1 is a lot more effective than player 1’s, but player 1 is more versatile, because he can also use B.

Again, player 1 has paid for his versatility.


Versality has no real price in the current game. This allows cherrypicking which in turn makes FotM builds much more powerful than other builds.