Zileas’ Guide to Avoiding Basic Formation and Firepower Pitfalls

Many times I’ll see players attack me with a massive line of single file troops.  Although many players often have the good sense to form up outside a base before attacking, often you can hit them when they are forming up, or even then, they may not attack with an optimal formation.

A lot of players don’t understand what constitutes a good formation, and what constitutes a poor formation.  Basically, think of it all in terms of firepower.  You want to deal out as much firepower as possible, but you want to take as little firepower as possible.  Single file marines going into a full group of hydralisks in line formation is very bad.  Single fire marines going into a full group of hydralisks in a shallow encirclement formation is even worse. Similarly, bottlenecks (narrow passages) can be used to your advantage with ranged troops especially, but with melee troops as well.

Melee Troop Formations: Shallow Encirclement and Flanking Actions

When attacking with melee troops, you have to be of course within the enemies firing range to attack.  The key when attacking enemy troops is to surround them(so that you apply maximum firepower), hopefully on an edge of their formation so that they cant bring all of their firepower to bear. Idealy you want to hit them such that their formation is effectively a single file line in relationship to your shallow envelopment formation.  Against other enemy melee troops, this can be a bit problematic if you are outnumbered, but if you are even numbered or outnumber them you can get far better kill ratios than you normally could by doing a shallow encirclement formation.  Basicaly this is a 30 degree portion of a circle, or a crescent.  This will cause them to be enveloped and take slightly more damage than they are dealing.  Do the geometry of two circles one on the outside and one on the inside (the length of the lines) if you dont believe me.  Deep encirclement, which is nearly a semi-circle is also possible, but if they manage to move a littlebit and encircle a bit along one of your edges of the circle it will really suck.  That and range troops decimate it.

Sometimes you will see formations where the opponent will put melee troops in  front to deal damage and act to hold your troops back, while dishing out damage with missle troops (hydras and zerglings are a perfect example. so are firebats and marines).  The simple way to counter this is a flanking action.  Engage their melee troops with significant number of your troops, but take the rest of your troops and walk around their melee troops, striking at their missle tropps.  since missle troops generally die in close quarters you should get much better results.  It all comes down to a question of firepower applied and firepower dealt.  Also firepower being applied to a  low hitpoints, high damage unit is far more useful than that applied to a low damage high hitpoints unit.  This is why ultralisks and sunkens should be low priority targets compared to say, hydralisks.  Similarly templars and infested terrans are some of the highest priority targets, as are defilers.

When to Use Bottlenecks

Bottlenecks are a way to max out unit mixes, as well as melee units.  The ways in which you can use a bottleneck effectively depend a lot on the power and type of attack of the units involved.  Generally its best for your melee  units (zerglings, zealots, firebats) to be surrounding the oppening of the  bottleneck, so as to concentrate a huge surface area (high firepower) on  their 3 across column of troops coming through.  Similarly, you want your ranged troops to catch ANY Troops while they are in the narrow bottleneck.   The best combination is using melee troops to delay their advance, all the while beating on them with ranged attacks.  This is particuarly effective if they are attacking uphill and are sustaining penalties to hitting your attackers, especially your missile troops (as a side note, attacking something on higher terrain in starcraft makes you suffer a 30%  chances to miss each shot…).  Keep in mind you can also create bottlenecks with good building placement, and you should.  A complete blockade works well too, and many terrans have held against large rushes using supply depot wall  offs and bunkers…..

Also as a side note, make sure you always use shallow encirclement formation for your ranged attackers against melee troops (at least before you use the disco tricks i will describe below…)

Ranged Attackers, Fire Control, and Xaanix’s Marine Disco Tricks

Ranged attackers are somewhat like melee attackers in that you want to concentrate large ammounts of firepower and take as little damage as  possible.  Firstly, you want to make sure that you shift que your attacks. This means attack a unit, then shift-attack another unit (Before the first one is dead) and so on going through their units that are near.  This will concentrate firepower on one of their units, and effectively cut down their firepower far faster than your is being cut down. This isnt especially practical in larger battles, but even doing this with a moderate number of your troops will help tons.  The idea here is that normally when people attack, damage is randomly distributed and only near the end of the fight do units start dying.  If you attack and concentrate firepower from one unit to the next, you will be cutting down their ability to do damage to you well before they do this to you.

You also need to make sure that your ranged attackers dont get attacked by melee troops, as melee troops will tend to decimate them (particularly zealots).  To get around this, you have to split your group, and “dance”. Basically run with both groups and whichever group doesnt get chased should fire.  Dont try this with dragoons — it doesnt work well, but it works VERY well with hydras and AWSOME with marines.  Xaanix has this really ugly disco routine with his marines where he kills often equal numbers of non speed upgrade zealots. “oh.. oh.. oh.. oh.. Stayin’ alive! Stayin’ alive!” (disco  quote of the year for me).  Beefcakecartman/~jolly~ is also known  for this sort of thing.  All good Zerg will dance around zealots in this  manner.  This is why if you actually do use these units, you need speed upgrade — it nullifies the disco tricks, at least to a certain ammount.  Remeber: Unit Control owns you.

So In conclusion:

1)Shift queing owns you.
2)Unit control in general owns you.
3)Why take all the damage when you dont have to?
4)Always think in terms of surface area and firepower concentration
5)Bottlenecks are yours to abuse.  Never assault a well defended bottleneck — go in from the side.  So many Zerg dont even know to hydralisk drop  when people use the bottleneck defense, and few protoss even bother going in the side with reavers.  Terrans of course can use siege….
6)Flanking actions, delaying actions, and shallow encirclement should be household terms.

or.. Starcraft Economics 101 by Professor Zileas

One thing that even the best players often overlook is the simple fact that technique and tactical skills can only bridge so much of a gap. With all things being equal, good techinque wins.  But, as I was painfully reminded when I lost to Fistantilus not too long ago, 2:1 kill ratios, even between identical units, wont matter if they have a much stronger economy.

Understanding when you are winning and when you are losing is very important. Generally speaking, You could mathematically generalize this something like this:  (ill put it in plain english)

Zileas’ Theorem of Starcraft Economics

If your kill ratio multiplied by the ratio of your production to their production is less than 1, you are losing. If their economy is gaining speed, and yours is stationary, and this number is close to but over 1, you are still probabaly losing.

When I say kill ratio I do not mean units killed/units lost; I mean RESOURCES killed/RESOURCES lost both in terms of unit production, miscellaneous upkeep costs (scarabs) and building production/loss.  Buildings also carry a lot higher cost in some cases when they cost them time — especially when the time prevents them from making cost-effective advanced units (i.e. killing a greater spire, killing templar archives).  For example.  Zerg X has 3 expansions, and has an income of say 75 mineral/second, 30 gas/second. I have 1 expansion and an income of 15 gas/second, and 40 minerals/second. I take one templar and kill 2 mutalisks, losing the templar.  I destroyed 200/200.  I lost 50/150.  I actually lost here because although the actual cost of the units may seem “Advantageous” to me, I’m actually overall behind on production such that this is a loss.   You need to realize what you can and cannot afford to do.  When i have the upper hand I throw away units a LOT more (i.e. sacrifice 4 zealots to kill a hatchery on island expansions, etc), which leads me on to another point…

The Importance of Time, as Well as Opportunity Costs and Opportunity Denial

Time is so so very key in starcraft.  Delaying your opponents ability to expand by a few minutes, or striking away the capacity to produce advanced units for a while is CRIPPLING at times.  Not only will this often force them to produce cost-inefficient low tech units (zealots, wraiths, hydras (in late game)), but it will put them on the defensive.  Also you need to think in terms of opportunity costs and opportunity denial.  For instance: If you delay someone from expanding for 2 minutes, youve not only killed their new command center/nexus/hatchery, but youve ALSO stripped them of 2 more minutes of mining minerals at that point, effectively killing 2 minutes worth of mining worth of troops. Moreover, youve also given yourself the opportunity to build up for 2 additional minutes, which makes this benefit even more.  Time is extremely valuable. Killing large numbers of overlords, or supply depots is very very crippling at times, particularly in an early game, as these structures/creatures take quite some time to replace, as well as costing resources.  Just as a side note, KGOR/Maynard who is one of the few players I fear, does this thing where he almost constantly builds probes/drones/scvs so that when he expands, he instantly has full scvs/probes/drones on the expansion…This strat has costs, but its a wise idea, especially later on…  A great player for sure

Conversely, there is a resource cost associated with attacking, especially with larger attacks.  When you attack someone, if you are just beating on their troops, you are effectively wasting your time unless you are doing insane ammounts of damage in terms of the resources formula I was talking about earlier.  There are hidden costs associated with this. The biggest cost is the money you COULDVE spent on expanding instead of those combat troops (assuming you maintained a reserve in your base). Attacking to break their concentration though, in addition to killing their units may be worthwhile… Which leads me to the ULTIMATE hidden resource: Concentration.

The Third Resource: Concentration

Minerals and Gas are the resources that most players think in terms of.  Although these are central to the game, you also need to think in terms of concentration.  I define concentration as time that a player has to spend focusing on a task during the game.  Expanding is a high concentration task, especially if you are protoss.  Attacking certainly has a high concentration level, and the more concentration you put into an attack, the higher the effect.  Even scouting carries a high associated cost.  One big difference between “Someone who is really good” and someone who is #1 is knowing when you need to watch a battle, and when you dont, and recognizing that your opponent also has a finite ammount of concentration to draw from.  There are a number of techniques for minimizing concentration costs (i.e. hotkeying buildings, using magic spell hot keys, queing attacks, etc.), but every thing you do has some intangible concentration cost.  I would argue that as you get better at starcraft, you go into a match with a larger innate concentration income/second   It is very possible when doing multiple cooridinated attacks at different locations to use your superior concentration reserve (if you have it) to decimate an enemy who is tied with you in terms of unit control and tangible resources. Although I’m sorry to say this, concentration is basically talent.  Playing a lot of games slowly raises it, but its something some people have a lot of and some people dont.  Its kinda like fast sprint ability in running: You can train up and become a great long distance runner, but for sprinting, theres always that talent based barrier — you can slowly improve it, but everyone has a limit.  Im sure that someone will push me off #1 who has more innate talent, along with the same skills…

The best way to train concentration, as Visage has reminded me just now over MIT chat, is to play 2 on 1s and 3 on 1s (multiple opponents vs you). I can often pull 3 on 1s, and certainly 2 on 1s, and really the only reason I can do this is my ability to multitask.  Also, team melee is an interesting game as it involves doubled concentration reserves on both sides… well almost doubled since its not one mind thinking at once and they have to communicate…