Flat Balancing- Putting the good bits at the start

[Saktoth]

Ive been thinking bout some of the games ive been playing recently, and some design methods most RTS's follow, so i thought id take the time to write down some general comments on those designs as well as to explain CA's Flat Balancing doctrine, which is one of those Concrete Design Goals that people sometimes claim CA doesnt have.

TL;DR: Why save up the fun strategies and units for later in the game, especially if most games dont even get there? Why deliberately truncate the options a player can pursue?

In almost all RTS games, the complexity of the possibility tree begins very low, and branches slowly. You start with minimum options, perhaps a few choices of starting strategies or build orders. Usually its about getting your economy up, and this is generally done in a precise and efficient fashion for which there are very limited options. From there you get a basic unit, and you gradually unlock more units, and more specialized units, a few at a time, over time. Often, you must unlock a unit you dont want in order to get units you do, giving a clear sort of progression through the possibilities of the game. Often the most exciting, unique and interest units are near the middle or end of the tech tree, and the most interesting strategies come late in the possibility tree.

There are several reasons this is done.

Firstly it is because as the number of viable choices increases the game gets more complex. It gets harder to predict the enemies actions, harder to counter him, and harder to pick the correct actions to take yourself.

The first two of these are mostly a good idea, because scouting, tracking, and predicting enemy behaviour within a set of options is a large part of the game. Progression through these options allows a broader range of possibilities to bring the game to a conclusion (more on this later).

This last one is important though. If you can start krogoths, some players will. Thus, removing choices that are inappropriate for the situation makes it easier for new players. Faced with fewer choices the correct ones become more evident. However, you can still make Depthcharges and Wind Generators on Comet Catcher, so...

Almost all RTS's take the approach of a strict tech tree, allowing the gradual unlocking of units. The game starts simple, for new players, and then gets increasingly more complex as it goes on. Single-player campaigns follow the same trend, but large scale- each mission unlocks a new unit or technology.

This is inappropriate for a competitive multiplayer RTS, it is singleplayer legacy stuff. The first 10 minutes of the game should not be dumbed down or made predictable, options truncated, when they are perfectly viable routes to take. Having more options earlier makes the game richer and makes every game unique from the first few minutes. I personally find the first 3 minutes of *A utterly intolerable because they are basically the same thing, over and over, with no interaction with the opponent- 3 minutes can be a whole 10th of the game! Its like having to watch an ad you've seen 100 times before a movie. There is no reason the same principles of truncated options should extend to the first 10 minutes of every game too, when beyond that lies a richer, more diverse game that you rarely even get to play or explore if the game ends before then (This is especially the case in 1v1 that usually ends before t2).

The Second reason tech trees are used is progression over the course of the game. Having units that are 'unlocked' later in the game requiring an investment (an upgrade, a building) allows you to make the game scale in power to better, more cost efficient units the longer it goes on (note i dont say 'bigger' or 'more expensive'- i expressly mean better cost efficiency). You start with small skirmishes, then eventually get up to huge battles with armies that can level bases in seconds- escalation. Essentially this, along with the increasing number of specialized possibilities, ensures that the game ends later, rather than early. It ensures that you cant just field your best units at the start, and at least lets the game last for a little bit, but ensuring that it doesnt drag on forever because armies and strategies get so deadly and hard-counter that you become punished more for mistakes.

This is a good thing. However, a tech tree is a sub-par way to achieve it. Having units that are better or more cost efficient higher up the tech tree obsoletes units further down, basically cutting off a whole range of possibilities and mandating progression to these more cost-effecient units regardless of circumstances or if you genuinely need the tactical and strategic options those units offer. If the units are just more cost effecient but otherwise mostly identical versions of the earlier units then why even have them in the first place- its bloat. If units are not any more cost effecient, and just fill a different niche, there is no reason you shouldnt have access to them right from the start of the game- they wont ruin anything, they're not any better than other units, just different.

This is especially bad in mods with just two tiers though its also present in any mod with strictly separated tiers (Rather than many smaller, more fluid upgrades like *craft). When there are two tech tiers and the game is cut strictly in half, with a huge leap between the tiers in the middle. If a player techs successfully (IE, you dont kill him because of his large investment which takes time to pay off) you are generally forced to tech in response, as you cannot compete against more efficient units. Whats more, due to the large investment, this can create a 'gap' in the game, a long boring pause, as players pour their spending into tech and see little point in making an aggressive move with sub-par or inefficient units.

So do you still get progression without a tech tree? If you offer all options (equally balanced to an even level of cost-effectiveness but with important differences in their situational application) to the player from the start, but require that the only pick a limited number of those options (IE, one factory) you get an early fork in the possibility tree that means there are many more games that came be played from the outset- but the player must commit himself to one of them to begin with, allowing his enemy to know his unit set and understand the options he is dealing with. As the game progresses he can add more of these, mixing and matching as well as switching to options that are strong against his opponents current strategy/units. Thus rather than investing in more cost efficient units, you invest in more options in your arsenal to augment your current ones.

Economic escalation also provides more than enough of the 'investment for a better army' mechanic that you get when you buy a factory, and it is gradual and progressive (especially in CA, with the beautifully elegant Overdrive mechanic). It also isnt tied to any particular set of units or strategies, allowing you to mix your economic escalation with any number of other strategies or units you might wish to employ.

Finally, if you have all units being all balance to the same level and all being available at the same time, it allows you to balance the game a lot better. Its a given that t1 is always balanced better than t2, because t1 is played a lot more and the variables that might muddle the results are a lot less pronounced at t1. By the time you get to t2 there are so many differences in territory and economy that its hard to know precisely whether it is a unit, or something else, winning the game (add to this mohos, which come at the same time as t2, and its very difficult to know if it is the t2 units or the t2 economy doing the winning). Thus with flat balancing, 't2' can be balanced properly as it can appear at any point in the game.

Since these units appear on the battlefield at the same time, rather than one replacing eachother or being 'segregated', it also allows you to focus on accentuating the differences between units. This is another core principle of CA. If 't2' is no better than t1, just different, it allows you to ensure there are meaningful differences and to focus on creating and accentuating them. This adds a huge wealth of new options for a player to persue, and differences important and subtle.

For game designers and mod developers, this is a plea that, when designing a game where you play many iterations of it over and over (Rather than a directed, singleplayer experience that is played through once, say), ensure that the possibility tree branches early on, as this makes the game much more interesting from the outset and dramatically increases replayability. It fosters strategic thinking, versatility and creativity over the mere perfection of the execution of a pre-ordained structure. Dont save the richness of the players choices for the end of the game, as you're basically sticking most of the content in the part of the game that is explored the least.

[El Idiot]

Get a Job.

No really, a good read even for advertising. Glad someone is thinking multiplayer game design from the bottom up.

Single-player campaigns follow the same trend... inappropriate for a competitive multiplayer RTS

One of the major complaints I've always had with games.

[Pxtl]

I skimmed it, but one problem I always see: you don't need tech trees to get the tech tree gameplay. Look at L1 Ba gameplay. Your only real options at the beginning are the usual mess of mexes, energy, radars, and jeffies.

Any game with exponential growth is going to go through phases, simply because different units are useful in different numbers.

1000 metal of A is better than 1000 metal of B
5000 metal of B is better than 5000 metal of A, and unit C is still useless.
and so on.

Jeffies are great when nobody yet has money for a decent number of units - even more expensive units are inferior to jeffies in the early game, simply because jeffies can cover the ground better.

On the other hand, when defenses and combat becomes unavoidable, jeffies are useless and flashes win the day... until the swarms of units become too large, and the flashes get blown to bits before entering firing range, and get in each other's way - only a handful can bring their weapons to bear. Then you need stumpies.

No tiers, but a definite progression exists. You can't avoid progression of units without completely changing the nature of your game.

and as for the proposal: yes, I've often thought of it too - a mod where you have X factories that are all equal, and start with enough metal that effectively the first one's free... but diversifying isn't really possible until much later, so by picking one, you're committed. This allows simply a second factory - ANY second factory - to be effectively the "l2" game.

[Tribulexrenamed]

I played CA on castles with my newbie friends, and won every game with spiders and terraforming walls around my base while they spammed zeus.


I like it.

[FLOZi]

Just you wait for the next S44 release...

[[Krogoth86]]

@Saktoth:
Well I think your observations aren't that solid. While you have a point for most things you say there's a - let's call it "flaw" in your argumentation...

You limit yourself to the CA idea of not getting superfluous or useless units lategame imo. That's why you describe the situation of having limited options on the first levels and extended ones on the later levels. But that's not the case...

Well let me choose an obvious example and that is Maximum Annihilation. You pretty much have all the tactical options on all the tech levels because the units are made of roles and the basis scheme of scout, fast raider, skirmisher, AA and artillery are present for every tech level. So you roughly always have all the principal tactical options. There of course are difference like some special class types and some factional changes (like for example Sumo vs. Fatboy and Penetrator vs. Goliath show a mirrored behaviour of Tanks & K-Bots for area-damage and single-unit-damage between the two factions). You of course also don't at once get the possibility to fire your rockets over the entire map because that would be just lame and boring and so you get those special things as add-ons to the basis later on...

So you for example criticise that situation where there is a "cut" in the game when one player techs and the other one then is forced to do so too and "everything gets boring & slow". Imo that's one of the most interesting part of the game. That is not this exact point but rather the timespan before and after. Probably defending on the one side against a player who invests much more ressources in combat units while you tech on the one side and the strong effort to do some serious damage on the other side as you're obviously under time pressure is pretty exciting imo. Then when the "tides turn" and you can push back with your better units it's just a great feeling that "you did it" and that probably is what teching is about: It's a reward. It's this point in the game when you see someone say "Wow - he's T2 already!". It also doesn't get boring for the player who didn't tech - the time where you suddenly have to defend and win enough time to achieve a draw again is about as exciting as that of the first defender...

Then there is this argumentation of a flat-tree being more easy to balance. Ok - that's true but it's a dull argument for promoting one of the "techniques" to be more interesting. It should be about the game you play in the end - not about the effort you need for a good balance...

So to sum it up:
Applying what you call the "core principle of CA" leads to this T2 not being better than T1 and so having all tech-levels available right from the start probably is the logical conclusion following. But you now can't go ahead and rate things the way you do like calling more cost effective units of the same role "bloat". This might apply for the system CA uses but not for the multi-level system. Having a linear increase in levels which unlock "new" units is a reward system - something you completely removed for your model. You get rewarded for your performance (which probably consists of not being defeated) and at the same time is a tactical decision. It also is a good indicator for the players' skills. With teams / players of about even skills you'll see a game which will show multiple tech levels leading to more challenge as you don't have to "just" know this one flat balance of everything but have to know about the special possibilities of your new tier and how to use it. Maybe the teams are on par on T1 & T2 but Team A is a bit more experienced and so wins on T3 - such an incremental advance isn't really given in your system where you decide to go for a tactic and use the units which fit for that...

Once that is played out your system is a bit stuck imo if it didn't end already (i.e. your primary tactic worked) which might often be the case as you (I think it was you) told of CA having a way lower average timeframe per match. Once there is a draw it gets a bit dull imo. You can go change your tactics but you won't see those changes like in a multi-level system. So imo your system delivers a nice and interesting starting phase but on the other side lacks a bit of lategame experience (having to do about zero work about higher-tier eco adds up to this) and maybe it isn't that important here because of shorter timeframes...

[Warlord Zsinj]

I guess this is partially directed at me, given IW's design.

I really don't feel like being dragged into a huge balance discussion-fest (interesting as it is, there are better ways to spend my time).

But, to be brief:

- I don't think that simply lumping all the options at the player's feet necessarily equates to "yay, our players can build everything at once - they have so many options to choose from! Our game has heaps of depth!"

- There is not a huge difference between having an enforced tech tree (you can't build Z till you build factory Y) and simply pricing units such that the player can't afford the unit until their economy reaches a certain escalation point. Essentially they are attempting to achieve the same thing.

- There are elements of difference - having a flat build tree is purely linked to economic expansion. However, a branching tree is capable of this too - the price of the advanced builder, the factory, etc, is linked into this.

- Having a branching tree allows players to target their enemies' technological progression as well as their resources, etc.

- There is a natural evolution in games, and a reason why some units are not worth building at the start, or even at all in a game (say on a small map), while others are. This does not by inference mean that all the 'fun' is at the end of the game. If it is boring at the start, that is a flaw with the game design in general, not with the mechanic of the branching tree.

- In a design that is not based on a redundancy model, that is, all units are intended to be useful at some point, some units will only find their niche roles are produced through a surplus of some other unit type. This is part of a good game, as I mentioned earlier, and as you seem to reflect on.

- Yes, a flat build model can still have evolving niches, you just wouldn't build that unit until it's 'prey' is being used in proportions enough so that it becomes a worthwhile investment

- However, for intuitive design - that is, a design that is easy for players to pick up - a branching tree is a good way to maintain a design that is conceivable by a new player. By making the branches designed so that they follow the natural escalation process, you don't necessarily make the game frustratingly limited in options for better players.

- This is part of the craft of design, in my opinion. I think you can get away with it in CA because 90% of your players are either coming from a TA base or another *A base - they don't need to be taught the fundamental basis of the game.

- However, I think that providing a game design that follows a branching tree creates a more controllable game for new players - and good design allows people to not feel inhibited by this. If players find that they want to build something earlier (that is, there is a niche that needs to be filled, but they can't without first jumping through branch hoops), then the design is at fault - but not the branching tree mechanic.

I don't think that every RTS is wrong in making branching trees. I agree with you that some have not done it very well - but I think it is somewhat aloof to think that all else were foolish and blind to keep the idea of a branching tree, or that the concept of making everything buildable from the word go is so groundbreaking that people simply haven't considered it.

For example, could you imagine booting up age of empires or C&C, and having every single unit available to you at the first stage of the game? Players would be totally overwhelmed by the options. I suspect you would find noobs starting on mammoth tanks as their first units, and sitting around bored for 10 minutes as their inadequate economy struggles to build one of the most powerful units in the game.

If players have the economy to build the most powerful unit in the game, they will, whether there's a flat build tree or not. A branching tree is just as effective in this situation, so long as it has been designed to evolve with the player's evolving war on the ground.

Bah, dragged me into a long post. Hope you're happy now

[Saktoth]

Krogoth

Probably defending on the one side against a player who invests much more ressources in combat units while you tech on the one side and the strong effort to do some serious damage on the other side as you're obviously under time pressure is pretty exciting imo.

This can be done with economy. Building a fusion, for example, is a large investment with a big payoff, that inreases the size of the army you can field. Personally though, im not a huge fan of porcing or econ whoring, i admit this is partly a preference and playstyle thing but investment options exist as a part of the economy and share a similiar structure to what you describe.

You do have all options available to you at the start, Krogoth- but you can only pick one, out of 6/10 or so. This allows the game to branch down one path dependent on that choice.

Later in the game, you can switch strategies, by adding more options to your arsenal. There are some things labs do better than others, some things unique to each lab, and having a better mixed army lets you attack the enemy in new ways.

Since each lab, and each distinct unit within that lab, requires a different response from the enemy, every time you switch units, labs, or strategies you are opening up new possibilities- the game doesnt get dull because there are so many to pick from and each is designed to be unique.

Then there is this argumentation of a flat-tree being more easy to balance. Ok - that's true but it's a dull argument for promoting one of the "techniques" to be more interesting. It should be about the game you play in the end - not about the effort you need for a good balance...

T2 will never be properly balanced if it only ever comes about once other variables have already altered the game so much that the effects of t2 units cannot be precisely determined. Equally if only a smaller percent of the game is played with those units. T2 will always be less balanced than t1.

I guess this is partially directed at me, given IW's design.

Im not naming names. Its a problem in everything from Command and Conquer to EE. A large part of that is the singleplayer legacy which doesnt apply to spring.

- There is not a huge difference between having an enforced tech tree (you can't build Z till you build factory Y) and simply pricing units such that the player can't afford the unit until their economy reaches a certain escalation point. Essentially they are attempting to achieve the same thing.

Agreed, but this emerges dependent on map, on player playstyle and preference, on economy, on phase of the game, on choices made by players. And since this happens naturally as a result of the game, it shows you dont need a tech tree.

- Having a branching tree allows players to target their enemies' technological progression as well as their resources, etc.

You can have this with labs that take an initial cost investment (your 1k start), ala CA. The difference is you can start with any one (but only one) of the options, expanding to more later.

A branching tree is just as effective in this situation, so long as it has been designed to evolve with the player's evolving war on the ground.

The tech tree exists basically to follow the natural progression of the game into more niche units and to make it easier for new players to understand- this is precisely what i said, and i agree. I feel that is inappropriate for a competative multiplayer RTS and i believe that the reason it exists is a legacy from singleplayer. Importantly this is restricting the game artificially, dictating its pace regardless of the actual game situation- it is the designer enforcing how and when units should be used, rather than letting the players decide and letting it emerge naturally. I believe this is something you feel is a bad idea...

Being overwhelmed by the decisions or making the wrong ones- this is just something new players have to learn. But they WONT learn it in the space of a single game! You can get every single unit in the game, potentially, in the space of half an hour. This is why singleplayer campaigns unlock new techs gradually over each mission. How many times do players try to rush krogoth or mammoth tank or Elephants or whatever ANYWAY, seeing the tech tree as an obstacle they must slog through to get the 'best unit'. They rarely see it as an issue of 'make the bad units for now, work towards getting the good units' as they should.

With an early fork in the branch of the tree, IE, that initial fac choice, you only need to make one choice out of 7 or so, and then from there chose from 6-10 units. However, that first branching makes for dozens more tactics, right from the first few minutes. Players can just pick one fac, learn that one fac, specialize in that one fac, and branch out to learning others once they feel comfortable.

I dont think flat balancing is the only way to do things. I just think rigid tech trees are a trend in RTS's that deserves to be challenged. You dont have to do it, but this is why CA does it and i think its a good idea that makes a good game, though probably not a revolutionary idea. In truth CA does have a tech tree, it just branches sideways a lot more.

I think tech trees that allow gradual unlocks and incremental upgrades to old units that ensure they stay useful (*craft etc) are probably a better structure than these huge leaps and lurches across factory or tech tree you get in most spring mods.

I also think that some of the issues raised here apply to any game, and they are things we should probably all think about. I think its important that even a game with a rigid tech tree offer a range of possibilities early on, that it isnt a good idea to put most of the content in the game in the part of the game that is the least played, or to make the start of the game just a restricted going-through-the motions preamble to the good stuff.

[[Krogoth86]]

This can be done with economy. Building a fusion, for example, is a large investment with a big payoff, that inreases the size of the army you can field. Personally though, im not a huge fan of porcing or econ whoring, i admit this is partly a preference and playstyle thing but investment options exist as a part of the economy and share a similiar structure to what you describe.

Well the comment you quote was primarily about your statement of that teching up situation being boring as a classic example against a linear tree. I wanted to state that in fact I think this imo is one of the most exciting aspects of the game and not boring because that small "pause" that might occur between the two teams is part of the reward for good playing and can be very exciting because of the reasons I listed...

You do have all options available to you at the start, Krogoth- but you can only pick one, out of 6/10 or so. This allows the game to branch down one path dependent on that choice.

Well the only addition I see is that you can go for things that have high costs which is something you can't do in the multi-level system as you have to "earn" the right to build that advanced and expensive stuff first. Apart from that I see no extraordinary new tactical possibilities...

Since each lab, and each distinct unit within that lab, requires a different response from the enemy, every time you switch units, labs, or strategies you are opening up new possibilities- the game doesnt get dull because there are so many to pick from and each is designed to be unique.

Yeah but the difference is everything just changes when you go for another tactic. The multi-level system adds in another factor you are enforced to adapt to what makes everything more interesting imo and that is the difference between the tech-levels you'll encounter as the game goes on...

[det]

The whole concept of "tech trees" in OTA/BA is largely a mental construct. You are by no means limited to going adv kbot if you have tier 1 kbot, it's very fast to make a veh/veh con/reclaim/go adv veh. I used to do it all the time. If you have a few nanos, you can do it in under 30 seconds. The _only_ barrier to entry for t2 is the high fac cost. There is no "tree", there is just stupid micro. Is it possible to implement real trees? Within the OTA model, probably not easily, you can always reclaim. Maybe it could be implemented as a non-refundable 'research" model that unlocks new units in t1 labs. That has the potential to be interesting.

[[Krogoth86]]

Maybe it could be implemented as a non-refundable 'research" model that unlocks new units in t1 labs. That has the potential to be interesting.

Imo that's the concept SupCom went for...

[AF]

No because upgrading one factory from t1 to t2 doesn't upgrade all the other factories. The same applies to factories with morph commands.

[[Krogoth86]]

Well that's just a detail of that "keep your factory and make it build more but still let it have the old stuff" concept...

[AF]

If we're allowed to make that sort of leap of logic then a lot of the dictionary could be merged together resulting in a small pamphlet.

[[Krogoth86]]

Or you need to pay more attention to words...
I didn't say "That's exactly how they did it in SupCom" but it's "imo the same concept". And that concept is having labs that stay and are able to get new build options...

If this gets done by unlocking single units or doing one big upgrade on the building or doing one big upgrade for all buildings of that type isn't that important as they all are variations of that one single concept...

[AF]

Maybe it could be implemented as a non-refundable 'research" model that unlocks new units in t1 labs. That has the potential to be interesting.

Imo that's the concept SupCom went for...

Your sort of under the same problem yourself ^_^ Please practice what you preach! =p

[[Krogoth86]]

Hmm - well I guess I should have made this more clear in that posting towards det as it seems to be hard to understand but it's still as I told ya - I didn't say "Supcom did it exaclty that way" and to show you my train of thought I'll correct that quote you just used "against" me so you might finally understand how this was meant:

a non-refundable 'research" model

Imo that's the concept SupCom went for...

All clear now?
If yes let's get ontopic again...

[KDR_11k]

One problem with flat balancing is that some things just would be overpowered in the early game. Sometimes you have a unit that must be cheap to be cost effective at its job but must not appear before the mid or lategame because of the average play situation at these times. An example here is the Flow in KP.net, a fairly weak and cheap unit that can deal only with small groups of weak units effectively. Mid-late game that's not an issue as swarms of units or heavies are common and the Flow is limited in its uses but in the early game it requires a player to cripple his ability to expand to get a counter up. Since KP cannot have teching we had to make up a more creative solution but usually you won't want to resort to that.

Also teching creates a point of obsoletion for units, a player building up huge armies of T1 stuff risks having his units get obsoleted if his enemies produce less and tech more but of course they are vulnerable from it so he still has a chance to use his force to decide the game.

Overall tech upgrades may move some actions into the later game (though I'd argue a strategy requiring late-game gear is still valid, it just takes a few more steps to do) but also add the teching dynamic that adds actions.

[Gota]

When it comes to game designs i think both the branching and flat design have their positive and negative points and while u gain something u also lose something by choosing a certain game design.

Some of you talk of balance and game design in the wrong way.
IMO Balance is the frame and the game's limitations.Things like map sizes maximum time spent per average game by a player,game engine proporties,lua capabilities etc...
a few examples.

In ba for example you could have a unit that costs a trillion metal and be a mega giga super unit but it will never be built--->you need to limit your unit's proporties to exclude crazy options.
map sizes:Map size is very important because faster units will gain advantages on bigger maps while heavier yet slower units will ans should gain advatages on smaller maps.So you have to limit your unit proiporties to fit into an arrey of potential maps.
That set of maps should allow all units to be used effectivly decent ampunts of times..this should and can opnly be based on common sense.
gameplay design is how you operate within those restrictions you have set to yourself.
If the restrictions are clear they can guide you and make your game coherent and consistant if not you will have to struggle by yourself to make sense.

IMO in CA's case the resatrictions are not clear at all and perhaps a bit too restrictive for the gameplay ca devs are trying to achieve..
In order to take so many units and make them interestingly different but not in an ota fashion(making a few units that can achieve the same exact goal but require different gameplay/micro)i think you need much more veriants like innovative lua features that will allow you to alter your units enough to allow for ota quantities(and im not talking about replacing moho converters with a widget...(as you yourself see you are removing units,you have some other reasons to do it like IP and such but i think those are not really important ATM and surely do not justify the removel of units ATM).

You talk about same units doing the same things..but that is that OTA way that you fail to grasp..nothing wrong about it,its just different and not so much rock paper...

I also think you also rushed t osome bad desacisions in the name of flat balancing and trying to not be OTA like.
Your plat balancing is hardly complete and id say that if there will be a light at the end of the tunnel it will take ALOT of time to reach it..

About porc.
I think that porc should also be a factor to some extant to whether you win or lose and it can not be all about constant expension..i mean there is no written rule,experiance or logic to tell us that there should be no porc for example or some other play strategy.
Porc can be made usefull and skillfull and appropriate at some times..

About haveing tons of options from start and having to mix tons of units.
The human body is limited and cannot perform endless tasks that is why there are no games that make there units move faster than the speed players can select or see them.
Mixing of units is a 2 way street.
If you want force more mixing of units you have to slow down gameplay.
If you have to mix more units you lose the ability to move forces in some ways you could without such mixture cause u had more time to micro..like dividing your forces of same unit type to flank of foll ur foe..there are many maneuvers that can be performed to achieve victory over your opponent even if you both are using the same unit type.
Tell you the truth most players dont even use simple maneuvers not to talk about moving seperating your forces to many different groups and still coordinating them all perfectly ..So when you leave the unit numbers and speed yet force peope lto mix units what you get is chaos and randomplay whne players are forced to create groups of different types of units in hope that they will have the correct quantities of each unit so as to beat their opponent but it is very hard to guess what you will need by the time a unit is built to the time it reaches its destination for example..
in a 1 unit spam you know what unit u need and will need thus u can plan ahead better and game focuses about unit maneuvering and so much trying to mix your group's units just to the right amount..

At a small amout of units pregression like in ba for example where u preogress from weasel to gator and from gator to raider and leveler and than to t2 the winning conditions are generated by much more macro descisions the micro descisions you make like unit mixing and the strong consequences that follow those mico desacisions.

IN ca you are punished most for not choosing the correct set of units in a small time zone while in ba u are punished harder for making bad macro descisions like not switching from weasel to gator at a much bigger time frame...this makes the game smoother and not so suddden as when your much more metal worth force was totaly descimated by a much cheaper metal force cause it was the correct counter..this again means less focus on meneuvering and more on sheer unit mix for the sake of mixing..

[clumsy_culhane]

I have read all the the posts in this thread, all the way through. Except for Gota's. Dude, learn how to type properly with punctuation and correct spelling. Its impossible to read and I'm guessing most other ppl will read a huge post that's so badly formatted/written.