Decision Cycles (Tactical Theory)

[PzKw]

Forewarning:
This guide is a wall of text. There's some good stuff in it though.

Foreword:
In any competition where tactics are involved and the tactics can be changed by both sides at the same time the two players will compete to make better decisions. Every action taken will be the result of a process called a decision cycle. There are a few models for these cycles, but at this level of analysis there is probably only one worth looking at - the OODA Loop (also known as the Boyd Cycle or Boyd Loop). Tactical theory is generally very poorly understood by people who play video games, often even competetively and as such development of sound tactical acumen is very slow in players as they are required to learn it by intuition without any theoretical framework to allow them to understand it. There are other, more important aspects of tactical theory which have to be learned before something like decision cycles become important, but decision cycles are probably amongst the worst understood and least known areas amongst those not formally trained in tactics, that's why this is the first guide on tactical theory I'm writing.

This is a layman's guide to decision cycles and the OODA Loop with reference to HoN.


Introduction to decision cycles:
Decision cycles contribute to several tactical phenomena you are probably familiar with like tempo, initiative, aggression and mind games. Tactics are divided into two distinct parts: the science of tactics and the art of tactics. The science of tactics is the capabilities of the tools at the disposal of the commander or tactician - the ranges, damages, status effects, speeds, skills required and so on form the science of tactics. The art of tactics is how the commander interracts with his counterpart to trick him, deceive him, outsmart him, outposition him and dislocate his forces or his tactics. Decision cycles form the one of the most basic layers of the art of tactics. Understanding decision cycles is key to understanding how to defeat your opponents and their tactics and not just out attrite their heroes or outskill them in a manual sense.

THE OODA LOOP
As previously mentioned the model for understanding decision cycles is called the OODA Loop. The principle of the OODA Loop, and the most important thing to take away from this guide is that if you are able to **** your opponent's loop at any of it's stages, or produce a loop that goes faster than his, you will render his tactics ineffective.

Behold, the OODA loop:




Behold, the OODA loop dumbed down:





OODA stands for the following:


Observe
This stage of the cycle is the one where the actors take in the data and information that makes up the tactical situation. This includes data on enemy and friendly positions, dispositions, compositions, terrain, capabilities, movements and combat effectiveness (in HoN this means HP, MP, stats and cooldowns if known).

In some decision cycle constructs the Observe stage happens after action, but these are more complex and riskier models. For all intents and purposes both you and the enemy MUST observe before you can do anything and a failure to do so will result in a failure in your tactics, Fundamentally this stage of the cycle is the reason for the dominance of Wards - wards allow you to have more information to feed your cycle with to produce better plans and also allow you to conduct cycles faster as you may observe enemy actions before they detriment your freedom of action.

Note that you will never observe everything and even if you could you are not able to take it all in. There is too much data to collect quickly enough and even if you could you would not be able to process it quickly enough in the next stage. Working with a fragmented and incomplete enemy picture is a reality of tactics.


Orientate
This stage is where the data gleaned from observation is analysed and given meaning. Those dispositions, compositions and movements are turned into the enemy tactics and intentions at this point. Here, the analysis becomes about what people are trying to make happen with their actions and positions, how likely they are to succeed and what it means for you and them if they succeed or fail. It extends to what the friendly and enemy strengths and weaknesses are based on what they're doing and what they have, likely points of failure in their plans and their freedom of action to change them.

Finally all this information is used to generate all the possible courses of action you can take and tactics you can form. You will usually filter them internally before you are consciously aware of it to get rid of the obviously wrong ones based on your prior experiences, but you won't always get it right.

This phase is the one that takes the longest to develop in HoN. All the mechanical skills of HoN can be learned in a matter of weeks with some effort. There is no good reason a player can't learn to aim his hooks and arrows like a pro in a few days, his LSAs and Split Earths in a similar length of time, how to chain stuns without loss of stun time or how to last hit more or less perfectly in the same length of time. The reason it takes a minimum of months for most to learn to not fail at HoN is because they can't Orientate themselves based on the information from their Observations because they lack the experience to work out what what they're seeing means and what people are trying to do and so can't assess chances of success or likely points of failure to shape their plan with. Every skill which is considered difficult to develop in HoN is one that is fed from the Orientation phase: positioning, map awareness (which is basically the art of working out what the enemy is doing when you don't have any Observations to work off apart from their absence), gank awareness, lane control, actual juking/counter juking (ie. jukes that rely on you actually tricking the enemy and not just losing him in fog with multiple choices) and timing are all reliant on sound orientation.

Practice is required to build this and it is also dependant on who you are playing against. Organised teams and teams from different regions will be trying to achieve different things with very similar compositions and dispositions. You will form relationships with the enemy and start to be able to predict his likely actions. This is a skill which must be built to feed the rest of the decision cycle.

In real life as well as HoN the Orientation phase of the loop is the one most suceptable to subservsion by mind games and psychological biases. A wealth of sayings describe these tendencies and they aren't hard to find. The one I find most pertinant in the context of HoN is the saying "Once bitten twice shy." - you can establish control over a person by making them fear you with only limited bold play or a few instances of play where you completely counter everything he tries=. Once he has learned to lose to you he will stop trying to win and start trying not to lose resulting in more freedom of action for you (in psychological terms this is usually learned helplessness).

It is said that Orientation is the most important part of the OODA loop, but I disagree, it's just the hardest and most likely to fail in any tactical endeavour. Every stage of the loop is equally important.


Decide
In this phase, based on everything you can observe and all what you've assessed that everyone is trying to achieve and what their strengths and weaknesses and based on the possible courses of action you generated, you make another assessment on which one you should take. In real life tempo is slow enough in combat that you can normally conduct some kind of appreciation process to theoretically test which one will work best, but video games and HoN in particular preclude any kind of wargaming phase, forcing you to make the decision very quickly or not at all. Normally speaking a failure at this stage will be directly the result of a failure in the Observation or Orientation phase which results in the information and assessments feeding the Decision being incorrect and producing a plan that has weaknesses that the enemy is more able to counter than expected. There is a second way that this phase can go wrong which is discussed in the next paragraph.

In real life, this is the phase of decision cycles which seperates ordinary people from leaders and tacticians from commanders. Making a decision is in and of itself a skill which has to be built through exposure to decision making, a wide variety of situations, and most importantly innoculation to stress. The way this phase ****s up most often in real combat is through the commander being overwhelmed and being unable to make a rational/reasonable decision or not being able to make a decision at all. In my experience (of which I have a little, so don't take this as a slight, merely an observation) women and those lacking self confidence are more likely to fail in this area than others. The other way this manifests in HoN is when one team or individual(s) on a team give up and refuse to make decisions or stop caring about the decisions they make.


Act
Dead guy quote:
"A good plan, violently executed now, is better than a perfect plan next week. "
Gen. George S. Patton

Once a course of action is picked it has to be put into action as quickly as possible and executed as well as possible. The quote above becomes more important later in the analysis of the OODA loop as a whole, but it is most pertinent to the Decide and Act portions of the cycle.

Plans should be actionable as close to immediately as possble. Contingency planning is an important thing, but is almost always less important than planning for the current situation - a bias for action is imperative.

This phase covers how the plan is actually put into action (hurr, pretty simple, right?). It covers how well it is executed, whether it achieves its goals and what effect it actually has on the battlespace. The skills involved in actioning plans well are probably the easiest things to develop in HoN, with the hardest bit being learning how to work with teammates to get best effect out of these efforts.

Now you're at the end of the loop. But wait, a loop has no end so you start the loop again. You keep making decisions using this loop continuously so long as the game continues opposed and things keep happening. At no point should nothing be happening and at no point should the situation not be changing - if it isn't it means the design of the game is flawed and the so called "game " isn't truly a competition with another sentient being. You're probably playing in a farm meta and should relocate to an MMO which doesn't pretend to be anything other than an excuse to dry hump AI with limited interactivity until S2 fixes the game.

***

Application to HoN:

It's worth looking each of the parts of the construct in isolation because it will help you develop them and produce better plans and tactics. This mostly fits into HoN at the level above actual fights but below grandiose strategy and drafting. Players are usually aware enough about what their hero's best use is in a fight to preplan what they want to do when it is actually going (for example almost every Hag knows she doesn't want to accept engagement from three heroes at once and so she'll maneuvre to avoid that without anything more than cursory OODA looping to make her aware of the simple actions she needs to take to do that), and at the level of drafting and picking lanes you're not dynamically interacting with your opponent - they can't significantly alter your freedom of action and the tactics of these steps are mostly deterministic (if I pick this and ban this he will be forced to pick either this or this which will allow these picks and favour these picks).

The layer it most keenly applies to is the layer where your team is deciding what to do with their time in the game. Pushing towers, ganking, farming, chasing, running, covering, warding, positioning and baiting in every phase of the game are where OODA Loops come into play. All the actions where you attempt to have an effect on the enemy and the enemy can do something about it. Improving each step of the OODA loop will ultimately produce better plans which will have better effects on the enemy, produce better outcomes for you and thus more wins.

The (more/most?) important bit:

However the second, and much more important reason decision cycles are studied in tactics are because you aren't the only one making decisions and certainly not the only one with priorities that you are acting to fulfill. Every action you take against the enemy will shape his freedom of action and every one he makes will shape your's. The most significant thing in considering this is that it isn't just who makes the best plans and executes them, but who makes the plans the fastest and adapts them the quickest.

Consider a game of chess. Chess deliberately eliminates the OODA loop's concurrent nature by forcing the players to take turns in acting. Neither player can produce a higher tempo by making faster decisions (outside of lightning chess) and so only the quality of the plans matter. HoN is a real time game, and so both sides take their turns as fast as they can - they are limited only by the characteristics of their heroes and the battlespace so there is, as in real life, vast potential to make and act on decisions faster than the opponent.

Let's move back to the OODA construct to consider the implications of this. Let's say I take my steps 20% faster than him. On the first round, I Act as he Decides. He then moves to Act as I start to Observe how the battlespace is changing. By the time he is Observing, understanding what is happening I am already somewhere between by Orientating and Deciding on course of action adapted to his tactics and the new situation. By the time he understands what is going on, I am already Acting again and his understanding/Orientation is starting to become obsolete. The more we loop the more irrelevant his plans become until they become so critically weak to my adaptive plans that he loses his ability to control the outcome of the battle and has his freedom of action curtailed to insignificance, resulting in me deciding the outcome. Colloquially this is called getting "inside" someone's OODA loop, in technical terms it is called "decision superiority" and is the fundamental reason decision cycles are studied in military tactics.

Everyone has played a game where every single thing they tried failed. Ganks ran into counterganks with more heroes. Other ganks arrived to find no one there to kill anymore. Pushes got countered before they even got off the ground. Carries got ganked all the time. Every time you tried to do something the enemy had already countered it. The enemy was inside you OODA loop - his decision cycle was so far ahead of your's that he was effectively predicting your decisions before you were making them and shaping his courses of action to destroy your's before you had met the prerequisites for them to function. Your decisions based on Observations and Orientations that were out of touch with what the enemy was doing and were resulting in bad Decisions and Actions which didn't effectively cater to what he was doing now (as opposed to when you made the decision).

Context of decision cycles in contempory tactical theory:

The OODA loop is basically a tool of maneuver warfare which is the counterpoint to attrition warfare. Games you won by attrition you won because the enemy had **** synergies or no defeat mechanism - you won because you did more damage, had more disables than him, farmed faster, killed towers faster - whatever - you won by making bigger numbers than him. There was maneuver along the way but it wasn't what won you the game - what won you the game was your heroes, not your tactics. You butted heads with the enemy your's was stronger.

Maneuver warfare is different. It's about picking the enemy apart by targetting his weaknesses with your strengths until his strengths cease to function or be relevant. A draft for a maneuver based strat is quite different from an attrition based strat. Maneuver and exploiting decision superiority requires heroes that rapidly change situations and which are flexible enough rapidly change between different tactics to exploit fleeting opportunities and continue to counter the tactics that the opposition may try to change to. A draft based around decision superiority is not always going to be optimal, however achieving decision parity or superiority is extremely important no matter what draft you are running.

Decision Superiority in HoN (an example draft):

The sample decision superiority draft I like to run is as follows:
:fors:
(eg. Match 13942839)

It's pretty interchangable, and there are a few other heroes that are excellent in it, but that's how I like to run it. Core items? None on any hero - pick what suits the game and let's you target the enemy's critical vulnerabilities and critical requirements best. The lanes? I like to run it with Andro roaming (because it forces the enemy to work with incomplete and inaccurate information a lot as well as setting the tone for the rapidly changing situations) DS Hammer and then two solo lanes, but you can mix it around however you want. Almost any combination of lanes will work out of that as everyone has something to offer everyone else.

As a strat it's good at everything and bad at nothing but not exceptional at anything. There are good stuns. Good magic damage. Good AOE. Good -armour and physical damage. Good mobility. Solid carry (3 solid semicarries). Solid early game. Solid mid game. Solid late game. Plenty of utility. You have laning options that range from passive leach lanes all the way to high intensity dual stun kills lanes all the way to trilanes and defensive double lanes. It's not picked for tier, and it's also not picked for direct synergy - it's picked for flexibility and the ability to exploit decision advantage.

Because there's a lot of utility in every hero and all the heroes are good at changing situations in unexpected ways it lets you rapidly change situations to **** OODA loops. Being a jack of all trades strat it isn't tied into a limited set of courses of action - if it wants to and the enemy's strat is susceptable to it, you can choose to farm, stall, gank, push or control. It's not a great play from behind strat, and it also doesn't have a hard carry, and it also can't outright stall against a pushstrat, but it can sort of do all these things and can sort of do what it wants to be doing as well.

The most important thing about decision superiority strats are that they're a lot of fun not least because they are usually built around interraction with the enemy on as high a tempo as is sustainable with the intention not of denying or securing farm, but of beating the enemy's design for battle one action at a time.



Closing Points
Make sound, timely decisions. Improve the quality of each step of your decision process to produce better tactics and plans. Make your OODA loop faster and have a bias for action. Keep the enemy on the back good and always be planning your next move. Once you get inside his OODA loop exploit it ruthlessly. As soon as you've done one thing transition quickly to another while he's still formulating a response to the first. Counter the response you think he is planning. Be flexible, be adaptive and always try to seize and hold the initiative.

[PzKw]

Reserved for questions (please feel free to ask any).


***

I would just like to point out that 99% of people use the OODA loop all day everyday. It's the process to succesfully interact in any environment. Obviously the 'data' is not hon related, but this general concept can be applied to almost any gaming situation, or real life situation.

Also, in the big picture its more like a finite state machine than a loop.

At any point you can jump to act, because you may need to act on observations immediately because orienting and deciding take too long. assume lines are bidrectional arrows <->

can't get the picture to load on the page for some dumb reason.

http://img709.imageshack.us/i/ooda.png/

(The picture doesn't load because it's a .png file).

Thanks for the insightful reply.

You are completely correct about the OODA loop being applicable to other decision making processes (it's taught at a lot of business schools and I learned it in the military), but I would add one caveat - it's only a useful model when you're dealing with situations involving decisions being made against each other (ie. opposed decision making - tactics and strategy in any context) or in any other time sensitive decision making where the situation changes of its own accord. Rather than looking at a model of how decisions are formed (the OODA loop) in order to improve the decisions when it's not time sensitive it's far better to look at different appreciation processes to just improve the outputs (this is just in my opinion).

The Orientation and Decision phases can be collapsed into an intuitive step when no appreciation time is available - even in a reflexive action the mind interprets the Observation and skips higher thought processes altogether running through the Decision and Action automatically once it Orientates to the situation.

You are correct however, the cycle isn't necessarily a conscious one where each of the steps is followed in a rigid and linear manner - subjectively it jumps between stages. The most common deviation from a typical loop I'm aware of is when during the process of Orientating or Deciding the actor Observes a change to the situation and starts the process again - it's one of the ways someone's OODA loop can be subverted.

The point you raise about Action coming at a different point of the cycle is an interesting one and a valid one. My hands are unfortunately tied in the models I can and can't raise here because I can only post things I can find in public space. I know a couple of other models off hand that deal with this better and it's fair to say that while the OODA loop is the simplest model for explaining it it's also one of the least complete and least adaptive.

Edit: Please note that manoeuvre has been spelled incorrectly in every instance in this post and thread.

[yyr_]

Interesting post, will reflect on it

[Raskalnikov]

The section titled 'FORWARD' should read 'FOREWORD'

[[TECH]YawningAngel]

This needed writing, I almost launched into it mid-cast before deciding I was better off not putting my foot into my mouth. Genuinely insightful and useful.

[Drasha]

hm my gut seems to be running this loop constantly and sending the information to my brain! Interesting read only finished the first half will go for the second at some point.

[painkiller`]

Definitely an interesting read; appreciate the effort. I like the theoretical look at mind-games. Too bad that the average HoN player won't care, but oh well.

[pwn_U_fast]

I would just like to point out that 99% of people use the OODA loop all day everyday. It's the process to succesfully interact in any environment. Obviously the 'data' is not hon related, but this general concept can be applied to almost any gaming situation, or real life situation.

Also, in the big picture its more like a finite state machine than a loop.

At any point you can jump to act, because you may need to act on observations immediately because orienting and deciding take too long. assume lines are bidrectional arrows <->

can't get the picture to load on the page for some dumb reason.

http://img709.imageshack.us/i/ooda.png/

[StVicious]

Hello Sir,

I rarely post, but this guide is so well thought out that i couldn't stop myself from complementing you on it. I hope everyone who plays HoN reads this. (or rather I hope only people on my team read it...) the point is, I hope to see more tactical theory from you soon!

- StVicious

P.S. - should this be moved to the relatively underwelming "stratagy" section.

[Drasha]

Hello Sir,

I rarely post, but this guide is so well thought out that i couldn't stop myself from complementing you on it. I hope everyone who plays HoN reads this. (or rather I hope only people on my team read it...) the point is, I hope to see more tactical theory from you soon!

- StVicious

P.S. - should this be moved to the relatively underwelming "stratagy" section.

moved to the underwelming article section. It is now slightly less underwelming.

[PzKw]

The section titled 'FORWARD' should read 'FOREWORD'

Good pickup.

hm my gut seems to be running this loop constantly and sending the information to my brain! Interesting read only finished the first half will go for the second at some point.

Every good gamer in almost any game has a very streamlined OODA loop and consistently produce sound and viable tactics, however only a tiny fraction of gamers know the theoretical basis of tactics which means that it can take longer than it really should for counters to tactics that don't fit into their conceptual understanding of the game (ie. the metagame they're used to) to arise and secondly that it takes new players a very long time to reach a point where they are able to produce viable tactics and make sound, timely decisions.

Tactics are best learned theoretically first, then put into practice, then internalised and made intuitive. Without that initial theoretical framework people just try to learn tactics by hit and miss trial and error which wouldn't be such a problem except that there are so many factors involved that when they hit or miss it's very easy to misinterpret why they hit or miss that time and learn an incorrect lesson from it which later has to be unlearned and replaced (even then they could learn the wrong thing again).

I'm hoping that eventually there'll be enough theory on tactics out there that newer players won't have to try to run without walking or crawling first.

[[Poni]nobozumic]

very good read, thank you for taking the time to write and post this

[Drasha]

Finished reading all of it now.

Any ways when you are talking about maneuver warfare and attrition warfare its basically a ganking strat vs a turtle start. Pushing would probably fall under blitzkrieg or what ever term the military uses for some thing like that.

[PzKw]

Blitzkrieg is a form of maneuver warfare, but what the Germans practiced in WWII was actually closer to attrition by penetration. The Germans used to concentrate forces over a tiny section of the front and penetrate it, the idea was then to destroy resupply and command nodes to stop the other forces from being able to fight. In practice they actually just encircled those forces and fought what was called a cauldron battle in which the other force would try to break out over and over again and just get ar****ed in the process. The main similarity between pushing and blitzkrieg is the concept of having a single focal point of effort the germans called schwerpunkt.

The reason pushing isn't like blitzkrieg is because blitzkrieg involved bypassing defensive strongpoints (ie. towers) in order to strike at weaker units and areas. Pushing is more like the Soviet concept of deep war or some of the bold offensive actions undertaken by the Allies in the west like Overlord and Market Garden. Edit: Examples of maneuver based campaigns are Gulf Wars 1 and 2 and the 6 Day War. Modern examples of attrition include the Second Battle of Fallujah and the Second Battle of Grozny.

There are ganking strats that could be described as attrition based and some which are maneuver based. Old fashioned blink dagger heavy strats are basically attritionist in nature relying on being able to position unilaterally with and then dump huge damage on heroes. Likewise some turtle strats are highly maneuverist - 3 2 and a lot of Chinese DotA, much as I hate it, is extremely maneuvrist and relies very little on attrition in the main, however the Chinese tend to maneuver heavily in the early and midgame in order to create preconditions for successful and very, very simple, if overwhelming attrition later, that's the sad thing.

One of the things I most hate about the hero PotM is that every part of her looks elegant and maneuverist, when in reality she's just a fast hero with a lot of damage and a long stun. The potential for dynamic maneuver is there, but in reality she usually wins just by skill based attrition.

[Drasha]

So you would consider pushing more of a attrition style strategy? Pick strong early game heroes and just hit them with that until they fold basically.

[[S2]ElementUser]

PzKw, why don't you talk about this stuff in LN IHs? ~.~

[PzKw]

I'm probably contradicting myself all over the place. I'm not entirely clear on what I mean anymore which probably means I'm getting too wrapped up in semantics. Edit: I just reread some of the stuff in previous posts and I'm almost sure I'm contradicting myself all over the place. Just disregard whatever I say.


Attrition warfare is warfare where you seek to win by degrading the enemy's material effectiveness to the point that they can't effectively fight you anymore. In HoN I suppose that means running a strategy that seeks to gain an economic or experience advantage and hold it to seal a victory as an inevitability as opposed to running a strategy that seeks to beat the enemy by critically outplaying him. In practical terms I take that to mean that inflexible strategies that only do one thing (eg. a gank strat, a pushstrat, a turtle and carry strat) are all basically attritionist in that they seek to create a decisive capability overmatch (by degrading the enemy's, by securing tower gold early, by investing gold better in carries) as opposed to a more balanced strat that goes into the game with only the intention of outplaying the enemy.

Attrition and maneuver don't translate well into HoN because as I was taught, strictly speaking, maneuver war seeks to destroy the enemy's will to fight by shattering his morale and cohesion. It is important to note that maneuver doesn't preclude the use of attrition, it's just that it isn't the ultimate goal.

When I brought up attrition and maneuver in the post, I think I was more just alluding to the real life application of decision cycles. Armies that still cling to attrition based doctrine generally do so because they don't have the requisite materiel, logistics, intelligence gathering, communications and training to execute the advanced tactics maneuver war requires. Most notably however they lack competent junior leadership and mission command which means they have really, really slow decision cycles. Short decision cycles and a high degree of mission command are essential for maneuver war, but aren't important for a wide variety of attrition based tactics.

Again though, I'm starting to become unclear myself. It's late and I'm tired :'(

Back to find another IH, as long as it's not being hosted on EU or FR or something (who wants to hear about this stuff in an IH anyway?). (Edit2: I missed an IH while writing this)

[[REBL]Vodka]

PzKw you're my hero. Fantastic article, and a good read.
Now we just need an article on how to apply The Art of War to HoN and everything will be right with the world, haha.

[[REBL]Vodka]

Actually, I have the gut feeling that HoN is fundamentally about attrition just because the only way you can "properly" win is by killing the tree/shrine. Really, since your main goal is to assault a base and destroy a key structure, the lategame will most often come down to one team having some overwhelming advantage over the other, so you should focus on (skill-, farm-, whatever-)attrition to build up that advantage.

The emphasis on maneuverability and all is good for defeating the enemy in teamfights, but that's really secondary when you consider that the main goal is not defeating enemy heroes but killing a structure. I dunno, this is just my gut feeling. I'll probably refine this idea after a couple of read-overs.

[[BooC]KSlidz]

i agree with vodka in that the final win of a legit tournament game only finally comes from attrition, thats why you almost never see a comp team without a strong late game carry. however to gain this advantage maneuverability can be used and in pubs 15 min concede boil down to maneuverability