# Measuring complexity in 4e characters?

This google doc contains my current work.

I am working on a set of quantifiers that can reliably compare complexity between two different characters.

At present, I've operationalized 5 factors:

• Positioning (battlefield movement, where you have to be, and how much you can and do rearrange the battlefield)
• Attack Sequence (how complex your turn is, if you need a flowchart, or if you need to recompute strategy every turn)
• Out of Turn Actions (how much attention you have to pay to the battlefield when it isn't your turn)
• Status Effects (inflicting and removing status effects, buffs, and debuffs on others)
• Buffs (How many external factors or conditionally active factors require effective changes to your character sheet. Rages, THP, Polymorphs, Auras, etc... This measures how fiddily it is to achive "normal" tactics)

Is there a better set of factors that can more accurately capture the complexity of a 4e character?

Complexity defined (thanks to @Acedrummer_CLB) as: "a measure of the amount of things a player needs to do and be aware of to run a character adequately in a combat situation."

-
Knowing the ultimate purpose of the complexity measure would probably help. Are you looking at complexity relative to teachability, enjoyability, understandability, etc.? Are you focused on the player perspective, the rules perspective, difficulty in balancing, and so on? – AceCalhoon Feb 8 '11 at 4:08
Player perspective, difficulty of understanding and playing the character. – Brian Ballsun-Stanton Feb 8 '11 at 4:45
Wow. Huge undertaking. What you have outlined is (in computer programming terms) a 8 dimensional array (Race, Class, Level, Position, Sequence, OOT, Effects, Buffs) with weighted valuation system. Now my head hurts. – Acedrummer_CLB Feb 8 '11 at 16:02
It looks like you are defining complexity as a measure of the amount of things a player needs to do to run a character adequately in a combat situation. Do I have that right? – Acedrummer_CLB Feb 8 '11 at 16:17
Race and class are two separate dimensions, as is level I think. They express certain degrees of freedom that the particle can 'explore'. That is itself a 3 dimensional system. – Pureferret Jan 3 '12 at 22:01

## 5 Answers

As a computer programmer and someone who has done statistical analysis on survey data I have a suggestion.

Your parameters seem to be sound. You have broken out what seems to be the key point/values that have a bearing on your analysis. The suggestion I have at this point would be to alter your mathematical methodology a little. You are using a 10 point scale (1-10) with an even ending number. While this gives you a large range it forces an error into the math. When using an even number scale there is no middle value. Your middle number is either 5 or 6 depending on a judgment call. This forces a false evaluation of each rating that you would consider 'average' which skews the data results you get.

If you use a 3, 5 or 7 point scale you will find it easier to set up a middle value and from there determine variations from that. A 7 point scale would seem to be the optimum for your purposes. It can give enough range for variations while at the same controlling the level of variation between rating levels when choosing a rating for something. Since you are using a tenth percentage scale this should work well for you.

I am hoping that when you publish your findings you let the group here know about them. I for one would love to read them.

** added per request I would determine which factor is the most universal of those you are using, likely this is movement (since without movement you can't really do anything else). Select the next most universal factor and setup your math between those two items. Repeat this for each factor. This way you can tweak just the portion of the formula that applies to a given factor with out rewriting your whole formula, should that be necessary. Once this is done the overall target value of the formula should make itself apparent.

I had a thought after reviewing your factors. Setup each with sub-factors and then determine an overall value for that factor before any math. Example: Positioning would have maybe 4 subs (of Self, of Party, of Creature, & of Object) Each would get a complexity ranking possibly weighted and then an average for the would factor could be determined. This will capture all the data within each factor without complicating the formula.

-

My approach is inspired by The Bill James Hall of Fame Monitor.

Instead of a set scale, assign points to a wide variety of categories. It would probably end up being a lengthy list and I don't know if I'll make the effort to attempt to fully define it. I will give some examples though.

• Movement

• How often does the character have to move?
• 0: Only if threatened (typical ranged controller)
• 1: Rarely (bow ranger)
• 3: Frequently (fighter)
• 5: Nearly every turn (charger, melee rogue)
• Can the character move though opponents?
• 1: 1/encounter
• 2: > 1/encounter
• Can the character ignore opportunity attacks?
• 0: Strict yes or no
• 1: Sometimes (either encounter power or conditional)
• Out of Turn Options

• +1 per Immediate power
• +1 if has both Immediate Reaction and Immediate Interrupt
• +1 per triggered opportunity action beyond the usual
• Marks
• 2: The character can mark (count mark-like powers as well - Favored Enemy/Warlord's Curse)
• +1 per additional mark

And so on and so forth. Once you have an initial "complete" list, run through a few characters and see if it feels right. Tweak numbers as appropriate, add new categories to address missed areas.

-

Don't forget variety of action types. Some of my characters only use standard action powers. Others have uses for their standards, moves, and minors. Guess which kind of character is more complex.

Other things to count/ways to count them:

1. Healing should be a buff
2. Count marks as a status effect if you weren't already.
3. Hybrid and multiclass characters should count extra. They usually have powers that synergize well. Alternatively, count how many powers are intended to be used together as combos.
4. Possibly count each source book beyond the first. Not sure if this matters in the day of Character Builders though.
-

I am not sure if there is right or wrong way to break down character complexity into quantifiable data, let alone how to do it accurately.

Character complexity has to do with the role play aspect of a character persona, how hard is it to be a dwarf fighter and how does that effect the player not just the base stats. I have a Knight in a game I DM for, who is one of the most complex characters, Role play wise, in my group.

Speaking strictly rules though, I would count up all the different options that a class or build gives the player to use per action, including free actions, opportunity actions, reactions and interrupting actions for what ever level. It would be a by character comparison kind of a way to help choose what feat, power, trait, ext. you might want to look into for your next level up.

If you are really building a unified counter to compare all classes it could prove to be quite hard, there are a lot of options for each class, and there are always more coming in some form or another.

In short what I am saying is if you are going for bare bones by the books simple make a character that has two or three different things they can do per turn action (IE Standard, Move, and Minor) and one maybe two things on an Opportunity action. A simple character should never have to worry about any other type of action. Effects (if any) should be easy to understand and easy to use, all buffs should be always on, or have a light switch style on/off. The only real conditional thing the simple character should need to concern themselves with is Combat advantage.

Hope this helps some what. if not please let me know. -zach

-

For a given situation (terrain, nearby creatures and allies, conditions it might have), you could calculate the number of distinct actions in a turn that the target character could take. You could use that number (or, more conveniently, its log) to describe the complexity of that character for that situation.

You could then create a gamut of situations and calculate the complexity for various characters.

The branching factor is pretty big. For example, just considering a single move action of a character of speed 6 on hinderance-free terrain.

Just the fact that a higher level character has more powers would mean that it's inherently more complex.

-