At long last, perhaps the most overdue writeup of all Paper Mario: The Thousand-Year Door’s combat systems; I’ve been researching TTYD’s battle audience system on-and-off for the past several years, and here’s the most complete look I’ve ever been able to give on it.

Introduction: The Audience’s Duality – Crowd of Individuals or Mass Hivemind?

Let’s start with a fundamental question; what exactly is the battle audience composed of, mechanically?  As it turns out, the game’s exposed view of it paints a pretty decent picture:

On the one hand, you have a colorful cast of individual audience members, each with their own seat and picked from a handful of different species, forming a crowd with varying composition (usually a relatively few distinct types at a time, as shown here, but potentially a potpourri of different Mushroom Kingdom denizens).

And when getting into details on how Star Power regeneration, thrown items, or area-of-effect events work, the specifics of who’s who do matter. But deep down, the audience secretly wants to be something much simpler; a simple decimal number, ranging from zero to whatever the occupancy of the auditorium Mario’s rank gets him booked permits. And it’s this numeric side of things that offers the simplest place to start gaining a full comprehension of the system; so let’s begin there, shall we?

Audience as Mass

There are a few different numbers one needs to be cognizant of that affect the trajectory of the audience’s size, one of which is fixed and the other two freely varying amid and between fights.

The game’s desired base target amount of audience, which will be referred to as the base, is a constant value based on your level, and doesn’t impact directly how much audience you start and leave a fight with (though it will play a heavy role into how quickly you gain or lose audience, as will be detailed later).  Specifically, this base value is equal to (3 x Mario’s level) + 5, up to a max of 200.0.  As such, your stage’s maximum capacity (ranging from 50 at levels 1-9 to 200 at 30+) far outstrips where the game “expects” your audience level to be until you reach ridiculously high levels (only finally meeting up at a whopping level 65).

The actual size of your audience at any time is a sum of two values: your “permanent” audience target value (referred to henceforth as the “target” audience value), and a “temporary” value on top of that (referred to as “bonus” audience) which is generally 0, but can be positive or negative depending on specific circumstances.  The “target” audience value is what’s saved between fights, which leads to a reasonable amount of consistency across multiple fights if you’re generally doing well or poorly at keeping the audience entertained.  By contrast, the “bonus” audience is used typically for exceptional cases (such as BINGOs, stage hazards or the packed house you start with for most boss fights) that shouldn’t reflect on your general performance, and as such won’t stick between fights.

At any point in the fight, the individual members of the audience will aim to fill a number of seats dictated by the sum of the “target” and “bonus” values; if something happens that shifts this total, the game will gradually have members fill in or vacate seats accordingly until the visible audience and invisible target values align again.  Now that the numerical stage is set, let’s look at the individual types of audience that make up the crowd…

Meet the Cast

Now that we’re familiarized with the audience types and their collective numeric underpinning, let’s look at how the audience gets filled out at the start of a bout.

Part 1: Getting the Battle Started – Audience Initialization

First off, it’s worth mentioning that as long as you don’t have any Star Powers obtained, the audience and BINGO slots will be disabled, rendering the entire initialization (and the rest of this article) moot.  Assuming that’s not the case…

Picking a Target Value

First off, the fight has its target value copied from the previously saved target value, and adjusted by a random amount.  Specifically, that amount is determined by picking two random numbers in the range [0, 10), and adding the following quantity to the target:

(randVal1 + randVal2 – 13) * 0.01 * (“target” value – “base” value)

This serves to fluctuate the starting audience amount slightly between fights, especially when it’s far from the expected base value.  If the resulting value is smaller than 0 or greater than the stage’s capacity, it’s clamped to that range.

At this point, if a fight should have a maxed audience (generally the case for most boss battles, notably excluding your first encounter with Doopliss, Lord Crump in Chapter 5, and Bonetail), your “bonus” audience value is set to the stage’s capacity minus your current “target” value, making their combined target equal to the max capacity.

Determining Composition

The rest of the setup has to do with picking the member types the audience is generally composed of.  For most “boss” fights (all but Red Bones and Hooktail), this composition is fixed, as follows:

• Magnus von Grapple – Only Punis and X-Nauts (initially filling the left and right sides of the arena, respectively)
• Grodus – Only X-Nauts
• Grubba, Doopliss’ final battle in Ch. 4, Cortez, Smorg, Magnus 2.0, all other Ch. 8 bosses – Mix of all types

For all other fights, the following process is performed to select the types that should appear and their relative frequency; each individual battle has its own table of base weights per “type” of audience determining which types can appear, and how often they should appear if their type is one of the types selected for the composition. (Note: Luigi doesn’t count as a “type” of audience and is handled separately.)  Here are some examples:

As a rule of thumb, the types of audience available in a battle are generally introduced one at a time as you progress through the game (This isn’t a perfect predictor, especially for Pit battles, many of which have a chance of anything showing up).  Generally, the weights are also higher for audience types related to the enemies in the battle, like Goombas in the first battle above; this doesn’t impact how frequently that audience type will be selected to show up at all in the battle, but does mean that it will typically make up a larger proportion if it is selected.

Given this set of weights as a starting point, the game does the following to pick which types should actually appear every time you start a battle, as well as their relative frequency:

• Pick the number of different types (N) the audience should contain, based on the following weights (As one might have experienced in-game, generally battles are composed of only 1-3 types of audience members.):

• Order a list of the twelve audience types randomly, and select the first N types (as determined in the previous step) in the shuffled array, skipping ones that can’t appear in the particular battle (e.g. Boos in the Ch.1 or Ch.4 examples above).  If there are N or fewer types available to choose from for this battle, then all of those types are included.
• Multiply each selected type’s base weight for the battle by a random number from 1 to 10.
• Check to see if Luigi should appear in the starting audience (which he will do 5% of the time).

Allotting Seats

The audience is then filled up to the target amount, roughly matching the composition determined above.  For instance, if the composition was Goombas with weight 6 and Koopas with weight 2, a starting target audience value of 80 should produce an audience of roughly 60 Goombas and 20 Koopas.  These weights are used to determine the type of new members entering throughout the rest of the battle (e.g. a new member in the aforementioned scenario would be a Goomba roughly 75% of the time).

Of note, Bulky Bob-ombs’ weights determine the relative number of individual members at the start of a battle, not the amount of space they take up; for example, if a battle composition has Dayzees with weight 8 and Bulky Bob-ombs with a weight of 3, a target audience value of 140 would produce an audience of roughly 80 Dayzees and 30 Bulky Bob-ombs (since they take up 60 seats).

Now that we’ve filled in the house, let’s take a look at all the various functions the audience performs throughout the course of a battle.

Part 2: Audience Participation – Events Overview

Once in the heat of battle, the audience’s main passive function is to provide the player with Star Power to power Mario’s Special Moves.  The details of calculating how much SP is gained from attacking or Appealing is already covered in detail by this previous article, so I won’t belabor them here.

Beyond that, there are two additional mechanisms where the audience has an impact on battle: throwing items during the player attack phase, and various audience events occurring after the enemy attack phase.  (For a refresher on where exactly these phases fit into the turn structure, see my previous article on the topic.)

Item Throws

Before each turn begins, the game determines a maximum number of random items the audience can attempt to throw during the following turn — this has an 80% chance of being 0, a 10% chance of being 1, and a 5% chance each of being 2 or 3.

During the player attack phase, while the player is idling or making menu selections, the game checks once every frame whether to spawn a random item if the max has not yet been met for that turn.  For the first item on a given turn, this has a ~1/1800 chance of succeeding (meaning the item will spawn after ~30 seconds on average); for the second it has a ~1/6000 chance (spawning after ~100 seconds on average), and for the third it has a ~1/16000 chance (spawning after ~4.5 minutes on average).  Note that these random audience throws will only happen one at a time, and they will not occur while any audience members are executing a scripted item throw based on an end-of-turn event (as described in the next section).

If the check is successful, the game randomly shuffles the audience members’ IDs and checks each until finding one that’s able to spawn an item.  Reasons a member might not be able to spawn one include being in certain states (e.g. entering/exiting the battle, asleep, etc.), being on the extreme left/right sides of the stage, or being a type of audience member that doesn’t have any items in its drop table.  If such an audience member is successfully found, it will pick a random item from its drop table (based on their corresponding weights) and hold it above its head for a few seconds.  If the player does not attack them during that time, it will then throw it at one of the player characters at random.

Post-Enemy-Attack Events

All other audience events are handled by a single list, which will have possible events placed in it under various circumstances, as such:

* The battle is considered to have “dragged on” if the turn count reaches 2x the sum of the starting enemies’ levels divided by Mario’s level, or 5 turns; whichever is higher.  The X-Naut rock throw event is placed in the list on the turn this count is reached.
** At the very end of every turn, there is a chance of all of these events being placed in the audience event list at once; this chance is equal to 3% x the number of turns since the events were last queued.  The Koopa event is included only if the “dragged on” turn count has been reached.

After the enemy alliance’s attack phase, any Bulky Bob-ombs’ lit fuse timers decrement if applicable.  If no Bulky Bob-ombs explode, and there are any events in the above list, the list is shuffled, then each event is checked until finding one with eligible audience members available to execute it.  After the event ends, or if none are eligible, the remaining queued events are cleared.  Finally, the turn counts for shelled Koopas and sleeping audience decrement (meaning those events actually last a turn shorter than advertised).

A few details about specific audience events:

• Boos are not picky about which audience members they attempt to turn transparent; they choose between all actors (including normally-untargetable Bomb Squad bombs) with equal chance.
• Shy Guys have an equal chance of triggering a falling object-type stage hazard or the stage jets / fog (unless Mario is a Rising Star, in which case it will always be a falling object).  The side of the stage it runs to doesn’t make any difference in how likely each actor is to be targeted.
• Thrown harmful items have an attack power of their base ATK – 1 + the number of items thrown, up to a maximum of 3 ATK; the can, rock, and bone each have a base ATK of 1, and the hammer has a power of 2.

Part 3: Ebb and Flow – Audience Gain/Loss

Finally, coming back around to the audience’s numeric representation, let’s take a look at how various events in battle cause the “target” and “bonus” audience values to change.

Fixed vs. Flexible Added Audience

Audience value changes can be influenced in two ways:

• Fixed value changes, which are typically part of events that are out of the player’s control.
• Flexible value changes, which generally are meant to reflect the player’s performance.

Fixed value changes always increase or decrease only the “bonus” value by the full change amount (clamping it such that the “target” + “bonus” values sum to a valid amount of audience for the stage, if necessary), and as such, don’t have a lingering effect on the audience size in future battles.  Flexible value changes are handled in a much more complex way:

• First, if the “bonus” value is non-zero and has the opposite sign of the value to add, the added value is used to cancel it out.
• If there is additional flexible value left to add/subtract, it is broken up into units of ±1.0 (plus one last unit composed of the remaining fractional part, if necessary).
• For each of these units, perform the following steps:
  • Calculate the following quantity:

((“target” audience / “base” audience) – 1.0) / 0.25

• • Round it down to the nearest integer; this is essentially a measure of how far ahead / behind the expected “base” audience value you are (going ±1 away from 0 for every 25% difference).
  • If this quantity and the value to add have the same sign, divide the unit by the absolute value of that quantity + 1.
  • Add the unit (or fraction thereof) to the “target” value, clamping it to the range of valid audience capacity, if necessary.

All this complexity serves a couple of purposes: first, it prioritizes cancelling out unusual circumstances over changing your “permanent” audience value, making (e.g.) the jump from a battle with a BINGO boost to one without it less jarring.  Secondly, it ensures that you don’t continue gaining audience at a constant rate even if you’re far above what the game considers a “reasonable” target for your level (and likewise, makes it harder to lose audience if you’re far below it).

As an example exercise, imagine Mario is level 15 (and as such has a “base” audience of 50), has a “target” of 74.0, and a “bonus” value of -2.5, and the player is slated to gain 7.5 flexible audience.  The following calculations occur:

• The game first tries to cancel out as much as of the -2.5 bonus audience as it can; in this case, it can cancel it entirely, leaving 5.0 remaining flexible audience to gain, which it splits into 5 units of 1.0.
• Calculating ((74.0 / 50.0) – 1.0) / 0.25 = 1.92, rounded down to +1; as such, the first unit of +1.0 is divided by 1+1 = 2, granting you a “target” of 74.5 and 4.0 remaining flexible audience.
• Similarly, ((74.5 / 50.0) – 1.0) / 0.25 = 1.96, rounded down to +1, granting you 75.0 audience and 3.0 remaining.
• Since ((75.0 / 50.0) – 1.0) / 0.25 = 2.00, rounded down to +2, the next unit is divided by 2+1 = 3, granting you an additional 0.333… audience; this is true for the remaining 2.0 points as well, granting you a final total of 76.0 “target” audience (and 0.0 “bonus”).

Audience Gain / Loss Scenarios

Here’s a comprehensive list of all the actions that can result in gaining or losing audience value in the middle of a fight: (Amounts marked as “x base” are in terms of the base audience value for Mario’s current level.)

* For “winning a battle quickly”, calculate the following value (rounded down), and if the current turn count is less than it, use a multiplier of the difference x 0.05:

(2 * sum enemy levels) / (3 * (Mario’s level + 1))

Attack-Based Audience Gain

Landing and receiving attacks has a varying effect on the audience, based not only on the performance of the individual attack, but also the player’s general performance in battle (each battle tracks counters of successive “crowd-pleasing” and “crowd-displeasing” actions).  The audience value multiplier induced by an attack is always “flexible”, and is calculated as such (starting from 0.00):

• For player attacks:
  • If an Action Command was performed successfully:
    • Increment the “crowd-pleasing” streak (to a max of 2), and reset the “displeasing” streak
    • Add 0.01 * the “pleasing” streak * (0.5 + 0.25 * Action Command level) to the multiplier, where the “Action Command level” = 0 for 2xSimplifier-Nice, 1 = 1xSimplifier-Nice, 2 for Nice, 3 for Good, etc.
  • If an Action Command was performed unsuccessfully and did not defeat the target enemy:
    • Increment the “crowd-displeasing” streak (to a max of 5), and reset the “pleasing” streak
    • Subtract 0.05 from the multiplier.
  • If the attack did nothing (i.e. dealt no damage/status, or secondary effects like flipping a shelled enemy):
    • Subtract 0.05 from the multiplier.
  • If one or more Stylish commands were executed:
    • Add 0.02 * the Stylish SP multiplier for the attack (always 1) * the number of Stylish commands (to a max of 5) to the multiplier.
    • Separately, add 10 Punies to the audience (fixed) if Punies were included in the audience’s composition.
• For enemy attacks:
  • If a guard was performed successfully:
    • Increment the “crowd-pleasing” streak (to a max of 2), and reset the “displeasing” streak
    • Add 0.01 * the “pleasing” streak to the multiplier.
  • If a guard was performed unsuccessfully:
    • Reset the “crowd-pleasing” streak.

Here’s an example scenario that walks through examples of gaining audience from attacks, using the Red Bones battle in Ch. 1, with Mario having only a Power Smash equipped and a base-rank Koops as his partner:

Concluding Remarks

This article was a heck of a long time in the making, and I realize covers a lot of ground; if anything audience-related seems to be missing or insufficiently explained, please let me know and I’ll consider making changes to it or making follow-up posts in the future.  Huge thanks go out to Koop and Kappy of the “Glitz Pit” Paper Mario challenge running Discord server for helping proofread this and the previous article, as well as providing suggestions on how to better illustrate stuff.

I would have liked to include more info on stage hazards in this post as well, but it’s long enough already that I feel it would probably be better suited to a separate look later.  In the meantime, thanks for reading!

Well, after a long hiatus (a non-trivial portion spent on general reverse-engineering Paper Mario: TTYD, a few SMF videos on unused content / specific trivia, as well as my two gameplay mods, the Shufflizer and the Infinite Pit), I’ve finally got more in-depth TTYD battle mechanics to share.  Let’s start off with one of the most fundamental and under-explored bits of the combat system; what exactly goes into the “turns” of The Thousand-Year Door’s “turn-based” structure?

Overview

At the 10,000-foot level, a battle consists of this basic structure:

• The “First Act”, i.e. anything that happens before the first turn.
• Main combat – a series of “turns”, each consisting of various phases where the actors (generally Mario, his partner, and enemies) may take turns using moves (or “attacks”, used more or less interchangeably for the rest of this article).
• Once a move, or some other battle event, results in one side meeting its win condition (either Mario falling without a Life Shroom, or all the actors on the enemy side being defeated), the battle goes through its end sequence.

That’s basically it; the first act and ending aren’t particularly interesting by comparison to the main “turn” sequence, so I’ll wait to touch on their specifics later.  For now, let’s dive a little deeper into what a turn consists of.

What is a “Turn”?

From a practical experience (casual or veteran), one generally might consider a turn in TTYD’s battle system to have a few distinct parts: the player’s characters move, the enemies move, and then some stuff happens between turns (and occasionally enemies might do something in reaction to a move).  In reality, it’s a fair bit more generically structured than that; every turn is broken down into a fixed number of phases, with a select few things occurring outside of them at the start / end of turns:

• Turn structure:
  • Start-of-turn stuff
  • Phases (1-5)
  • End-of-turn stuff

Well, then what is a “Phase”?

Now we’re getting to the real meat of the structure.  Each of the five phases looks like this:

• Phase structure:
  • Start-of-phase stuff
  • “Passive phase” – All actors execute “phase events”, which are generally used for scripted threshold-based form / behavior changes in boss battles, or passive things that happen at the start of a new turn (such as Mini-Yuxes spawning or Puffs changing position).  In the majority of cases, these don’t do anything outside of specific phases / circumstances.
  • “Active phase” – Actors execute “attack events” depending on their unit type, which allow them to perform moves (attacks).
  • End-of-phase stuff

All actors run their “phase events” in each of the five phases, and “attack events” in only one phase (Mario and his partner attack during Phase 2, and enemies attack during Phase 4).  The priority in which “phase events” and “attack events” are executed differs; both schemes will be discussed later on.

Putting together these roadmaps, we get this slightly more detailed overview of a fight’s structure:

• First Act
• Turn
  • Start-of-turn
  • Phase 1 – Start, passive phase, end
  • Phase 2 – Start, passive phase, player attack phase, end
  • Phase 3 – Start, passive phase, end
  • Phase 4 – Start, passive phase, enemy attack phase, end
  • Phase 5 – Start, passive phase, end
  • End-of-turn
• Ending sequence (triggered by one team meeting its win condition)

Now with a full sketch of what turns and phases consist of, we can go over the entire fight structure in detail in chronological order.

End-to-End Guide

First Act (“Turn 0”)

• All actors execute their “entry events”, if they have them (this includes the player characters walking in, and boss introduction cutscenes).
• If a character performed a First Strike outside of battle, that attack plays out before anything else.
• If Mario has Lucky Start equipped, a status will be picked at random between HP-Regen, FP-Regen, Dodgy and Electric and applied to him for 3 turns.
• If there was no First Strike at the start of battle and Mario is a B-List star or higher, the stage jets will be initialized to one of the four types (fog, ice, explosion, or fire jets), and fired if anything other than fog is chosen.  If there was a First Strike, this will not occur until an attack or audience member attempts to trigger their effects; presumably this was done so the first attack doesn’t have a chance of turning the jet nozzles towards the stage, making what was meant to be a cosmetic “start of battle” effect into an immediate hazard.

That’s all the pre-battle-proper stuff, now into the turn sequence:

Start-of-turn

• On the 2nd and subsequent turns, if Merlee’s curse is active, her turn timer counts down; if it reaches zero, one of the four curse types is chosen and stored, then applied at the first available opportunity (stored across turns or multiple battles, if necessary).
• Actors have their move count reset, generally to 1.  If they have the Fast status active, it will be set to 2, and if they have the Slow status active, it will be set to 0 or 1, based on the number of turns they have been active since they received the status. (These move counts will be recalculated if an actor is switched in/out, or is inflicted by / loses either status mid-turn).
• Actors have their move priority determined based on their unit type.  This will always be 0 in practice for all but a few actors (Grodus has a priority of “1” to ensure he attacks before his “X” satellites, and Shadow Queen’s hands have higher priority than her body in her demon form), but there is unused functionality for each type of actor to have different a base priority and variance (essentially acting as a “Speed” stat w/a range).
• If fog is active, its turn count ticks down; it will be dispelled on the second turn after it spawned.
• After this, statuses / status-like effects are processed in the following order:
  • Poison damage
  • Burn damage
  • HP + FP-Regen ticks
  • Happy Heart and Flower ticks (for turn 2+ only)
• Followed by their turn counts being reduced in the following order*:
  • Sleep, Stop, Dizzy, Confuse, Electric, Dodgy, Burn, Freeze, Huge, Tiny, Poison, Attack-up, Attack-down, Defense-up, Defense-down, Allergic, HP-Regen, FP-Regen, Invisible, Fast, Slow, Payback, Hold Fast, Disabled Command(s)
• After this, the five main turn phases occur.

  * Note: Since this happens even on Turn 1, you only get the Electric / Dodgy effects from Lucky Start for 2 turns compared to HP/FP-Regen’s 3.

Start-of-phase

The primary purpose of the start of the phase is to determine the order in which “phase events” and “attack events” will occur.  In both cases, this is determined mostly by X-position on the battlefield, but the phase and attack event (move) queues handle it differently.

For phase events, an actor’s priority is determined by the following formula (with smaller numbers going first):

Priority = abs(X position) + (1000 if X position ≥ 0, otherwise 0).

In essence, this means that actors will first be processed from the center of the stage leftward (to -1000.0 units, which is over two screens’ width offscreen), then the center of the stage rightward.

Since enemies never spawn on the left half of the screen (generally, the valid range of enemy positions is 10.0 to 170.0 on the X-axis), this generally means that Mario and his partner will be processed first (with the character closer to the center going first), then enemies, from left to right.  Since the player characters’ phase events never do anything, you can basically just treat phase events as always occurring left-to-right (barring a specific move that’ll be touched on later, at any rate).

For attack events, the priority calculation is simpler, using the move priority calculated at the start of the turn, then breaking ties by the absolute value of the actors’ X position.  In theory, this would mean that attackers’ turns could alternate between sides of the screen, but in practice since enemies are always on one side and player characters on the other, and they attack during different phases, this will never occur.

Once these queues are populated, if this is the 4th phase, Merlee’s Defense curse will activate if available, and will last through the “passive” and “active” phases.

Passive / Active Phases (Phase 1-5)

• First, all actors execute their “unison phase event” simultaneously if they have one; these are rare and usually only have any effect in Phase 1, e.g. Puffs or Piders changing position.
• All actors activate their normal “phase events”, in priority order.
• All appropriate actors (players during phase 2, enemies during phase 4) activate their “attack events” to move in priority order.  Each actor gets the number of moves determined at the start of the turn (or if they were switched in/out or received/lost Fast/Slow status prior in the turn).  In addition, actors inflicted with Rally Wink status get to perform an additional move.  If Merlee has an ATK spell queued, it will activate the next time Mario uses a jump or hammer attack, and deactivate after all events resulting from that attack end (this includes Bulky Bob-omb explosions set off by Fire Drive!)  Notably, audience item throw events will only occur during the player’s attack phase.

End-of-phase

Pretty much the rest of the events from here on out have to do with the battle audience.  There’s way too much depth there to dive into everything here in detail, so that’ll have to wait for a future article (hopefully less than three years from now).

• At the end of phase 2, any audience members holding an item will cancel their item throw event.
• At the end of phase 4:
  • The audience picks a new limit for the number of random item throws for the following turn, from 0-3 (this is also done when initializing the audience at the start of a fight before the “first act” starts).
  • Bulky Bob-omb audience members’ fuse timers decrement, and they explode upon hitting 0.  If none are present or explode on a given turn and there are any possible audience phase events queued (again, more on this in a later deep dive!), one of them is chosen at random and the rest are discarded; if it’s an item-throw event it will happen on the following turn, otherwise it occurs right now.
  • Sleeping audience members and Koopa audience members in their shells have their “inactive” turn counts decrement.
  • If the crowd is in an excited state from a BINGO, the turn count for that effect decrements by 1.

End-of-turn

• Audience members leave if the battle has lasted too long (5 turns minimum, potentially longer if fighting higher-leveled enemies).
• A list of audience phase events has a chance to all be queued (to be randomly selected at the end of the next turn’s 4th phase).  This chance starts at 0% at the start of a fight, and increases by 3% every turn until it occurs (after which it gets reset to 0%).  This is what causes all non-item-throw audience events, such as Bulky Bob-ombs’ fuse getting lit, or Dayzees putting the audience to sleep.

End-of-battle

Practically any individual step in the battle that can result in an actor becoming defeated will perform a check to see if the battle should end.  The end-of-battle sequence is fairly nondescript, as pretty much nothing impacts gameplay, with a couple of exceptions:

• Merlee’s EXP curse, if available, activates right before receiving the EXP gained from enemies.
• Grubba / Jolene’s battle conditions are checked at the very end of battle, after the level-up sequence if applicable.  This doesn’t have any bearing on conditions that were used in the actual game, but there are some unused conditions (notably “Mario finishing under/over HP threshold”) that might result in unexpected outcomes; I actually ended up changing where the conditions are checked in Infinite Pit to sidestep this issue.

That covers the whole overview of the general battle structure, but now it’s time to get into some of the unexpected ramifications of how the systems combine together.

Edge Cases & Addenda

Damage Events + the “Reaction Queue”

When one actor’s “phase event” or “attack event” damages or inflicts a status on another actor, this will cause them to start their “damage event” concurrently.  This will interrupt any other events the damaged actor is currently running, and may result in that actor being defeated, otherwise it will generally return them to their idle state after a short animation.

Actors that cause exploding chain-reactions (Bulky Bob-ombs, Bob-ulks, and Bomb Squad bombs) use a separate mechanism called the “reaction queue”.  Each entry of this array holds an actor id + another integer value, the former determining which actor is reacting, and the latter determines what kind of reaction event to look up in the actor’s data table.  Whenever one of these bomb-type actors is hit by a damaging move that can cause them to explode, their actor id and respective data table value are added to the back of the queue (unless they’re already there, in which case nothing happens).

Once the main phase/attack event and its resulting damage events have ended, the game checks to see if the reaction queue has any entries.  If so, the game checks first to see if either team has met their win condition, ending the battle if so; otherwise, the first actor in the queue’s reaction event is run, then that entry is removed from the queue, shifting all the remaining entries up (if necessary).  This process repeats until all reactions are removed from the queue, and only then can the next phase/attack event occur (or whatever the next step of the battle flow should be).  Note that the game will check for win conditions before every reaction event is executed, meaning the battle might end before all the Bomb Squad bombs in a chain have exploded.

Spawning New Actors

Plenty of things can cause new actors to be introduced midway through a fight (form changes, enemies calling for backup or respawning limbs / helper actors), which might raise the question of when they first get to attack.  Ignoring player partner switching (which has its own set of mechanics that I haven’t really looked into), there are three supported options for handling the move queue when spawning new actors, which are to insert the enemy at the first available spot, the last available spot, or not at all (which is what’s always done in practice).

Recall, however, that the move queue order is only determined at the start of the phase where the actors in question get to attack.  As such, enemies spawned during phase 4 (e.g. Green Fuzzies splitting or Koopatrols calling for backup) won’t get to move, but actors spawned during another phase (e.g. Mario splitting Embers with Fire Drive in phase 2, or Crump summoning the third kind of X-Naut platoons in phase 1) will get to move when their attack phase comes.

One small quirk is that actors spawned in the middle of a turn won’t have their move priority set for that turn, instead defaulting to 0.  In most cases this doesn’t matter since most actor types have a move priority of 0 anyway and ties are broken by position, but in the specific case of the Shadow Queen’s hands this does make a difference.  Her hands have a base priority of 100 (Left + Right hands) or 90 (Dead Hands) compared to the body’s 80; normally they would have a higher priority than SQ herself and as such would move first, but if they respawn or are switched by SQ, they will attack after her instead!

Just… Bomb Squad.

Bobbery’s Bomb Squad move is, in short, one big manifestation of the quirks of all the aforementioned systems working in concert.  As an overview of how the move works, Bobbery spawns three bombs that explode 2 turns later.  Specifically, they’re created with a turn countdown of 2, and their phase event decrements that timer by 1 during the first passive phase, immediately calling their explosion event if it ticks to 0.

The combination of the unintuitive priority of phase events, the fact that Bomb Squad bombs can be placed basically anywhere on the player’s and enemies’ sides of the battlefield, and the fact that their explosions can cause chain reactions makes predicting the exact sequence of events incredibly hard to predict.

For instance, compare these two subtly differently positioned sets of bombs (all set to explode on the same turn):

In the first case, the middle bomb’s X position is -2.0, barely left of center, so its phase event has the highest priority and it will explode first.  In the second, the middle bomb’s X position is +2.0, barely right of center, so its phase event’s priority will have 1000 added to it, making it higher than the left one, and thus it will explode second.

Things get even more complicated when you consider chain reactions:

The middle one will explode first, since it’s left of and closest to the center, but which chain reaction occurs first?  Well, it turns out the left one explodes before the right, but that’s only because the Bomb Squad explosion attack determines its targets left-to-right, and that’s the order they go into the reaction queue.

How about this setup?

The bomb slightly left-of-center will go first again, and will hit the ones on either side of it, but none of the others.  Since the explosion’s targets are processed left-to-right, the one on its left will explode first, and that bomb will itself set a chain reaction off with the far left bomb.  However, since the bomb to the first one’s right was already queued, that one will explode before the far left one.  Clear as mud, I’m sure?  But it gets worse…

No, seriously, Bomb Squad is bizarre.

Since Bomb Squad’s detonation is based on a phase event, what happens when you combine it with other actors with meaningful phase events?  The sky’s the limit; for instance, Grodus and Yuxes respawning their satellites is also something that happens in their phase event during Phase 1.

In the case of Grodus, that means that placing your bombs carefully can result in you knocking out his Xes either before or after he respawns them, potentially leaving him with fewer than he started with by the time your attacking phase starts (or even none, if you place them all barely to his right!)  This sort of behavior is well-known and abused in some of the harder TTYD challenges, notably Impossible Mode.

Perhaps most famous, though, is how Bomb Squad interacts with the Shadow Queen, specifically when she taunts you three times while in her invincible form in order to progress to the mid-battle cutscene.  In a casual playthrough, this would normally happen on three separate turns, but in fact, she checks during every passive phase whether she was hit by anything intended to deal damage or status since the last time she taunted.  As such, you can end this part of the fight in a single turn if you have Bomb Squad bombs on her left and right sides set to explode the turn she assumes her invincible form:

• Phase 1:
  • Left bomb(s) explode
  • Shadow Queen reacts once
  • Right bomb(s) explode
• Phase 2:
  • Shadow Queen reacts again, since she was hit since her last phase event
  • Mario or his partner attacks during their active phase
• Phase 3:
  • Shadow Queen reacts a third time, ending the fight before she gets an attack phase!

Funnily enough, this isn’t even the only way to cut this part of the fight short; if you’re lucky (or a TAS) and a Boo or Shy Guy from the audience attempts to hit the Shadow Queen as the post-phase 4 audience event, she’ll react to that as an attempted attack in Phase 5, cutting the fight short by a turn.

Concluding Remarks

Hopefully this gives you a bit of a taste of what TTYD’s general combat structure looks like under the hood!  I’m hoping to follow this up with more posts in the not-too-distant future doing deep dives on more specific parts of the battle system (including a long-overdue look at the stage hazard + audience systems, and one on all the different parameters that go into making attacks distinct).

Truly 2017 was a banner year for Paper Mario challenge runs, with the Glitz Pit Discord server surging in popularity and several players popping into the scene.  Discussion of strategies, loadouts, and mechanics of the Paper Mario series has abounded, but the badges, arguably the series’ combat’s defining trait, have always been a particular focal point.  Some badges are loved by all, some ridiculed by all but a few dedicated apologists, and many fall everywhere in-between.

Since stacking badges is what this blog was built on, I’d like to do something a little different here, and exposit my current thoughts on the matter at length, giving my personal ranking of every badge in Paper Mario: TTYD and some justification as to their placements.

To start, I won’t be considering these badges (the “FX-Esque Tier”) in my tier list:

• The Attack FX badges and W Emblem / L Emblem are merely cosmetic and cost nothing to equip, so they’re obviously a matter of taste (even if W Emblem alone is the clearly supreme stylistic option).
• Seeing as TTYD gives you many options to optimize strategies to minimize or maximize focus on HP, FP, SP, items, etc.;  HP and FP Plus are an invaluable convenience for adjusting your stats as you see fit, without needing to ever level up anything but Badge Points.  However, as they’re functionally equivalent to level-ups in HP or FP at any given time, they’re not really possible to rank alongside the majority of badges with unique effects.
• Timing Tutor is useful for building mastery of Stylish Action Commands for a nominal BP cost, and on the other hand, freely skippable if you are well suited to perform them.  An excellently designed badge for beginners with no use (but no downside, other than slightly clogging Charlieton’s inventory) to experts.
• Double Pain is good for adding extra challenge, and rarely (but occasionally) for manipulating Mario’s HP value.  No major positives, but it’s free BP-wise and is really only meant to be an extra challenge.
• Finally, Slow Go is obviously meant as a joke, nothing more.  Only shame is that it’s not hidden somewhere needlessly cryptic like its predecessor in Paper Mario.

As for the rest of the badges, I initially grouped them roughly into four categories – well above average, above average, below average, and well below average usefulness.  I then tweaked their scores slightly, adding more granularity, and comparing badges with similar niches against each other.  Eventually, I ended up with a fairly balanced seven-tiered setup – Tiers F, E, D, C, B, A, and S.

Coincidentally, the higher and lower tiers ended up being clearly delineated by the middle tier (Tier C).  Most of the ones in that tier are decently generally usable badges with modestly useful effects, whereas ones above it (S/A/B) are much more frequently usable and/or heavily useful, and ones below it (D/E/F) are only usable in special cases (and often less useful even in those cases).

Of note, I generally didn’t consider stacking more badges than are naturally obtainable in the game (assuming only one of each available badge – and no Power Rushes – are bought from Pianta Parlor).

Without further ado, here are the tiers from bottom to top (badges in each tier ordered alphabetically):

F TIER

Generally badges I’d never dream of using seriously in a battle.

• Bump Attack is really placed this low largely out of spite; it’s undeniably of some use by the time you get it, but that’s really a symptom of going through the Pit leaving you unfortunately overleveled.  Furthermore, it makes a very unsatisfying reward while in the Pit, especially considering the difficulty jump between the floors 61-79 and floors 81-99.  Arguably, First Attack is just as usable anyway for much less BP (and badge setup churn), and is available from the shop about as early as it’d be usable.
• Chill Out is useful for dodging a very few enemies’ First Strikes (basically just Z-Yuxes and maybe Chain Chomps / Moon Clefts), and being in Dazzle’s shop with many useful badges as alternatives doesn’t do it any favors.
• Head Rattle is a big missed opportunity. Confusion is a very fun status, but it only targeting the frontmost grounded enemy limits its use considerably.  Not only are you taking out most of that enemy’s potential weaponization in dealing damage to it, but you’re then unable to use a number of moves afterward (including all Hammer moves save for Hammer Throw) without doing further damage to it.
• HP Drain and HP Drain P are massively nerfed from the former’s Paper Mario incarnation; dropping attack power by 1 in exchange for at most 1 HP restoration per turn is virtually never worth it when there’s so many other ways to mitigate damage.
• Peekaboo is a waste of both BP and Star Pieces.  Experienced players should have enemy HP values memorized and keep track of them, whereas inexperienced players already have access to Goombella’s Tattle for the exact same information.  Knowing Koops’ Shell Shield and Doopliss’s partners’ HP is the only unique benefit it provides, and that is questionably useful at best.
• Pity Flower is not the worst badge in concept, but unless you’re sponging a good deal of low-damage hits on Mario specifically, the 1 FP you’ll regain 30% of the time is not worth the HP you lose.  If the FP restoration was guaranteed, or the amount restored was tied to the damage taken, it could see a bit more use, but otherwise it’s filler for Pre-Hooktail Pit Bonetail setups at best.
• Return Postage is of minimal use for much the same reason, on top of being a somewhat lackluster reward for defeating the Pit.  At least it’s highly salable (not that that is likely to be a concern after a Pit run anyway).
• Simplifier‘s hit to Star Power regeneration far outweighs its use for easier Action Commands in general.  It is useful for scouting in the Pit of 100 Trials as it makes the run-away meter much easier to fill.  It’s also debatably useful for getting an extra bounce or two out of Power Bounce, but rarely does the difference between 8 and 9 hits equal night and day (and if you have good timing you’ll end up capping at the same point on bosses anyway).
• Tornado Jump‘s tornadoes do fixed damage (depending on the number of badges equipped), and doesn’t do enough ancillary damage to aerial enemies (which are rare in the first place) to be worth it over a battle item, or Fiery Jinx, or Earth Tremor, or what-have-you.  It is useful for clearing out Mini-*-Yuxes or Grodus Xes, but that’s about it.
  

  E TIER

Badges I find occasionally worth considering (or consistently in specific cases, but with relatively small effect).

• All of the P badges (Defend Plus P, Feeling Fine P, Happy Heart P, HP Plus P, P-Down D-Up P) are in this tier for pretty much the same reason; partners are generally tanky enough on the whole that it’s not worth the BP to make them safer. Not really a lot to say beyond that.
• Hammer Throw is outclassed in nearly every situation it could be potentially useful.  Swoopers are just as easily hit with Earth Tremor or Quake Hammer, as well as some partner attacks.  Spiky Parabuzzies have far too much DEF for it to be viable.  For virtually everything else, a standard Jump or Hammer will suffice.  It can theoretically have a niche against airborne Ruff or Ice Puffs, but I’ve yet to be in a situation where there wasn’t some reasonably inexpensive alternative.
• Lucky Start‘s effects are too short-lived, underwhelming and unpredictable to be worth the BP it costs.  It can be used to farm HP/FP by repeatedly running away if you really hate your “A” button (and Sweet Treat).  It’s particularly unusable if planning to use Danger strats, given the chance of getting the HP-Regen status.
• Refund doesn’t yield all that much salvage since it rounds 75% of the item cost down rather than up (as in Paper Mario).  Nonetheless it can be a decent fill-out badge if you have a spare BP and a packed inventory.
• Soft Stomp is reasonably useful (though not 100% reliable) on some very late-game bosses, but certainly not indispensable.

D TIER

Mostly badges that fairly typically get use, but aren’t all that useful, or are very useful in a very limited set of use cases.

• Charge and Charge P pretty much never see use outside of boss battles, and quickly fall on the massive-overkill side if they are used flippantly, essentially adding 2 damage per hit per charge in the turn their charge is expended for negligible FP (and you’ll have plenty of multi-hit options from early in the game).  Arguably these could be placed in a higher tier on those grounds, but they’re really nearly wholly unnecessary considering the existence of power / Danger badges, Power Punch, and Power Lift.
• Damage Dodge P is generally a bit more useful than the rest of the defensive partner badges, since it’s cheap to equip, available early, and combining it with the Defend command offers a lot of control over partner HP values.  It won’t do any good against some late-game bosses due to piercing attacks, though.
• First Attack isn’t quite as convenient as Spin Attack from Paper Mario, but it isn’t as inconvenient to get or equip as Bump Attack, and its effect is appreciated when backtracking through previous chapters’ areas, especially in completionist playthroughs.
• Hammerman‘s extra point of power comes at the expense of Mario’s generally more versatile and powerful Jump moveset.  Could be more useful depending on your preference of partner strategy and HP/FP consumption.
• Ice Power is completely invaluable against Lava Bubbles, useful against Grodus and Bowser in Chapter 8, and completely pointless otherwise.
• Ice Smash‘s status ailment can be devastating if it works, but the Frozen status doesn’t last too long without two copies, and there aren’t a lot of targets that it can hit and work reliably on that can’t be dealt with more easily in other ways. Clock Out is generally a far better choice for reliability, or Sleepy Stomp / Sheep for turn count.
• Money Money is too expensive and appears too late in the game to be as useful as it should be, but it can be useful for farming extra Pianta badges or recipe ingredients for a completionist playthrough nonetheless.
• Pretty Lucky P can be useful for mitigating partner health loss in the long term, or bolstering their evasion rate in Danger, but isn’t generally as useful as Heart Finder for the former purpose.  Can be helpful or actively harmful if it activates against moves that hit the front and back player-controlled actors in sequence if they’re in front, as Mario will be forced to perform the last defensive action taken (including a miss) even if it would be undesirable.
• Shrink Stomp can be fairly useful on the handful of bosses it works on, and perhaps particularly tough enemies, depending on the player’s preference. However, Mini-Egg has the same chance of working and gives you multiple attempts, so it’s almost always a better option.
• Super Appeal P can help make the most of an unspent or otherwise-unspendable partner turn, but only one copy appears naturally, and by the time you have it, the extra 0.25 SP units restored aren’t of much consequence.
• Unsimplifier is sometimes effective in boosting SP restoration, but makes a handful of moves, as well as Superguarding, a much less viable option due to the increased difficulty of their Action Commands.

C TIER

Also the middle tier, and now we’re definitely in the realm of generally useful badges.

• Defend Plus is a solid, obvious, but slightly BP-expensive way to mitigate damage. Particularly useful against fast multi-hit attacks like Magnus 2.0’s audience launcher and the like, and particularly useless against some very late-game bosses with mostly piercing attacks.
• Double Dip and Double Dip P are incredibly versatile tools, particularly for efficiently strategizing boss battles. Their FP cost (especially with Triple Dip) is fairly high, but that can be mostly mitigated by using at least one FP restoration item.  Late-game this can remove the need for Charge badges entirely, and earlier in the game it allows setup with Power Punch, Point Swap, Trial Stew, etc. in combination with recipe items to hit pretty much any stat breakdown you want.  Their biggest drawback is being located fairly late in the Pit and in the main story.
• Happy Heart is a decent filler badge for picking up occasional extra health in long battles (for instance, in Pre-Hooktail Pit), particularly if you prioritize heavier-hitting in-battle badges to the Finder badges.  Their effect is negligible in comparison to a Slow Shroom if you can afford the inventory space (and detour to the Deepdown Depot), though.
• P-Down D-Up is the only defensive badge (alongside its partner variant) to work against piercing attacks – including all stage effects and enemy items – which can be a game-changer against a number of enemies if you can afford the loss of attack power (or the badges to offset it).
• Piercing Blow is invaluable for Pre-Hooktail Pit runs against anything highly defensive and immune to fire (particularly Chain Chomps and Moon Clefts), or as a low-FP alternative even if Fire Drive does work.  It is generally overshadowed by Earth Tremor, Quake Hammer, or various item / partner attacks, and is equivalent to or worse than Power Smash for enemies with 2 or less defense.
• Power Jump and Power Smash are the go-to very-early game options for dealing a lot of damage in a single blow, and remain effective for that purpose during the entire game if you have the FP to spare for a double- or triple-stacked version of the move.  They fall behind the upgraded boots’ Jump moves quickly on low-DEF enemies, though.
• Spike Shield makes Spiky Parabuzzies tremendously easier to deal with, as well as allowing you to use Jump’s often superior firepower (and Jumpman) against lower-defense spiky foes.  It also makes Bristles approachable, though Quake Hammer is plenty effective at neutering them on its own.  Plus, there are spiky enemies it’s not terribly effective at dealing with (notably Clefts).
• Super Appeal is nice for quickly regenerating Star Power, particularly early in the game or when stacked.  It’s particularly nice for dealing with Dull Bones audience against Bonetail, as it guarantees a 1.00 SP fill-up in no more than two Mario Appeals (and often one Mario and one partner Appeal).
• Zap Tap completely blocks leeching attacks, and is a good failsafe in Pre-Hooktail Pit, particularly for direct attackers that would be otherwise impossible to kill if you’re inconsistent at Superguarding them.  It also has an edge-case use in allowing you to use contact attacks on enemies with the Electric status (not charged Puffs, though).

B TIER

Getting into the upper tiers. These badges are generally fairly potent and usable a sizable amount of the time.

• Close Call P is a great, very inexpensive safety net for Danger/Peril partner strats, activating nearly a third of the time (or over half the time with both copies).  It’s not the best way of ensuring their safety, though.
• Damage Dodge is another inexpensive defensive boost if you’re not opting to Superguard everything to oblivion, and a fantastic tool for targeting specific HP amounts with Mario.  Naturally, still it’s of no use against piercing attacks.
• Feeling Fine is far more useful on Mario than his partner, removing the need to worry about nearly all potent status effects (Freeze being the notable exception), and freeing you to go for Superguards on nearly anything.  It’s only available very late in the game, but it’d be most particularly useful on late-game bosses anyway.
• Flower Finder‘s 1-3 guaranteed FP drops per battle are fantastic at maintaining high FP across multiple field battles.
• Heart Finder is useful for the same reason, and is unique (as opposed to Happy Heart / Lucky Start) in being able to restore HP to Mario and his partner at once.  Being able to choose when and whether to pick up each individual drop also means being able to target specific HP values for Mario or specific partners, so you can keep partners in Danger or Peril as desired. Furthermore, it can uniquely (aside from items) bring 0-HP partners back to life in a Peril (or Danger) state.
• Item Hog is useful for the fairly substantial increased chance of held and random items dropping, but perhaps better for the ~5% chance of a Dried Shroom dropping from a battle with no enemy held items, allowing for resuscitating a 0-HP partner to Peril.
• Last Stand P is useful for targeting partner Peril, and more generally for allowing partners an additional attack or two after reaching Danger status.  Unlike Paper Mario‘s Last Stand, it (and the Mario variant) takes effect last in damage calculation and rounds up, so it won’t outright prevent damage unless none would have been taken before it factors in.
• Mega Rush is obviously incredibly potent in dealing damage, but it’s relatively risky to attempt maintaining it through enemy attacking turns.  It’s at its best if you can manipulate Mario’s HP and use it to finish out a boss battle.
• Lucky Day isn’t quite as cost-efficient as Pretty Lucky (and nowhere near as much as Close Call if in Danger), but without farming extra evasion badges it’s still a decent evasion boost.  Where it obviously shines is in Pre-Hooktail Pit runs, substantially lowering the average damage taken on floors 91-99 and against Bonetail.
• Multibounce is useful for grounding multiple winged enemies in a single turn, or dealing modest damage to all enemies in a set for minimal cost (or more than modest, with sufficiently increased attack power).  Also likely the go-to cheese badge for unambitious Danger Mario “strategists”.
• Quake Hammer, despite a none-too-flattering fixed damage and FP cost compared to Paper Mario‘s array of similar badges, is nonetheless especially effective against a set of high-DEF enemies, particularly if they’re flippable and/or immune to fire (Beetles, Bristles, Clefts, Chomps, Koopatrols).  It can still be decently potent in general with sufficiently increased attack power, but there are likely better options.
• Sleepy Stomp is unparalleled for taking a single enemy out of commission for an extended period of time.   It’s available very early in the game, and particularly indispensable for Pre-Hooktail Pit runs.  Typically loses out to Clock Out for general utility later in the game.

A TIER

These are generally incredibly potent at their best or very solid general choices.

• FP Drain is an incredibly underrated badge.  For a single BP and at the cost of a single point of attack power, it effectively acts as a Flower Saver that works on any of Mario’s attacks that deal damage, even 0- or 1-FP moves, so long as you can foot that extra point in advance.  The lost attack power is very easily recuperated in late-game, given the many cheap attack power-raising options that exist.
• Happy Flower is similar in appeal to Happy Heart, without the potential to screw up HP targeting if you’re planning to employ Danger / Peril strats.  Again, Gradual Syrup blows its inflow out of the water if you can afford to carry one (or two, if you have both Mario and his partner both use one).
• Last Stand is great as a last gasp to survive an additional attack or two, similarly to Last Stand P, though again not nearly as potent as Paper Mario‘s equivalent.
• P-Up D-Down is the cheapest way to get extra attack power without cutting off strategic attacking options (Jump / Hammer moves or intentional missed Action Commands), at the expense of taking an additional point of damage from everything, including piercing attacks.  Naturally, that downside can be mitigated with extra defensive setup, evasion, exploiting enemy AI, or Superguarding, or even be exploited to reach Danger/Peril earlier.
• Power Plus and Power Plus P are the most BP-hungry and blunt way to raise your attacking power, but are nonetheless effective and generally not too expensive to be worth considering.
• Power Rush P is doubtlessly useful, but is really the “poor man’s” Mega Rush P, as keeping partners in Danger is barely easier than keeping them in Peril for substantially lower gain.  The extra damage on top of Mega Rush P can be useful in theory, but is often unnecessary, especially with the existence of non-badge sources of increased attack power.
• Pretty Lucky is a fairly useful badge for filler (particularly with two equipped), or in conjunction with stronger evasion badges.  It often isn’t advantageous in cases where you need to deal damage to direct attackers with counterattacks or Zap Tap, as Sweet Treat should often be sufficient on its own to mitigate the lost health from failed Superguards if your partner Appeals.
• Quick Change‘s primary function is to preserve a partner’s status (be that Peril health, or Charge / other positive status effects) indefinitely by switching to another partner before the enemies’ attacking turns.  Aside from that, there aren’t many compelling use cases for it, and it has a steep BP and Star Piece cost to boot.  It’s not all that necessary for partner Peril strats in Pre-Hooktail runs either, unless you have extra restrictions (such as a speedrun or low-Mario health run).

And finally,

S TIER

The best of the best; unmatched in utility and/or viability.

• All or Nothing is a cheaper Power Plus with virtually no downside; you lose only the ability to strategically omit Action Commands (or the damage from the last, unsuccessful jump in a Power Bounce).  Obviously it is not a good choice if you’re not solid on Mario’s Action Commands in general, but most of them are fairly straightforward and easy to land, even with Unsimplifiers equipped.  Its only flaw is its only being obtainable terribly late in the game.
• Close Call is a cheap and effective safety net when Mario is in Danger, particularly when stacked in duplicate (~55.1% misses), or in duplicate with the other evasion badges (upward of 70% misses).  Adding Last Stand(s) to the mix can allow near impenetrablility of Mario’s defenses in Danger, for a measly handful of BP.  Both natural copies are available very early in the game, and they can be farmed (albeit not terribly effectively) basically at the start of the game as well, if you want to truly overkill your evasion.
• Fire Drive is nearly always a viable attacking option for non-aerial enemies, and is absolutely indispensable (daresay, the linchpin) in Pre-Hooktail Pit runs.  In addition to doing massive damage early-game, it can inflict the burn status on pretty much anything that isn’t immune to fire attacks, dealing ancillary damage and completely confounding lingering Dark / Elite Wizzerds.  Stacking both copies is virtually never a good idea, by comparison; the boost in attack power is negligible for the whopping 10 FP and 3 extra BP cost.
• Flower Saver is the single most generally useful badge in the game, for my money; it is available incredibly early in the game via Dazzle, and there is virtually no time (especially after obtaining Spin Jump) that bringing Mario’s 2-FP moves to 1-FP is not useful.  The second badge can be useful as well, bringing Quake Hammer and Power Bounce to 1 FP and Fire Drive to a modest 3.  Combined with FP Drain and/or Gradual Syrup(s), nearly all of Mario’s moveset, natural and badge alike, can become virtually free.
• Flower Saver P isn’t quite as mind-blowingly useful as the Mario variant, but it brings down the cost of several key moves (notably Power Shell, and all of Goombella’s and Yoshi’s moves) substantially, especially if worn in duplicate.
• Jumpman‘s extra point of power comes at the cost of his defense-piercing, harder-hitting, and multi-target hammer moveset, but given most Jump moves have at least 2 hits, it raises damage dealt by at least 2 points for a negligible BP cost.  Star Powers and partner attacks can fill the void of multi-target and piercing attacks, and additional power-increasing badges (or Power Jump) fill the more-damaging niche of the Hammer reasonably well.
• Mega Rush P makes Pre-Hooktail Pit runs considerably less daunting (though it is by no means essential).  Using it with Heart Finder works wonders at maintaining partner Peril across multiple field battles, whereas Quick Change often makes it near-impossible to lose the status within a single battle.  In conjunction with Goombella’s Multibonk or most of Yoshi’s attacking moves, it is arguably the most broken badge in the game.
• P-Up D-Down P, in particular if stacked via Elite X-Naut farming, is essentially equivalent to having permanent Danger / Peril partner strats.  Even a single copy paired with Power Plus(es) P turns partners into destructive powerhouses, at a relatively nominal cost to defense for an Ultra-Ranked partner.
• Power Bounce can, as expected, take advantage of Mario’s slightly wider array of attack-power raising options than partners to blast through pretty much any foe.  It’s also (arguably more sportingly) useful for tackling annoying enemies in Pre-Hooktail Runs as soon as possible, often in conjunction (for high-HP enemies) or parallel (for multiple moderate-HP enemies) with Goombella’s Multibonk.
• Power Rush is a cheap and a safe option (compared to Mega Rush) to maintain raised damage output in repetitive field battles, as well as an easier and safer damage boost than Mega Rush to target from initially max HP in a boss battle.  Obviously stacking multiple copies from the Pianta Parlor, and/or using it in conjunction with perma-5 HP Mario, makes it the most boringly broken badge in the game, but it doesn’t deserve all the stigma it gets for that if it’s used creatively.

With that, this is my complete Paper Mario: TTYD badge tier list as it currently stands, in a single image for brevity:

Feel free to discuss your opinions on the matter (and no doubt lambast my disservice to Quick Change such-and-such badge) in the Glitz Pit Discord server or my own personal Discord server, Club Jdaster64; I’m generally around both frequently, and there’s plenty of Paper Mario series discussion to go around.  If you’re interested in this sort of qualitative discussion of Paper Mario mechanics in addition to my usual technical fare, Kappy (organizer of the Glitz Pit and Paper Mario wizard) has a blog for this sort of stuff that you can check out here.

Of course, if you’d like my take specifically on more stuff like this, feel free to shout out suggestions for similar topics in the comments here, or on my Discord server.  Might be worth pooling together some of the top challenge runners’ minds to form a “Paper Mario Back Room” badge tier list in the future, or something; who knows.