So in terms of balance, the first step is to establish a baseline of "cost in" to "power out" that elements of your game will broadly adhere to. This is generally referred to as the Power Curve. What's noteworthy is that the term "curve" and not "line" is used to describe it, because the Power Curve is non-linear, in fact it can be outright exponential. This is a function of Opportunity Cost.
Opportunity Cost is best boiled down to the question of "What else could I take?". This isn't exclusively limited to points though. In Malifaux, each category of model has a set limit on the number of upgrades it can equip, so every upgrade presents not only the opportunity cost of the amount of Soul Stones required, but also of all the other upgrades that it could be and that you're no longer taking. In Warmachine, you only have 1 Warcaster and they lead your force. In that regard, even though there is a cost element in terms of their Battelgroup Points, this is almost never a consideration and always gives way to the spells, abilities and feat a caster can offer to your force. We can see opportunity cost in action when looking at all the really cheap or free Elite Pilot Talents in X-Wing 1.0. Trick Shot costs 0 points, so why wouldn't you equip it? Well, if you really need Predator or Expertise for the offensive modification, Push the Limit for the extra action etc and you only have 1 Elite Pilot Talent slot like most ships, the opportunity cost of trick shot is quite high because even though it's free, your losing a critical element of your build. There's also a somewhat intangible opportunity cost in that, all elements being equal, lots of small things is generally superior to 1 big thing. An area buff multiplies faster across the little things, they can be in more places at once, losing 1 is less of a problem and they're less vulnerable to failures of whatever random element your game has i.e. a bad round of dice for 1 of your 30 infantrymen is no big deal, but for a giant robot that's a third of your points, that might be the game.
Opportunity cost is the reason why game elements at the low end of the curve are typically less efficient and elements at the high end of the curve are more so. There's not much competing with a basic Academy Pilot in a TIE Fighter in X-Wing 2.0 at 23pts, but Dash Rendar in a YT-2400 at 100pts is not just competing with individual elements, but combinations of those elements that can result in a more potent whole and give substantially more value. As a thought experiment, Imagine what a ship in X-Wing 2.0 would look like if it cost 1pt out of the 200pt standard list. Could you even give this ship and attack die or would that be too much? What about a ship that costs 200pts with no upgrades? How insane would that ship have to look like in terms of offense and defense to even be remotely playable? For examples in other types of games, look through the 10 mana cards in Hearthstone that have been playable in the past (things like Deathwing and C'thun) and how high impact they were. They don't quite say "win target game", but they do say "find an answer for me soon or lose" and if you're spending all of your end game mana for 1 card, that's really what they need to say. While many players believe the biggest problems occur at the high end of this curve, this belief is generally incorrect. Far more design teams have struggled with the low end of the spectrum, in essence, not being bad enough for their cost making fielding large numbers of these elements the optimal play.
For an interesting analysis, let's look at the very first wave of X-Wing 1.0. For all the praise the early waves of X-Wing 1.0 will receive from fans of the game, it had a major flaw that the game struggled with for it's entire lifespan. The cheapest ship in wave 1 of 1.0 was the TIE Fighter Academy Pilot at 12 and it was generally the most efficient in terms of cost -> power. Y-Wings were the second cheapest at 18pts and were actually pretty weak in terms of cost -> power. The X-Wing and TIE Advanced were both 21 and the X-Wing was decent enough in cost -> Power, but the TIE Advanced was absolutely terrible (probably overcosted by 3-4). You can probably already see the issue i.e. the highest point of efficiency is at the lowest point of cost and the highest point of cost is the lowest point of efficiency. Basically, your entire power curve is inverted. As to exactly why this happened, nobody 100% knows but it appears linked to the use of secondary weapons. The TIE Fighter had none, the X-Wing only had torpedoes that present very little advantage over it's primary, but both the TIE Advanced and Y-Wing, despite having a 2 attack primary, had access to 4 attack secondaries (Concussion Missiles and Proton Torpedoes) on top of the Y-Wing also having a turret slot. If you've ever wondered why TIE Fighters dominated the competitive scene so much in early X-Wing 1.0, this is why, they were just mathematically better.
So what options did the X-Wing design team have when they started arriving at the realization that the value they anticipated these secondary weapons having wasn't coming together on the tabletops and their game was being dominated by 1 ship?
1. Leave it as is. This wasn't really an option. To put it simply, if you go through 6-8 waves or so and there's still not much of a reason to break out anything but a dirt cheap wave 1 ship, players are going to start losing interest. It's cool seeing a TIE swarm once, it's less cool when you see them every game.
2. Power level errata. It's important to note now that a power level errata (an errata that addresses the power level of an element in the game rather than just correcting an error or something game breaking/not working as intended) was not commonly used by designers at the time. As tabletop games jump more into the digital age where these kind of updates can be easily pushed out to everyone through apps or other media, it's becoming super common, but 5 or so years ago, power level erratas were incredibly uncommon. The devs were probably also understandably hesitant about errating elements that would be the first thing players would see and use in the core set.
3. Right the curve. This was the most viable option at the time and for the game longevity, but where do you go? Well, really the only option without a nerf is adjusting off of the TIE Fighter, but that's going to lead to another problem. You see, the Wave 1 and 2 ships, where all of the iconic stuff that you want to see in a Star Wars game are located, would be at the bottom of this curve. Adjusting up off of the TIE Fighter would eventually completely kill the first few waves and push all of the iconic ships out of the game. This was probably seen as a lesser evil because, for all intents and purposes, a lot of these ships were already out of the game (both the Y-Wing and the TIE Advanced were basically DoA and the X-Wing was only fairing a little better).
This is where we get to talk about Power Creep. Power Creep is where the average ratio of expected cost -> power changes over time. Note that this is an average. Outliers will still exist, but don't necessarily point to power creep. Elements that are unplayable being superseded also doesn't point to power creep. Was the JumpMaster 5000 in X-Wing 1.0 and example of power creep? I'd argue "no" because it won worlds twice in a row despite a number of waves dropping in between both wins. If there was a substantial case of power creep involved, the intervening waves would've released ships that at least had comparable enough ratios of cost to power present an obstacle to the JumpMaster's dominance. Instead the JumpMaster had to be nerfed multiple times and still occasionally found it's way into competitive builds. This indicates the ship was an outlier to the normal power curve and not indicative of a general direction, up or down. Another example was the X-Wing 1.0 Harpoon Missiles vs Concussion Missiles. Harpoons were just better, but was it power creep? Simple question, how many competitive lists since the release of Concussion Missiles in Wave `1 used them? The answer was 0. They were not playable in any format or with any ship and there was no reason to believe they would be even at the end of 1.0 X-Wing. You can't compare something to an element that never sees play because that element is below the curve, so simply establishing that a new element is better doesn't indicate broader movement of the curve. Â
Now, Fantasy Flight Games did try to address this problem by releasing a variety of fixes for older and more iconic ships, but this highlighted major problem that both X-Wing 1.0 and Warmachine/Hordes struggled with. Both games had a model that generally revolved around a rotating schedule of symmetrical resales across all factions. Quick question for the audience: If this is your model, how do you go back to address problems with your power curve? Both Fantasy Flight and Privateer Press (and also Wyrd if you look into Malifaux and some of their fixes) tried a similar solutions i.e. releasing elements designed to make these less desirable components better. For X-wing, it was 0 or negative cost upgrades or upgrades that were specific to that platform and gave it a better role. For Warmachine, it was releasing specific buff elements (Unit Attachments, Weapon Attachements, buffing solos, casters). The problem is that this kind of approach has a bit of shelf life. You can only do it so long before it starts to wear thin. For example, my X-Wing list at the end of 1.0 had 4 X-Wings and a Z-95 with a total of over 20 upgrades across all 5 ships that net totaled 0pts in cost. After a while, this system gets overly clunky and bloated and it's still not quite going to produce the desired effect because it's not very well suited to dealing with the elements of your game that are overcosted, but not quite broken.
This was more than likely the reason that both Privateer Press and Fantasy Flight are moving to a digital model. Privateer Press has the CID system where each faction gets it turn with some new toys while older ones are revisited (essentially a digital version of the Codex system from Warhammer that incorporates community feedback). Fantasy Flight has put all point costs and upgrade slots in their app and it has received updates every 6 months or so. In short, both companies seemed to have realized the major flaw with trying to right your power curve with releases alone.
A lot of what we've talked about so far has been unintentional power creep i.e. the devs missed something or under/overestimated the value of certain elements. But what about intentional power creep? Well, as you can probably guess by now, I play a little bit of everything. The field where you see the most egregious and definitely intentional power creep is in online, free to play, CCGs. Some of them are pretty solid and have interesting mechanics, but far too many are so bad that they basically become unplayable beyond a certain level. Malicious power creep is less prevalent in miniatures games just because the cost and infrastructure required to put out a miniatures game makes it less desirable if you're just looking to do a smash and grab and don't care about your customers. That being said, there is a situation where you might intentionally use power creep, but not to just take your player's money. That is, if your game is just too bloated with poorly designed or poorly costed elements to continue and your company isn't large enough to adsorb the cost of an edition change, you could just use power creep to render the problematic elements obsolete and they would slowly filter out of the game because even the best and most problematic elements would eventually lose out in efficiency to the new curve. This is not a good idea or ideal on any level, but it may be the only option available.
It's also easy to confuse non-design related concerns for power creep. If the company is producing a new big model and it seems too powerful, stop to think how much it cost to produce that new model, package, ship it, distribute it and how much it costs to maintain an inventory of it. It's most decidedly not cheap especially if it's dealing with a scale or production techniques that are new to the company in question. In that regards, it's common for the new big models to be better than their in game cost might lead you to believe, but this isn't harmful to the game unless they're just so good and so much better for their cost that they dominate the competitive and casual scene. I don't think it's a coincidence that a lot of the CID cycles from Privateer Press have involved buffing underused Colossals or Gargantuans. Even with that effort, most of these large and impressive models don't see much table time (maybe a quarter of them see regular play and that might be generous). That's not to say this isn't beneficial to the player, the better these big models perform, the easier it is for the company to justify more of these large and impressive center pieces. I know from my experience that I get a lot more spectators who know nothing about the game watching with interest when I break out one of these big models, so it also serves to generate more interest in the game. That's not to say it's never exploitative or that you shouldn't be wary of new releases that are just stronger than existing options, but, from my experience, players are far too quick to sound the alarm of "cheap cash grab" or "power creep" against something relatively benign that doesn't dramatically alter the overall curve of the game.
The power curve might be the most cited, but least understood by players, aspect of game design. The next time you hear "power creep" look closely at what is being compared and ask yourself the key question "Is this element replacing an existing element that sees play or of demonstrably higher quality than existing elements in other factions that see play?" before jumping to a conclusion.