HMs are a genuinely brilliant design choice. Most JRPGs have a very clear line between exploration and combat, and this was especially true at the time Pokémon was originally released. But Pokémon was serious about adding some fun exploration elements to the overworld. It’s kind of like a minimal version of a Metroid or Zelda game in that respect. Like those games, you explore the world, you get tools that let you interact with the world in new ways, and then you explore even more. It’s fun!

The very clever thing is that, through HMs, you teach your Pokémon the skills that allow you to explore the world, rather than learning them yourself. And the skills are combat abilities, so your exploration tools interact with combat. It emphasizes that your Pokémon are more than just things you send out to battle – they’re companions that explore the world together with you. What a nice idea!

Mechanically, this also plays to Pokémon’s strengths. The games are often excellent at getting the player to consider making changes to their party. HMs play into that, because not every Pokémon can learn every HM. It’s easy to go through half the game without having a Flying type Pokémon on your team without even noticing it – but as soon as you get the extremely useful Fly HM, it becomes very attractive to give some flier a permanent spot in your party.

That’s why everyone loves HMs, and they remain a staple of the Pokémon games to this day.

Just kidding! HMs caused some serious problems, and fans’ complaints about HMs grew steadily louder over time. By looking at how HMs changed over time, it’s evident that Game Freak either capitulated to or agreed with the fans. At first, new HMs were being added every generation, but the compounding problems had clearly reached a critical point sometime before the release of Generation V (Pokémon Black and White). Gen V was not only the first generation to not add a new HM, it was also the first generation remove an HM without replacing it with another – and it removed two! Generation VI (Pokémon X and Y) was the last generation to feature HMs at all.

We’re now up to Generation IX, and HMs have been a thing of the past for three whole generations. If HMs have been non-existent for fully a third of the Pokémon franchise’s life, could it be that they were never any good to begin with?

To answer that, let’s start by summarizing the major problems of HMs:

HMs take up move slots. As I mentioned before, HMs take up a slot of your Pokémon’s combat moves. But, with only four slots available for each Pokémon, you don’t have many to spare. A full team of 6 Pokémon has only 24 total move slots, and HMs eat into that.

HMs are hard to unlearn. Further, unlike every other move, HM moves cannot be replaced by a new move move your Pokémon learns. This is a necessary evil to prevent a player from getting stuck by entering an area using an HM move, unlearning that move, and then being unable to get back. HM moves can only be removed at a specific location in each game.

Most HMs are terrible as combat moves. Here are all the HMs ever made:

  1. Cut
  2. Fly
  3. Surf
  4. Strength
  5. Flash
  6. Whirlpool
  7. Waterfall
  8. Rock Smash
  9. Defog
  10. Rock Climb

Most of these moves were terrible when they were introduced. There are some exceptions – Surf is a great water move, even to this day. Fly was great in the early generations. Waterfall was a worse version of Surf when it was introduced1 but got buffed in Generation IV (Pokémon Diamond and Pearl). Every other HM is a bad combat option in its respective game, even at the point in the game you get it.

Note that each of these problems compound on each other. If HMs take up a smaller share of move slots, it’s not as big a deal if they’re bad in combat, because you don’t give up as much to keep them around. If HMs are easy to unlearn, it’s not a big deal if they take up a move slot – you can easily replace that slot with a better move later. So fixing or minimizing even one of these problems lessens all the issues HMs have. So, did these problems get worse over time, or were they always bad?

Let’s start with the easy one: HMs were hardest to unlearn in Generation I (Pokémon Red and Blue), because it is the only generation in which you simply couldn’t unlearn HM moves. Score one for HMs getting better over time. But, even in generations with Move Deleters, the Move Deleters are always found late in the game. At best, you have to wait until badge #6 to unlearn an HM. So, even generations that mitigate this problem still force you to commit to any taught HMs for the majority of the game. This specific problem got better over time, not worse – but it didn’t get dramatically better.

What about the burden of move slots? Gen I had 5 total HMs. To have all HMs available at all times (which is not actually necessary), you would need to tie up 5 out of 24, or a little over one-fifth, of your total available move slots. But by Gen IV , there were 8 HMs, requiring a full third of your party’s combat slots! So this definitely got worse, and surely played a part in the decision to reduce the number of HMs in Gen V.

As for the relative power of HMs, this is where things get a little more interesting. Let’s start by taking a look at Gen I’s HMs, where this problem was the least bad. We have:

  1. Cut
  2. Fly
  3. Surf
  4. Strength
  5. Flash

2 of these (Surf and Fly) are contenders for best moves of their respective type. The other 3 are bad options even by the time you get them. Future generations had a worse ratio than this. This is as good as it gets!

But there’s more to it than that. This problem is the least bad in Gen I for more reasons than just the ratio of good HM moves to bad ones. The other mitigating factor is that Gen I move slots were less valuable, because there were fewer good moves.

Say, for example, you want to use Flash to get through Rock Tunnel. You’re gonna have to sacrifice a move slot for a basically worthless move. But what Pokémon can afford to have only three move slots for the rest of the game? Well, basically everyone. Let’s look at Pikachu’s learn set, for example:

Pikachu's Unimpressive Learnset

At the end of this little guy’s career, he will have:

  1. An electric attacking move (Thunder Bolt by TM or Thunder otherwise)
  2. Thunder Wave
  3. A Normal type attacking move (either Quick Attack or Swift, or something from a TM)
  4. Uh…. Agility, even though you’re already fast? Growl??

It’s hard to find a good use for that fourth move slot. That means the opportunity cost of sealing one slot away forever with basically an empty space is quite low.

It may seem unfair to pick an Electric type, since they have famously shallow move pools. Okay, let’s pick something more charitable, then. How about Charizard2?

Charizard's Red and Blue Learnset

Throughout all of Charizard’s life, it will essentially learn two varieties of move: Normal type attacking moves, and Fire type attacking moves. You need to dip into TMs to have any use at all for half of Charizard’s move slots! Adding in TM options still only gives us a few more viable choices for our beloved starter, bringing the total number of useful moves by endgame to around 8. But TMs distort our number of options a bit – since most TMs cannot be replenished, any TM we use on Charizard is one we can’t use anywhere else. So if there are, say, a total of 10 good TMs, they can’t use up more than 10 of our 24 slots – the other 14 options have to come from each individual Pokémon’s level up learn set.

Now, 8 options is still more than 4, so it’s definitely possible to fill out all four move slots! But in company like this, we’re not exactly missing out on much if we use a slot on Strength. Frankly, if we naively threw Cut onto Charmeleon in the early game, it wouldn’t be that great of a loss. We just don’t have that many ways to get all 4 move slots to put in work for us, so throwing one away on even 5 of your 6 Pokémon is just not that great a cost.

This is in stark contrast to future generations. For example, here’s Charizard’s Diamond and Pearl learn set:

Charzard's Diamond and Pearl Learnset

Charizard learns attacking moves from 5 different types3. You’ll have to choose one type you don’t want to have coverage for, because you only have four move slots! In an environment where each move slot has so many good options, even 5 HMs would have felt burdensome. But 8 HMs? That feels downright oppressive!

As the generations went on, Pokémon combat got better and better. Part of the reason for this is that there are many more moves, and many more good moves. By as early as Gen IV, not only was it easy to fill out all 4 of any given Pokémon’s move slots, it was consistently difficult to choose which moves to keep!

So, of the three problems we outlined, two were much less significant in Gen I, and the remaining one was only slightly better in future generations. It’s pretty conclusive that, although HMs always suffered from the same problems, HMs were much less problematic when they were introduced than when they were removed.

Were HMs better in earlier generations? Yes. But they were better because the combat was worse. Pokémon had fewer moves that were worth using. As the gameplay developed, the combat became more interesting, and part of that was developing more moves that interacted with each other in novel ways. While Game Freak didn’t do themselves any favors by making most HMs so bad in combat4, the competition for move slots got more and more fierce over the generations. As such, bad HM moves hurt more and more, until the combat was so good that we couldn’t afford to have HMs in the game anymore!

HMs should be remembered as the good idea they were, but one which the series eventually grew past. Because the types of problems HMs posed were always the same, it’s easy to misremember those problems as always being equally severe. In fact, HMs were a great idea, and they were implemented well in the context of the first game. It’s good HMs are gone, but that doesn’t mean they were a mistake to begin with.

  1. I just learned while writing this that Waterfall, while introduced as an HM in Gen II, was actually already a move in Gen I! It was the signature move of the Goldeen evolutionary line. Who knew! Well, people who had a Goldeen in their party knew, I guess. ↩

  2. Please note that, while I knew Pikachu’s move pool was very shallow before I started writing this, I didn’t cherry pick Charizard – I didn’t even know his Gen I move pool before I started writing this. I just thought, “Who’s likely to have a deep move pool in Gen I?” and looked up whoever I thought of. I was honestly surprised to only see two types in all of Charizard’s level up learn set, but I shouldn’t have been. This is just what Gen I move pools look like. ↩

  3. Including Flying! Charizard, the Fire/Flying type, didn’t have a Flying type move at all in Gen 1 – He couldn’t even learn Fly! ↩

  4. I suspect they feared players would teach good HMs to everyone on the team, making each Pokémon feel too similar to one another in combat. Still, I’m not sure making most HMs so bad was the best way to avoid that. ↩