What's the best way to add difficulty? Pros & Cons

[jojobear13]

In upping the difficulty of my own hack, I have learned that that there is no one "best" method. A good challenge is a carefully baked cake that uses a proper mix of all the ingredients you have identified.

1. raising the level for every trainer pokemon for every battle encounter.
2. boosting their evs/ivs.

These two are effectively the same thing. At the end of the day you are just increasing the stats of the enemy pokemon. The important thing to consider is enemy roster stats in comparison to the player's expected roster stats at any particular point in the game.

Consider trainers in a hack's first mini dungeon (viridian forest or what have you). The player team is going to be early-game basic pokemon around level 5 to 8 with very, very few EVs. If you make enemy trainers there have level 9 to 11 mons with perfect IVs and EVs boosting them to effectively level 12 to 15, then the player is going to have a bad time. It just becomes blatantly unfair to be constantly outmatched trainer after trainer for no reason other than the game's math. The answer is to grind some levels, but grinding levels is merely the illusion of challenge.

3. programing different move sets.

This is similar to increasing levels. If you just give better moves to enemy trainers that far outstrip what the player currently has access to then the challenge involved is just to, once again, grind out some more levels.

4. adding more pokemon to each trainer.

Okay, so you've carefully balanced every trainer so that their stats and moves are comparable to what the player team is expected to be when encountered. You've even added mons to their teams so that they have the about the same number as the player when everything averages out. Now you have another problem. Those routes and caves full of trainers? They become a massive slog!

Imagine every lass, youngster, and hiker fight being one where your teams suffers a KO or near-KO and the rest of your roster is hurting for HP. The player, whilst tearing their hair out, is going to be running back to the pokemon center unless you make healing items much, much more plentiful. That is why vanilla games keep regular trainers mostly at 2 or 3 pokemon. It's just enough to wear the player down little by little with each trainer fight. Ideally you want to have the player come out of that area barely hanging on but victorious. You want a gauntlet, not a death march.

Adding pokemon to enemy rosters can also work against difficulty. Each new trainer mon is a new source of boosted exp. Adding one or two mons to every trainer in the game can easily cause the player's team to over-level beyond expectation. Note that this can be done intentionally in order to rectify dead zones in the game where the player seems under-leveled and has to grind low-level wild mons (*ahem* looking at you Johto region).

5. changing battle type. (like more double battles or rotation battles)

Less of a challenge and more of way to break up stagnation. Do these in long stretches of trainer battles to keep things from getting boring. But don't overdo it. Moderation helps keep it special and interesting.

6. setting level caps through exp gain adjustment.

This is trying to fix a flaw in the game by forcing the player to play a certain way. I'm just going to out and say it. This is bad game design, it should be avoided, and is really just unacceptable.

I'm going to call out a specific example here with Gen 8 since it really illustrates my point. Sword and Shield flat out prevent you from capturing pokemon that are too high a level unless you have the right badge. The reason is because high level pokemon can spawn in the open wild areas, and it would break the game's progression if you could catch them. But that is a poor excuse for bad game design. A proper solution would be to have different encounter tables for each wild area based on how many badges the player has. This way the player never encounters wild pokemon that would break progression while also giving the player a reason to return at various points in the game to see what has changed.

7. or changes to the AI.

Pokemon games generally have three "groups" of AI opponents.

• The first is the know-nothing trainers that choose moves totally at random. Usually given to youngsters since they are supposed to be very young children, and to cueballs since they are rather unintelligent.
• The second is amateur trainers who avoid making dumb decisions (using sleep powder on a poisoned pokemon, trying to double-confuse an opponent, etc) but also don't use strategy like STAB, type-countering, or making proper use of movesets. These are your trainers like bugcatchers, rockets, bikers, and blackbelts.
• The third is professional trainers who know what they are doing. They do everything from the second group but also make use of the type chart and movesets. These are your leaders, ace trainers, jr trainers, and rivals as well as trainers like gentlemen and sailors.

I'll say it again. You don't want to turn your routes and caves into frustrating slogs. You need to have a good mix of trainer AI to prevent every trainer from being the same super-elite every 20 steps. It gets samey, dilutes the feeling of player progression, and takes away from the excitement of fighting the gym leaders. Use trainers from the first group as a warm-up; let the player feel their growing strength by smashing a few nobodies into the dirt. Then up the ante with some trainers from the second group, and intersperse some trainers from the third group throughout in order to keep players on their toes. Finish off with a one or two tough fights to act as a mini-boss (the SS Anne rival battle is a perfect example of this).


I'll close by stating this. Making a proper challenging pokemon game isn't about taking one aspect (like raising levels) and applying it across the board. Pokemon game difficulty is actually carefully calibrated. Albeit the whole issue is that it is always calibrated to be difficult for an easily-distracted 9-year-old. Making them more challenging for teens and adults means going back, understanding all the subsystems, and re-calibrating the whole experience for an older audience.

 

[Meister_anon~Master_o f_None]

II'll close by stating this. Making a proper challenging pokemon game isn't about taking one aspect (like raising levels) and applying it across the board. Pokemon game difficulty is actually carefully calibrated. Albeit the whole issue is that it is always calibrated to be difficult for an easily-distracted 9-year-old. Making them more challenging for teens and adults means going back, understanding all the subsystems, and re-calibrating the whole experience for an older audience.

*clapping*
very well said, and I like that you took both sides of the argument in some places just to flesh it out.
Thanks I'll keep these in mind.

 

[Salvation714]

From my own experience, I know boosting the enemy Pokemon to have perfect stats is too over powered.
All you need is higher level pokemon with a balance of moves.
ex.
-lvl 25 geodude with magnitude
not
- lvl 25 geodude with earthquake

Also make gyms have a full team.
give trainers at least 3-4 or 1-2 strong pokemon

Level caps are pointless.
-just go with open world.

Battle AI?
just give the trainers potions.

Playing Pokemon Radical Red shows us how broken full ev stats can be.

 

[Corveone]

Consider trainers in a hack's first mini dungeon (viridian forest or what have you). The player team is going to be early-game basic pokemon around level 5 to 8 with very, very few EVs. If you make enemy trainers there have level 9 to 11 mons with perfect IVs and EVs boosting them to effectively level 12 to 15, then the player is going to have a bad time. It just becomes blatantly unfair to be constantly outmatched trainer after trainer for no reason other than the game's math. The answer is to grind some levels, but grinding levels is merely the illusion of challenge.

Well, that's why level cap can be a good thing to prevent players from bypassing a challenging battle by just gaining more levels. But yeah, it has to be done and balanced properly as you said, if a game doesn't allow overlevelling, then it should be fair and reasonable so there aren't absurd gaps between what the opposing trainers have access to and what the player has access to at that given point in the game.

Otherwise the player won't grind levels, but might have to spend a lot of time grinding EVs, or catch and train Pokémon that can evolve and learn good moves before the next important battle specifically, so it's like solving a problem by creating another.

Level caps are pointless.
-just go with open world.

Battle AI?
just give the trainers potions.

How would an open world structure help the difficulty though?
If it's open world, that probably means you could challenge different Gyms in any order you want, meaning that if you have trouble with a certain hard leader you could just go do other gyms, routes, caves, etc., then come back 10 levels later and destroy it. An open world Pokémon game would be far more complicated to balance than a linear one.

And just giving more potions won't improve the AI in any way. Even if you give an AI trainer 20 full restores, that would just make battles drag for longer and become incredibly annoying, the game would basically revolve around outsmarting the AI and forcing it to waste all their consumables on a Pokémon you can take in two hits so it doesn't use them on its more bulky Pokémon.

 

[Meister_anon~Master_o f_None]

boosting the enemy Pokemon to have perfect stats is too over powered..

I mean if you keep levels in mind its not that bad, its all a linear scale.

and its not always hard to fight, you can have a bad allocation, like boosting hp without fixing speed or defense, the pokemon will still get one-shot.

Plus like jojobear said, boosted evs, just raises a pokemons effective level up about 3 or 4 levels. but only for that stat.


Now if radical red did both of those things, lol then they intentionally set you up to face pokemon that were probably about
10 effective levels above you at all times. It just tricked you into thinking they were easier because the number wasn't as high. I assume.

Because of how the formula works, perfect evs don't even create a significant difference (by itself) until about lvl 50.

 

[Salvation714]

^Thats not true. If you raise a pokemon stats it will cause unbalance to the point of being so broken. No pokemon is meant to have perfect stats just for the sake of difficulty.
ex. A Drowzee that out-speeds your Raichu and spams Psychic.
Adding in potions is useful, but obviously changing the enemy's move set would give a new experience. I'm saying that adding in potions is the most you need for difficulty without programing new AI.

I stand by my statement with open worlds. Putting in level cap areas is too linear and boring AND REPETITIVE. The only solution is adding in road blocks like Nintendo does with HM 1 trees and NPCs.

 

[wally-217]

I think most of these points have been touched upon already but I'm a big advocate of Flow Theory and Dynamic Difficulty Adjustment. One of the driving forces behind my rom hack is that I want experiment with how you could implement DDA into Pokemon (And hopefully, later, Open Worlds). One thing I really liked about the wild area is that it gave me control over my flow. So instead of grinding on low-level pokemon that you can OHKO, I could spend 30-60 minutes trying to take down a level 50 Rhydon with my level 25 Grookey. The amount of XP I recieved was actually very similar to what I got from grinding over the same period but it was so much more satisfying because I created that challenge and landing that critical hit when my pokemon was about to die creates good player narrative.

For me, I have always been the type to overlevel my starter and if the gym leader is too difficult, I just grind until I have a pokemon capable of sweeping their whole team. I started watching VGC last year and the games really seem to hold back on teaching the player the actual battle mechanics. I imagine this is done because the player is likely 7 years old but simply conditioning the player to think tactically would turn the "difficulty" from 'who is the higher level?' to 'oh this gym leader uses a physical team, maybe I should prioritise burn'. I would like to see Gyms used in this way, with the Elite Four being the real challenge. Things like soft level caps work really well here because they show the player that they can't just power their way through.

One thing I would like to try that's a bit more experimental is using the gym trainers to hone the difficulty for the gym leader. For instance, if the player defeats the first trainer and at the end of the battle they still have 6 pokemon with full hp, you could set a variable to increase the AI difficulty and/or buff the opponent's damage output. If the player gets to the last trainer and gets fainted, you could set the difficulty variable to be lower. If the Gym leader's first pokemon gets OHKO'd in the first turn, you could increase the level of their second pokemon by a few, etc. I am hoping that I can incentivize losing enough that they don't just soft reset every time they white out (maybe by increasing prize money with every attempt) as then you can track their attempts and give them positive boosts too! Disabling in-battle items would be essential for this to work though.

In my hack, the gyms are not tied to the story and HMs aren't tied to gym badges so there's no risk of stalling the entire game if a gym is too difficult. In this instance, soft level caps could be set based on gym badges (if required), with each gym having 3 difficulty levels (so with one gym badge, they might use 3 unevolved pokemon and at 7 gym badges, they might use 6 fully evolved pokemon with competitive items). If needed you can also scale their pokemon's level to around the player's level, preferably dynamically, using the player's performance and type matchups. It is more work but I also feel that it saves a lot of time in trying to perfect the level curve.

NPCs and Wild Pokemon could be handled this way too. You could even scale them by class or by area. Routes can be designed so that there are 'easy' paths and 'hard' paths. Skyrim does pretty well by just scaling NPCs to the player's level, until the player reaches the level cap for that NPC. They also have some enemy types like giants that have static levels. Difficulty settings typically just buff or nerf the player's damage input/output which is pretty doable for pokemon hacks/fan games.

Sorry for the long post! I've been wanting to do a write up on this topic for a while! And I should reiterate that this is all theoretical as far as pokemon goes but Jenova Chen (Journey and FlOw) has some good research on this on his site if anyone's interested in further reading.



TL/DR

Use gameplay to teach strategy. Scale trainer levels to player to encourage strategy over grinding. Cater game to different FLOW preferences.

 

[Meister_anon~Master_o f_None]

Sorry for the long post! I've been wanting to do a write up on this topic for a while! And I should reiterate that this is all theoretical as far as pokemon goes but Jenova Chen (Journey and FlOw) has some good research on this on his site if anyone's interested in further reading.


Use gameplay to teach strategy. Scale trainer levels to player to encourage strategy over grinding. Cater game to different FLOW preferences.

It's all good man, anyone that's gotten into this stuff can appreciate a well thought out argument/comment. Thanks for adding to the discussion.

 

[TriAttackDev]

Currently for my Crystal hack, I've implemented a new gym gauntlet system where you have to beat the whole gym in one go, as in, if you leave the gym the trainers reset. Because of that, they don't give money or exp, but beating the leader gives a lot of money and exp to the whole team

Usually people just use the trainers in the gym to level up, then leave to heal and then fight the leader. My goal with this system is to create a "mini elite 4" experience

 

[Meister_anon~Master_o f_None]

Currently for my Crystal hack, I've implemented a new gym gauntlet system where you have to beat the whole gym in one go, as in, if you leave the gym the trainers reset. Because of that, they don't give money or exp, but beating the leader gives a lot of money and exp to the whole team

Usually people just use the trainers in the gym to level up, then leave to heal and then fight the leader. My goal with this system is to create a "mini elite 4" experience

That sounds a little terrifying but also interesting, are you doing the normal only difference between gyms is type, or are you doing something different like maybe each leader has a preferred battle style/strategy that the underlings apply to varying/lesser effect? That combined with how your gym is setup would be really cool imo.

Since you can't grind exp with the early trainers, you can at least gain something from the fights, a bit of strategy.

 

[TriAttackDev]

That sounds a little terrifying but also interesting, are you doing the normal only difference between gyms is type, or are you doing something different like maybe each leader has a preferred battle style/strategy that the underlings apply to varying/lesser effect? That combined with how your gym is setup would be really cool imo.

Since you can't grind exp with the early trainers, you can at least gain something from the fights, a bit of strategy.

A little bit of both, it's an open world hack so if you fight Falkner at a higher level he'll have a more varied team, like an Aerodactyl or something. I'm trying to theme it based on type and the trainer's personality

 

[Meister_anon~Master_o f_None]

A little bit of both, it's an open world hack so if you fight Falkner at a higher level he'll have a more varied team, like an Aerodactyl or something. I'm trying to theme it based on type and the trainer's personality

Nice, keep us posted, or just me, I'm definitely interested. So the rom base is actually pokemon crystal?

 

[DoctorKoolMan]

Squirtleinatorinator got it right

Battle Style Set, No spamming items, and give the pokemon EVs/IVs (or removing EVs from your game so you dont overscale too hard)

Then give gyms and other rival battles full teams and you got something thats engaging but still lets you use a team you want to use, instead of it turning into a 'lets farm this section of map for the meta counter to this OP settup the enemy has'

 

[Meister_anon~Master_o f_None]

Squirtleinatorinator got it right

Battle Style Set, No spamming items, and give the pokemon EVs/IVs (or removing EVs from your game so you dont overscale too hard)

Then give gyms and other rival battles full teams and you got something thats engaging but still lets you use a team you want to use, instead of it turning into a 'lets farm this section of map for the meta counter to this OP settup the enemy has'

I can see that but I'm not sure about giving every important person a full 6.

just looking at gen 1 and its remakes, I'm pretty sure if brock and misty had full teams most people would have never beaten the game.

Those first gyms were rough, as they were meant to be, because looking back like 90% of the pokemon you'd encounter before them, are either weak to or lack an advantage against their pokemon.

The only useful pokemon you'd find on your way, would have been mankey, pikachu, and paras. All rare encounters (for the most part)
Lacking the type advantage was a smart move because it forced most players to strategize.

 

[Mighty Golduck]

Very interesting topic. I put a lot of effort into balancing my own hack (I have been working on it for more than 10 years now). You always have to keep in mind that there are different types of players. Therefore, difficulty settings are interesting. The main philosophy behind the creation of, in my opinion, best game experience is to keep the game's main fights strategic and challenging. In my game this means that a decent team needs probably a few tries to beat a gym. You will have to scout and try different tactics first before you get the better of the fight. This creates a very satisfactory game experience in my opinion. But I will also release a very casual version for people who are not really into Pokemon and strategy.

Keep in mind that the player has certain advantages over the AI, such as:
- You are more intelligent than the AI
- You have the option to switch in a hard counter after every KO for free (This is huge)
- You have access to an unlimited amount of healing items (This can be blocked though, but I personally never use medicines)

1. Extreme focus on Pokemon stat, ability and movepool balancing
I spend a lot of time in creating movepools for every Pokemon, including unique strategies. Every Pokemon is viable in my hack up to a certain point in the game. Don't expect to beat the Elite Four with your Butterfree. But Butterfree will definitely come in handy during the early stage of the game. To nail this you will have to test your game over and over again with different team members. I give Pokemons new options which they never had in the main games, so that you can play them with a different strategy. And I had to nerf a few overpowered Pokemon to avoid players from entirely sweeping the most epic boss fights in my game. Balance and strategy are key.

2. Interesting boss fights
I recently rebalanced the bulk of my boss fights. I made a list of overpowered tactics like "stat boost stacking" and tried to find a way to give the AI more options to counter it for a more balanced game experience. I also had to test the AI's capability to understand and successfully use certain moves. By understanding the AI's decision making better, I'm able to create better and more balanced game experiences.

3. Accessbility
I give the players plenty of options in terms of Pokemon and moves, because like I said every Pokemon is viable in my hack. You can try different strategies to win fights and balance your own team. You don't need specific Pokemon to win a difficult fight like in some hacks. In my hack TM's are reusable and move relearners/tutors are free or charge. You have an infinite amount of strategic options in this way and you can customize your team for every boss fight without being punished for it. Wild Pokemon are also generally the same lvl as your team members. So it is very easy to switch team members.

Used mechanics
- Bosses have higher lvls (Because you have access to EV's, but my trainers don't have access to EV's yet)
- I tried to make the AI as intelligent as possible (but it is still lacking up to a certain degree)
- Challenging movepools that are designed to counter common tactics to enhance strategic thinking
- Accessibilty in terms of Pokemon, movepools and strategies
- Reusable TM's and free move tutors/move relearners
- High lvl wild Pokemon to make it easy and comfortable to switch team members out
- I changed some Pokemon moves to create more strategic options or to promote them (for example, haze has priority to counter boosted sweepers)

Hopefully this post has added something to the discussion.

 

[Corveone]

Keep in mind that the player has certain advantages over the AI, such as:
- You are more intelligent than the AI
- You have the option to switch in a hard counter after every KO for free (This is huge)
- You have access to an unlimited amount of healing items (This can be blocked though, but I personally never use medicines)

Well, 2 & 3 are not a problem, they can be disabled either by player's choice (many do that voluntarily to prevent cheesing the game) or by the game locking you from doing it.

So the only real obstacle that keeps battles between human player and AI from being 100% fair is the intelligence gap. But to compensate that, you can allways give the AI access to somewhat better Pokémon, move sets, team compositions and stuff early on. Of course without being outright ridiculous, (i.e. Mega Garchomp with Earthquake on the third gym :)).

Optimal IVs/EVs are also a way to boost the AI Pokémon, but not everyone likes dealing with the RNG garbage to get the Pokémon you want with perfect or nearly perfect IVs + the right nature + the right ability, and on top of that wasting hours on EV training to catch up either. So more user-friendly team building features is basically a must these days, especially if you're making a difficulty-focused rom hack / fan-game.

I agree with your points and your ideas to improve the experience, so good luck with your hack 👍

 

[BoulderSmolder]

Be careful with SE coverage moves in the early game (And the rest of the game too, but especially the early game).

The player gets more options the further they progress, so the AI needs to get more options in turn. If you give the AI perfect coverage at the start of the game, then the first gym is the hardest because the player the fewest options then. Then the player gets a wider variety of Pokemon and TMs and the difficulty starts going down because you can't give the AI better coverage than it already has.

Also, if someone finds a way to cheese a battle... Consider whether that's actually a problem worth fixing. Even the most meticulously well-designed challenge hacks have exploitable battles once you know what you're doing. You can only go so far in mitigating this, and part of the point of making a difficult battle is encouraging people to find creative solutions.

 

[APPro0744P]

As I read I completely disagree
It would be bad to implement any of those as we are just forcing player to do a task again and again , we should add something new (not having any idea about that)
The creativity is required

 

[Amras Anárion]

This is an interesting discussion.
There is one topic that has not been addressed here: giving the player a choice of difficulty. I'm talking about more than a self-imposed Nuzlocke Challenge. Optional rules that the player can activate in the options or at the beginning of the game, such as increasing the level of opponents (this is the principle of the Challenge Mode in Gen 5) or banning items in combat.

Concerning the balancing levers, there are indeed many in a Pokémon game. I'll take the list from the first post.

1. Raising the level
The most instinctive lever for any Game master. It should be taken into account that the player will gain proportionally more experience, which will reduce the level gap.
Ex: You decide to increase the level of all your trainers by 25%. Before, the player would reach level 16 in the first gym against a level 16 leader. Now the player will reach level 18 against a level 20 leader (instead of 16 against 20).
Take into account that due to accelerated progression, the player will have better attacks and better evolutions more quickly. The same is true for trainers programmed by the game master.
However, be careful not to end up with trainers who are blocked by the maximum possible level, while the main scenario is not yet finished.
An alternative solution may be to decrease the experience gains by a global factor. This is a powerful lever to balance the player's progression.

2. boosting their evs/ivs.
Has the same effect as increasing levels, except that it doesn't give the player an XP bonus.
So it's a better idea.
The only warning I give is to be careful not to give overpowered IVs/EVs to trainers at the beginning of the game. Pokémon Mind Crystal is a good example not to follow. (The first youngster on the second route already has his perfect ivs/evs, big difficulty spike).
The automatic mode of Pokémon Essentials is quite well balanced: it distributes the IVs and EVs proportionally to the level until the maximum around level 56 for EVs and 62 for IVs. Thus, the first trainers have few IVs/EVs, but those at the end of the game have perfect Pokémons.
Just 2 flaws:
• EVs are distributed equally over the 6 stats, up to a maximum of 85/stats. A strategic distribution (typically 252+252+4) will have to be done manually. (This is not a problem for a generic trainer)
• The same formula is applied regardless of whether the trainer is a Bug Catcher or a Leader. For the same level, it makes more sense to give even more EV/IV to an elite trainer.

3. programing different movesets
Most trainers will have to use the default moveset. To be reserved for boss and mini-boss trainers, as well as those who have a special theme or strategy. It is even highly recommended to customize the moveset of such trainers: giving them specific strategies will give them contrast and set them apart from standard trainers.

4. More pokemon by trainer
Avoid on generic trainers. Chaining only trainers with 6 Pokémons becomes long and boring.
On the other hand, any self-respecting elite trainer should have a full team (at least 4 Pokémon). This emphasizes the contrast between the classic trainer and the boss or mini-boss trainer.
Be careful not to give too big parties at the beginning of the game: the hero rarely has a complete party at this stage. (Typical mistake: the first Gym Leader with a team of 6 Pokémons).

5. changing battle type. (like more double battles or rotation battles)
A mix of single and double battles (and occasionally triple) is a great idea. The player is encouraged to be prepared for every situation.
There are also boss battles with special conditions (like 2 vs 1) that bring a breath of fresh air.
Getting "gang banged" by two or three trainers at the same time and having to send so many Pokémon at the same time generates strong feelings.
Reverse and aerial battles should be kept occasional.
On the other hand, I didn't like the 1 vs 5 horde fights: it's long. 3 vs 3 is much better and tactical. I also find the rotation battles too slow.

6. Level caps
The ultimate weapon fol all Game masters! Here, three schools compete: no level cap, a soft level cap and a hard level cap.
No level cap is the default philosophy: overleveling is allowed in vanilla games (except for outsider Pokémons). If the player has taken the time to train his Pokémons, he is rewarded with a lower difficulty.
The hard level cap is radical: the player is forbidden to exceed a level decided by the game master.
The soft level cap is a middle ground: exceeding the level limit is allowed, but progress beyond this limit is slowed down.
I have a strong preference for the soft level cap: daredevil players who have the patience to grind will keep some of the fruits of their labor while encouraging all players to stay within a level range.
Note that the XP formula of Gen 5/7/8 has a soft level cap effect: light at high level, but quite violent at low level (< 10).

7. Better AI
One of the best levers: it's so much more fun to play against an AI that approximates human behavior. (Otherwise, PvP would not be as successful). Nothing is more frustrating for the player than to have to face an opponent who has been overleveled to compensate for his debility. It makes the battle very random (and there are already enough haxx in Pokémon).
After that, it is important to shape the AI according to the experience of the trainer. A beginner will have to attack randomly, then gradually, the trainers will have a better AI.
Singular trainers should of course have a better AI than their neighbors in order to stand out.
Oh, and a randomly encountered wild Pokémon must necessarily have the lowest AI. Nothing could be more frustrating than seeing the Rattata from Route 1 play strategically.

 

[Prof_Oatfield]

I love this kind of discussion, and I have seen a few interesting ideas thrown around here.

Many people have different opinions regarding levels and level caps. How do you feel regarding a game that has all Pokémon's stats set to level 50 from the start?
Difficulty would have to be achieved through other means, such as evolution and better movesets as the game progresses.
Whether IVs and EVs stay is debatable, but those could be used as further training rewards.

One thing that I do not like is the ability to solo a battle with a single Pokémon. There are two extremes for this: setup and sweep on one side, stalling on the other. I think that Pokémon would benefit from improvements to its PP system. Make it behave more like energy or stamina. Something that depletes fast and forces you to switch after a while. I have a few ideas on how to implement this, too.

Something you have mentioned is different types of battles, to keep players on their toes.
What is your opinion on making Gym battles more like what we see in alternative media (manga, anime). For example, 3 Pokémon each, the first trainer to faint an opponent's Pokémon wins. Do you go for a fast and offensive strategy, or a defensive and cautious one?
Another example would be for the Gyms to be a teaching experience. To have the player learn more about that particular type, by forcing both teams to have the same type. Brock would use 3 Rock Pokémon against 3 Rock Pokémon of your own (the game would have to provide ways to obtain them, such as access to Mt. Moon, or in-game trades).