kieranmillar
Level: Smitemaster
Rank Points: 2670
Registered: 07-11-2014
IP: Logged

How do you make a good RPG hold? Regular DROD has years and years of analysis and examples to draw from in terms of what makes good puzzles, but such a wealth of information is considerably lacking for RPG.

After having played pretty much all of the RPG stuff out there, optimised large amounts of it and topped the leaderboards, and made my own stuff, I'd like to think I've learned a bunch over the years about what makes for a good RPG experience. So I thought I'd share it rather than hoard the knowledge, because RPG needs more content and I want to play more! So here goes!

The Most Fundamental Point: It's All About Timing!

Distilled down to its essence, RPG is a matter of being presented a number of choices, then making these choices in the right order (or not making them at all) so as to minimise the amount of resources you spend / maximise the amount you gain. So all you need to do is emphasise this by coming up with a mixture of choices that all encourage different timing. This is most easily explained by demonstrating it with examples, some real, some imagined:

* Here's an example straight from my hold The 'Neatherhood, where really early on you can open 3 yellow doors for a green key and 3 yellow keys. At first this might seem weird, because the overall cost is actually just zero, you spend the keys and get them straight back again, but early on there are also choices to open single yellow doors to get stats. Now you want the stats early, but if you get them early, you can't stockpile the yellow keys needed to get the free green key. Do you want the green key early, or are you prepared to wait later to pick it up to get the stats now instead? At some point you want the green key because, hey, free green key, right? But it requires a commitment and when are you going to make that commitment?

* At the start of the third floor of the hold Trial Tower, you can make a trade where you kill a monster and spend a green key and 90 greckles, to get your green key back, along with some HP and two yellow keys. But if you choose to progress any further in the level beyond this point, you will lose access to this trade completely as a pressure plate will close a black door in front of it. The trade is good, but it's also really expensive, because green keys are at a premium and every green key trade up until this point has been one you want to make. Is it worth the cost of saving a green key until this point to get the extra yellow keys?

* You've got some resources to collect behind a gauntlet of Mad Eye gazes, but alongside those resources is also an expensive but good trade. Ideally you want both, but you can't easily afford to make that trade right now, and you really need that cache of resources right now. If you get the cache now, you'll have to pass through the gauntlet a second time in order to make the trade later, increasing the total damage you've taken from the gauntlet and effectively making the trade less efficient. Is it worth it?

* At some point in order to make progress, you need to open a row of three yellow doors. Behind each yellow door there is a small amount of health. Scattered throughout the level are monsters with a lot of health behind them. Imagine you need to get some health. Do you kill one of the monsters for a big chunk of health, or do you spend a key early in order to get the smaller amount of health? That key could have been used elsewhere right now for a better trade-off, but you need to spend it anyway at some point, and also want to leave killing the monsters until later when they will hopefully be cheaper. Do you push ahead with required progression earlier to pick up resources you'll be getting anyway? Maybe you could end up avoiding making some of those optional trades! Or maybe you are focussing too much on long term benefits, and will run out of resources in the short term if you open the doors now.

As you can see, there are lots of ways you can make decisions really interesting by focussing on their timing component. When making choices in your hold, try to think about it from this perspective, and come up with incentives to make some choices late, and some early.

You can then also layer these choices for more variety. Imagine a trade-off that's quite inefficient, but also opens permanent access to an area with more choices that are more efficient than the average. By paying that "toll booth" early, you get access to better trades. Is it worth the extra resources up-front?

Understand How Scoring is Calculated, Then Ignore It

10 points = 400 HP = 2 Attack = 1 yellow key. You should understand how scores are calculated, because it's really important for optimising RPG holds, and sets a reasonable baseline for many people when first tackling your hold. If you don't understand some of the more technical aspects of RPG, check out the wonderful Architectural Handbook thread stickied at the top of this forum. It explains scoring, among other things.

So of course, your yellow door trades need to be for 400 HP each, and you should make green keys roughly twice as expensive to obtain as a yellow key, right? No. No no nonononononono no.

Knowing how scoring is calculated is important because it tells you what an optimiser is going to do "around the edges". So for example if all of your green key for HP trades give about 1,000 HP each, then before triggering the scorepoint the player will trade any green keys they have spare for HP, and if they know they will end up with spares then earlier on they may prefer to gather HP by spending green keys.

But when it comes to designing your level there are other factors that are more important for deciding what things should cost. Is there an area of your hold where you expect green keys will be scarce? Then you can make any green keys you place in this area more expensive. Maybe by doing this you encourage the player to consider holding on to green keys earlier on so they can bring spares into this more expensive area and get some long-term savings. Maybe in the next area the trades are more efficient, so the player is encouraged to try and push forward and get there earlier. Great! It's all about timing, remember, don't tell me you've forgotten already!

It's also important to note that similar trades do not have to have the same costs either. You want to have the flexibility to make some trades more efficient than others to encourage the player to take those more efficient trades.

Go back to the first example I gave in the last section, where the player gets a green key "for free" if they stockpile yellow keys. This actually isn't free, because stockpiling the keys means delaying getting stats, making the fights they take cost more HP, so getting this green key actually costs HP. But how much? It's not easy to answer, and by making it not easy to answer, playing the hold is fun rather than rote calculation, with intuition and experimentation playing a larger role.

Ultimately you should come up with your own idea of what the various resources should cost based on the choices you want to make, and then shake it up depending on the situation and to enable you to make the choices more interesting.

---

[Last edited by kieranmillar at 09-22-2016 09:31 PM]

kieranmillar
Level: Smitemaster
Rank Points: 2670
Registered: 07-11-2014
IP: Logged

Percentage Damage

Hot tiles, Aumtlich beams and uh... stabbing bombs while you have a speed potion active. All of these things deal an amount of damage proportional to your current hit points. If you've listened to people talk about Rule 6 of Tendry's Tale, you've probably heard people grumble a lot about percentage damage, but that's because Tendry's Tale handles them quite poorly.

The truth is that percentage damage can work. There's a time and place for it. However bad use of percentage damage can also make your hold significantly less enjoyable to optimise.

There are two major issues with percentage damage. The first and foremost is that it can make optimising your hold incredibly tedious. Percentage damage can often be made to deal only single digit points of damage, but in order to do so you need to micromanage your health. In Tendry's Tale, it is optimal to take the percentage damage as soon as you get to it and is always much cheaper than any alternative, but to do so optimally you need to keep going back and forth across the entire length of the hold to go back to the early levels and pick up small health potions so you keep your health critically low at all times. Given that it can take over a minute to do this back and forth, it gets very tedious very quickly, and makes replaying sections an incredible chore.

The second problem is that often there is no interesting timing component to the decision. If you can go back and collect left behind potions at any time, it is always possible to reach the percentage damage source with minimal health as soon as you come across them. Always taking the decision right away simply isn't interesting. Opening a door or fighting a monster vs taking less than 4 points of damage is almost never going to be an interesting decision.

If you're not sure about whether or not to include percentage damage in your hold, it is safer to simply not include it at all. Yes I know the hot tile is right there in the editor on the main tab, staring at you and begging you to use it, but please don't. The risk of making your hold worse to play is quite high.

If you really do want to include it, you really need to focus on tackling the two issues outlined above. Here's how to do that:

#1: Making percentage damage not be tedious: What you need to do here is prevent players from micromanaging by walking long distances constantly. One simple way is to just make your hold short. Another way is to make your hold non-backtrackable, so for example you can't return to previous levels. Unfortunately that's pretty much all you can do. Including lots of shortcuts, while a good idea anyway, generally won't be enough to solve the constant topping up of health, which is what you're trying to avoid.

#2: Adding interesting timing decisions into when to take the damage: Percentage damage encourages you to play at low health all the time, if it's possible, which makes it uninteresting, so the trick here is to not allow that. If the player's health changes throughout the hold numerous times between small and large amounts of health, then there are interesting timing decisions to make. So what you really need to do is come up with ways to force the player to keep taking health in large chunks instead of allowing them to leave behind lots of small potions. Here are ways you can do that:

* Non-backtrackable levels encourage players to collect all remaining health before moving on as this is optimal for score (and generally sensible anyway). That way they will hopefully enter the next level with a large amount of hit points.

* Simply put health potions in the way. A bit sneaky, but it works. Could be health you're forced to pick up to progress, or when the player gets a reward of stats/keys and some health, put the stats/keys behind the health and force them to take the health to get the main reward.

* Don't let the player take potions individually. You probably don't want to have only large potions, so use force arrows and trapdoors to ensure a player picks up everything. (Obviously the latter only works if you can't disarm yourself freely)

* Make getting health or crossing hot tiles more expensive if the player leaves them until later. Want to move into the latter half of the level? That enforces you to trigger a pressure plate, locking additional doors in front of those nice health trades and hot tile trades, so the player is encouraged to get them early.

* Give the player a reason to cross the hot tile at high health. Your player has just entered the next level, and can't backtrack to previous ones so has a lot of health carried over from the previous level. There's stats and keys in the Entrance room. Take them now? Or wait until later? It'll be cheaper if you wait, but maybe overall it's better to take more percentage damage now?

You may have noticed that calculating the benefits and timings of these decisions can get quite complicated in a hurry because the damage you take is always in flux. Some people just don't enjoy that level of complexity and will not enjoy percentage damage no matter what you do. Might be worth keeping in mind. Again, if you're not sure, don't include it!

---

[Last edited by kieranmillar at 10-18-2016 08:04 PM]

kieranmillar
Level: Smitemaster
Rank Points: 2670
Registered: 07-11-2014
IP: Logged

Difficulty

Setting the difficulty of RPG holds to the right level is a bit more difficult than regular DROD holds for a couple of reasons. One is because instead of dealing with a single room, you're dealing with an entire hold. The other reason is that due to the way RPG gameplay tends to snowball in the long term with all these different decisions having long-term payoffs, it can be difficult to recognise what the current difficulty even is!

But a quick note that's important to mention first, unless your hold is completely trivial, playing to fully optimise and top the leaderboards is almost always going to be difficult. Even a small number of decisions can start to get very complicated very quickly when you're playing to eke out every hit point of savings. So if you want to make a hold that appeals to as many people as possible, just err on the side of easiness! The less skilled players will play to get through it, and the optimisers will do what they always do and make things hard for themselves.

What makes an RPG hold difficult is resource scarcity, usually health. Coming up with the long term plan is easy, but not being able to afford it in the short term massively complicates the planning in a way that trying to crunch the numbers doesn't solve very easily.

Let's explain this with an example. It's usually the case that you want to maximise gains in the long term so it's the first thing people try. You usually want to pick up the stat gems and equipment as soon as possible, and kill as few enemies as possible until you can come back at the very end and kill them all for the minimum cost. This lets you collect all of the resources as cheaply as possible, ending the game with more stuff and therefore more score. So there's this overarching plan that you follow, and sure there's some fiddling around in the short term to figure out the right order, but ultimately you're executing the same overall gameplan, right?

So you've picked up all of the stats and equipment in the order that seems cheapest, then picked up the cheapest keys to reach the boss, but he massively overpowers you. So then you go back and pick up all of the remaining health potions, but it's not enough! You can't even make the boss any cheaper to fight because you've maximised your stats at this point, so you're completely stuck. Does this sound familiar? It should do if you've played a hard hold like Cludo's Dungeon or Hold Anonymous, where this problem tends to happen early, but it also happens to people in the last level of the otherwise very easy hold Ten-Minutes Distraction, where people reach the final boss when it's dealing minimum damage and don't have the health. So what's going on here?

The key is to recognise that in DROD RPG, you are always trying to make trades where you pay resources up front in order to make savings that slowly pay off over time and then start to profit in the long term. So say you kill a monster for 200 damage to pick up a +1 defense gem. That extra defense is 3 points i.e. 120 HP-worth of score, so you're going to need to get hit 80 more times until you break-even on score, and after that you will start to profit from the trade, and will profit more and more as time goes on, with each future hit you take saving you 1 HP. So ideally you want this as soon as possible. But you have to pay 200 HP up front for it, and its not until 200 hits later that you are finally at the stage where you have the same HP if you had just not collected it. And this is fine if until you reach that point, you always have 200 health to spare. It's when you don't that you run into problems. So once you've been hit 120 times it looks like you've gained 1 score by making the trade (120 HP of savings + 120 HP worth of score in DEF = 240 HP, +40 more than the initial cost) but you're still under an 80 HP deficit and if you need that HP to survive, then you've caused yourself to fail.

If you want to properly understand difficulty in the holds you make, you need to understand this idea of payoff. How much does the player pay up front, and how long does it take to break even? Sometimes the payoff can be very quick, so if you have monsters where picking up small amounts of stats massively reduces the damage you take, then the payoff is going to come quickly, so the player benefits more from getting stats ASAP. If the damage from monsters doesn't change much, then it will take longer to pay off. Typically defense always takes a long time to pay off while attack is much quicker.

If you want to set the appropriate difficulty level for your hold, you need to primarily consider how generous you are with your health. It seems obvious that the less generous you are, the less room there is to make mistakes. But also you need to be aware that the less health there is, the harder it is to pay the upfront costs. Crucially, there exists a tipping point where health becomes scarce enough that you cannot make all the trades you would ideally like to, and as soon as you reach that point, the difficulty starts to ratchet up considerably as you remove more health.

You may have played some holds where the early levels seem quite hard, but suddenly you reach a point where the hold becomes insanely easy from that point on. The issue is usually that early on the resources are low and the architect wants to make tough choices so there isn't enough HP to do everything you want. But then the architect decides that as there might be a lot of variation in what the players are doing, they should accommodate for weaker players who might be in a worse position and need more health to keep going. So they add in a bunch more health, and in doing so push it beyond that tipping point where suddenly the player can make the ideal long term trades and get all the stats easily and it all completely falls apart.

If you want your hold to feel like it has a steady difficulty curve, you actually want to do the complete opposite to the above. Provide more health earlier on and then ramp it down as the hold progresses, reducing the spare health past that tipping point. Then the player will be forced to go back and optimise the early game more so they can bring more spare resources into the part of the game where there aren't enough resources being added to meet the costs. The slower you ramp it down, the more gentle the curve.

Of course, because RPG holds are one large puzzle that spans the entire hold, this idea of a difficulty curve doesn't really make too much sense if you think about it. After all, if the players reach the stage where they need to go back and optimise the earlier levels better, then those early levels aren't quite as easy as they first seemed. So RPG holds don't really have a difficulty curve so to speak. Instead, you provide a series of areas that are easy to stumble through at first and let the player explore, but as the hold progresses players need to find more and more optimisations throughout in order to progress further, which makes it feel like the latter levels are harder even though deep down the whole hold is just as hard as its hardest point because the whole thing is just one big puzzle.

So to summarise:
* Players pay health up-front for more score later.
* There exists a point where there is not enough health to meet all of these these up-front costs.
* To provide the feeling of a gradual difficulty curve, start above this point, then slowly reduce the health as the hold progresses until you meet and then go below this point.
* Optimising will almost always be difficult, so the more you force players to optimise your hold, the harder it will be.

---

[Last edited by mrimer at 09-24-2016 03:44 PM]

kieranmillar
Level: Smitemaster
Rank Points: 2670
Registered: 07-11-2014
IP: Logged

Choosing stats for custom monsters

The stats that the monsters have in your RPG hold are crucial to get the right balance and to construct the choices that you want to make. Because RPG gives you full freedom in making monsters with whatever stats you want, the architect can freely design any monster with any sorts of stats for any situation! Great!

But what numbers do you choose? If you've ever spent time fiddling with the stat values for your monsters but felt frustrated because you're not getting exactly the results you want, such as gaining a small amount of stats resulting in the threat they pose dropping considerably, then this post is for you!

There are 4 questions you need to ask yourself when designing any monster:

* What sort of damage do you want it to deal at the time you expect the player to be fighting it? Pick a single set of player stats as a starting point e.g. the player will have 30 ATK and 20 DEF and I want it to deal around 180 damage in total, hitting the player about 4 times.
* How quickly do you want gaining stats to reduce the effectiveness of the monster? Do you want the monster to remain quite tough throughout the time the player is focussing on that level, or do you want to it to start off as a behemoth and then quickly become a big wimp?
* Which stat do you want to reward the player for prioritising? Attack, Defense, or a mixture of both?
* How much GR and REP should the monster give?

The second question is quite important as it gives you another way to control the timing element of the choices you make, which (as you should know by now because I keep repeating it) is the most important thing.

Let's look at each of the stats and see how these affect the way the monster behaves:

*HP: The amount of attack it takes to go from dealing 1 point of damage to a monster, to killing it in one hit, is exactly equal to its HP value. Monsters with small HP values lose effectiveness very quickly as the player gains attack (e.g rock golem and mud mother), while conversely, monsters with lots of HP keep posing a threat for a while (e.g. gel baby). So we want to set the HP value in line with how quickly we want the monster to react to the player gaining attack.

*ATK: The major thing this affects is how much damage is saved by reaching the next attack checkpoint. So a monster with very high attack (e.g. the swordsman) strongly rewards the player for raising attack to certain values. In rarer circumstances, your focus will be using it to set how much defense you need to reduce the damage the monster deals per hit to zero.

*DEF: This fine-tunes where the attack checkpoints occur. It can also be set high like a rock golem or wubba to completely block the player from fighting it until a certain time.

Time to put all this information into practice. (Finally!)

----------

Let's say you know the player will have around 50 Attack. You want to design a swordsman-like monster that has an attack checkpoint fairly early on. So we decide on having it hit the player only once when we reach 60 Attack, and then the one-shot value won't be until about 90 Attack. We can't be too picky about deciding exactly where every attack checkpoint will fall, but knowing the distance you want to have between two specific attack checkpoints is very helpful.

So our basic starting point is going from 1 hit at 60 Attack, to 0 hits at 90 Attack, a difference of 30. We will assume the player has a sword and so strikes first, so we hit the monster 2 times at 60 Attack, and 1 time at 90 Attack. Therefore we must be dealing 50% of the monster's HP at 60 ATK, and 100% of the HP at 90 ATK. So 50% of the monster's HP is taken off with 30 damage. Therefore this monster will have 60 HP.

Plugging this HP value back into the attack values we wanted, we want to be dealing 60 damage when our attack is 90, therefore the monster must have 30 defense.

The attack value we set will then depend on exactly how much HP we want to save by reaching 60 Attack, and will depend on the player's expected defense. If we want to save 300 HP and the player will have about 40 defense, then the attack value of the monster should be 300 + 40 = 340.

Final stats:
HP = 60
ATK = 340
DEF = 30

----------

Let's run through a second example.

You want a monster that hits the player a fair number of times and you want it to be largely resistant to raising your attack. Say it hits you 5 times when you first reach it with 170 Attack. The player will be expected to reach 220 Attack when they can move to the next area, and we still want the monster to be hitting us at least 3 times by then.

So let's choose values of taking 6 hits to kill at 160 Attack and 4 hits to kill at 230 Attack, a difference of 60 Attack. You will be dealing 17% of its HP at 160 Attack and 25% at 230, so 8% of its HP is taken off by 60 damage. Therefore it has (100 / 8) * 60 = 750 HP.

At 230 Attack, we will be taking off 25% of its 750 HP. 750 / 4 = 188, so we will deal 188 damage at 230 attack. 230 - 188 = 42, so the monster has 42 defense.

When we are being hit 4 times, we want to take about 500 HP, and will expect the player to have roughly 150 - 200 DEF, so each hit should deal about 125 damage to us, so we want the monster to have about 300-ish attack.

Final stats:
HP = 750
ATK = 300
DEF = 42

----------

Note that there are important limits on some of these things when the player has low stats. You can't really have a monster that hits the player a small number of times but hardly reacts to raising attack at all if the player has very low Attack. This is because you simply need to have a monster with large HP in order to get gaining attack to scale slowly, but if the player has low attack, then a monster with large HP is going to take a lot of hits to kill. There is no way around it except giving the monster negative defense, which you can actually do, technically, but I don't recommend it. You'll know it's one of these situations if following the steps above produces a negative value for the defense the monster needs to have.

Instead you could start the game by dealing with bigger numbers, so high starting stats, but then you lose some of the feeling of having the character grow a lot throughout the hold. Alternatively, you simply have to accept that early in the game monsters will lose effectiveness quickly.

----------

So far we have only focussed on attack checkpoints. What if we want a monster that rewards gathering a lot of defense? Well as it turns out, they aren't actually very easy to make and balance in the way you might want. The problem is that if you want a monster to hit you a lot of times, then the damage per hit can't be very high unless you want it to be some sort of boss monster. And if the amount of damage it deals per hit is low, then it doesn't take very much more defense to reduce the monster to dealing trivial or zero amounts of damage. So if you want a regular monster that deals middling damage to heavily reward gathering defense by e.g. hitting you a double-digit number of times, then it is almost always going to rapidly lose effectiveness as you gather more defense. There's no real way around this.

Therefore it is often better to not specifically focus on making specific monsters that reward gaining defense, and not treat gaining defense as some sort of "alternative path" to gaining attack. Instead treat it as something that saves damage in a different way with different timing. It's often more interesting to have to focus on raising both anyway.

----------

Some other miscellaneous notes that I feel need mentioning:

* If a monster reacts very slowly to gaining more stats then the player is encouraged to kill them earlier. If you've just reached an attack checkpoint and the next one is a long way away, you will be waiting a long time to get further benefit, so you might as well kill it sooner rather than later for more benefit in the short term.

* Conversely, if a monster reacts rapidly to gaining more stats, then the player is encouraged to leave that monster for later as they will make larger savings for delaying these monsters than others.

* But ultimately, it's often all about the timing of attack checkpoints. Try to tweak the defense values of your monsters to ensure the attack checkpoints come at different times so the player keeps prioritising different things.

* As the player gains attack it begins to suffer diminishing returns against a particular monster. As the number of hits you takes reduces, the amount of extra attack you need to hit the next attack checkpoint goes up. You may have noticed that going from taking 1 hit to one-shotting the monster requires an amount of attack equal to 50% of the monster's HP. Going from taking 2 hits to 1 hit requires gaining 16% of the monster's HP in attack. From 3 to 2 hits its 9%, etc. Keep this in mind when deciding how many times a monster should hit you. If it's going to be a lot of times, then it also won't take very much attack to reach the next attack checkpoint and will then scale quickly with gaining attack.

* By comparison, gaining defense provides a linear cost reduction that does not slow down as you gain more of it. It only slows down if you raise your attack to reduce the number of hits you take, but savings from gaining more defense comes at a linear rate. You may look at a monster that hits you 10 times and think that as you gain defense, going from taking 50 damage to 40 is a 20% reduction, and then going from 40 to 30 is a 25% reduction and therefore defense gets better as you get more of it, but this is not true. The savings was 10 both times. It's a flat rate.

---

[Last edited by kieranmillar at 10-18-2016 10:55 PM]

Nuntar
Level: Smitemaster

Rank Points: 4592
Registered: 02-20-2007
IP: Logged

kieranmillar wrote: Therefore it is often better to not specifically focus on making specific monsters that reward gaining defense, and not treat gaining defense as some sort of "alternative path" to gaining attack. Instead treat it as something that saves damage in a different way with different timing. It's often more interesting to have to focus on raising both anyway.

Well, I have to correct you on this point, because testing Tetrahedron has made me think a lot more closely about ATK and DEF "paths".

A large number of holds have altars (again I'm using this as a catch-all name, no matter what they are called in a specific hold), which continually force the player to choose between gaining ATK and DEF. Even if there are no altars, there may be micro-decisions, e.g. a cache containing one or more power gems, and a cache containing one or more shield gems, and a trapdoor/bridge setup ensuring you can only take one. And even without decisions of the "lost forever" type, in most holds you constantly have to choose which stat to raise now with the resources you currently have.

And the thing is, the more the player concentrates on one stat, the more it becomes beneficial to continue concentrating on that stat rather than balancing the two. Let's say at some point you have to choose between a single power gem and a single shield gem. If you've concentrated on raising ATK, you've concentrated on reducing the number of hits you take from enemies, so the shield gem is comparatively lowered in value. If you've concentrated on raising DEF, you've concentrated on reducing the amount of each hit, so the power gem is comparatively lowered in value. It's inescapable, unless you have a hold with only one type of gem so that the choice never arises.

In my post on altars in the other thread, I suggested one possible way to encourage a "mixed" route: have the altar offer a mixed option, and make it so that taking the mixed option twice gives more than taking ATK and DEF once each. Now that I've remembered this, I'm considering adding such an option to Tetrahedron. However, even if I do that (or if someone does in another hold), this only creates an additional "path", it doesn't change the fact that the choice of paths is a very real phenomenon.

So, what does this imply for setting monster stats?

Some monsters strongly encourage gaining ATK (Rock Golem, Swordsman), while others strongly encourage gaining DEF (Rattlesnake, Slayer). However, monsters in the latter category can be tackled on an ATK path, and their damage can be reduced by gaining more ATK (or equivalents such as Invisibility Potion, Wyrmsmiter). Golem-type monsters are simply impassable obstacles on a DEF path, unless you can gain just enough ATK to hurt them.

That means that if you use too many of one type of monster, you are taking a potentially interesting choice away from the player, or punishing them for taking the "wrong" path. Look at how Tower of the Sorcerer / Tendry's Tale handles it: there is only one Rattlesnake, and it guards an optional extra. There are a few Swordsmen, but they can all be avoided by taking alternative routes, except the ones in the boss battle -- and those are placed so that they always get in one hit, reducing the difference between the paths. There is only one Adder, and completing the normal route gives you a Wyrmsmiter. There's plenty of HP available on the last level, so that the Slayer is viable on either path.

____________________________
50th Skywatcher