The most memorable battles are the ones I've had to try more than 3 times to beat and I'm not talking about fishing for better rolls. They are the ones who create a sense of accomplishment.
Battles I could beat easily (and by easily I mean on the first try; and naturally, I mean playing "near optimally") are the most forgettable and if the game has too many of those, they ruin it.
If you can "naturally" play them near optimally, then they are easy because finding optimal or near optimal solution is a no-brainer.
What I'm talking of are battles that can be won reliably without foreknowledge (player can obtain all the necessary information and knows of it) or counting on blind luck. Battles for which an optimal solution exists and can be figured out, but is difficult to figure out and involves a lot conditional stuff making it branch a lot depending on how the battle will unfold.
Reloading for better rolls isn't challenge, but neither is reloading for better information, nor reloading while blindly trying to bump into a viable tactics. In the first case you reload till the randomly behaving game does something desirable, in the last you reload till the randomly behaving player does something right. Both are as vacuous as they sound.
Second is simply using save&reload as cheap scouting tactics, which is rendered unnecessary by making necessary information available to the player, by implementing scouting, in-game sources, clues and so on.
Tactics (the way I see it) is about trying different things to overcome different challenges. You need to know exactly what you are up against, which requires trying and failing so that you can learn and fail better to learn and have a chance to win.
By that logic tactics is impossible IRL, because real life has no quickload functionality and doesn't allow for retries.
You DO realize how monumentally dumb such idea is, right?
I don't see how this makes sense. How is any player playing 'near optimally, exploiting available info', in a tactical RPG with character development? In most good RPGs, whether old school or relatively new, when you reload you're reloading because:
- (a) you made a judgment / reflex / control / etc error,
- (b) you weren't aware of the particularities of the challenge or did not react adequately to that challenge,
- (c) you have come up against a particularly difficult fight due to the game design and/or your own choices thus far, and
- (d) luck of the draw.
What you say would only be true if we were playing a rather linear RPG where the designers know what skills you have and you should be proficient at the rather limited number of abilities you currently have, and they decided to throw you a challenge which required the luck of the draw/roll to win. In such cases, we actually end up with only (a) and (d), which is similar to some challenges in linear FPS games or something. You either didn't shoot fast enough, or you need to be lucky and hope the enemy's sniper shot misses. i.e. it would be comparable to someone playing an IE game and reloading until their thief succeeds in rolling a 20 crit backstab at the start of the battle, because without that they would get wiped. That would indeed be bad design (or horrendous character building by the player), but that has nothing to do with reload-based combat challenges in general, yeah? So essentially you're building your argument about randomness on an extremely partial and arbitrary context.
In a well designed game b) reduces to a) because inadequate reaction is a judgement error, so is not obtaining relevant information beforehand.
If the information is unavailable, then the game is shittily designed, because it uses learn-by-dying paradigm, presenting player with a set of mutually exclusive survival tactics where only one is valid, and forces player to choose randomly by not providing necessary information (AKA mushroom design - keep player in the dark and feed him manure).
c) reduces to a), b) (which reduces to a) if the game is well designed) or d).
d) should be minimized if outright elimination is impossible without throwing the baby out with the bathwater. It isn't inherently fun, but some degree might be an inevitable side effects of producing random behaviour in game which increases entertainment factor and challenge of a).
It is irrelevant whether the game is linear or not and whether it supports variety of builds, because it doesn't affect what a)-d) are all about, although it influences viability of various design strategies. For example heavily non-linear game with very rich mechanics and variety of possible builds doesn't really allows to make any assumptions about player's character/party, so the only thing you can do is to provide relevant information and enough wriggle room so that player won't be, for example, forced to duke it out with a dragon in a narrow passage even if he plays a pacifist diplomat wearing no armour.
Fortunately if the game has very rich mechanics, it means that almost in any given situation player has access to a very large number of possible actions and, in terms of a mathematical model, it's difficult to trap someone inadvertently in a space with very many dimensions.
To sum it up, a) is the only point that should ideally count, and since it involves player error it means that the encounter could be handled without reloading if played flawlessly, without requiring any sort of additional information or anything.
d) is a side effect (possibly inevitable), of additional mechanics making a) interesting and non-trivial.
Even when we take relatively casual players (i.e. not your grandma, but someone who actually plays games), when they first come up against an enemy, and that enemy introduces a new ability, or spell, or a tactical maneuvre, and you die, the response isn't "Fuck this shit". You're surprised, and maybe think that's pretty cool, and now you know, so you reload to try and get around that.
But how about witnessing someone else getting shafted instead during enemy introduction, or getting a suggestion to flee, then learning of the introduced stuff from in-game sources?
Reloading makes for shit storytelling and any, but the most abstract games are telling stories. They may be non-storyfag, highly mutable stories, that simply flow from player's interaction with mechanics, but they are effectively stories.
"Johnny
gets incinerated by dragon's unexpectedly fiery breath and dies - the end knows for no reason that dragon's breath will be fiery so he gulps a potion of fire immunity"
makes for a shit story.
So take the Kraken in TW2, for instance. It varies depending on the player and difficulty level, but many people complained and even quit, not necessarily because the Kraken was too difficult, and certainly not because you needed luck of the 'draw' or randomness to win. It was because the Kraken introduced too many mechanisms and challenges that were too different from combat challenges in the rest of the game, and because there were not adequate feedback mechanisms to allow the player to take it in without many reloads. i.e. you not only need to learn to dodge the tentacles, you also need to learn to dodge or cut apart the sticky stuff, you need to learn to hit the tentacles in exactly the right spot, you need to learn to dodge the tentacles when they're flailing around
Those aren't the problems with Keiran battle.
The main problem is that you're not informed that the tentacles will start flailing around, you're not informed that the beast will suddenly chuck pieces of bridge at you and you can never be sure of what exactly are you attempting to do, due to the battle being based on shitty QTEs and having no unified control scheme of any sort, relying on same two buttons doing whatever arbitrary shit devs saw fit at that moment, so, for example you'll get flung into walls many times before figuring out when to let off.
The last one was also the reason why said goodbye to optional QTEs for good in TW2 - dear nonexistant god, they were so shitty.
Sluggish controls and iffy collision also didn't help.
Now, one solution is obviously to make the challenge a lot easier, i.e. every one of those challenges remain, but they deal less damage and are less punishing, so the player can take their time. I'd argue this is a suboptimal solution. Not only is the player still going to have to spend a lot of time figuring stuff out, as they're doing so, the generosity of the challenge takes away from the excitement. In other contexts, thus you have people saying "Oh just ignore all that shit, just go in and hack it with swords and it dies". It's not fun to reload 50 times to figure out the Kraken, but it's also not fun to sit there and let it hit you 20 times while it deals 3 damage as you figure out how to kill it. A much better solution is to retain or even strengthen the challenge, but reduce and/or make clearer and more intuitive the various challenges presented at you. In the context of the Kraken, that might mean allowing you to strike at the tentacle at ANY given point, rather than having a scripted spot and timing, so that people can stumble on their own ways of killing the thing. It might just mean taking out one of its special unique abilities, or even just giving the whole thing a wider camera angle.
How about providing player with responsive controls and clearer idea of what he's attempting to do?
Cut off the tentacles, then what? Charge it and poke it with a sword? Oh, it dropped a bridge on itself, how cute. Now it flung me against the wall. Wat.
If the hazards need to be presented in action, you can always make a warning shot - an event that doesn't kill the player but presents the risk adequately - for example sniper missing or hitting another target, tentacle breaking a trunk of an ancient tree like a twig, if all else fails, some exposition dialogue. After that player is a fair game.
Figuring Keiran's weak spots didn't exactly take an Einstein especially considering optional dialogue with Triss. Yrden was suggested by Sheila when you were rolling around trying not to get mashed, so it was also no-brainer.
What was really derp was that you didn't even have any clue what will you do for most of the combat - did Geralt expect that the thing will collapse part of the bridge on itself and that it will fall so conveniently that it will become a ramp? Is he a fucking clairvoyant?
Same with Keiran trap - why only the two outer tentacles? Shouldn't I be able to deploy it in any location and aim at any tentacle?
Another problem is that player is prevented from exploring the situation by non-obvious instakill hazards.
The last thing is unreliability of controls - it's no fun to have to do something that makes character completely unresponsive when certain death is flailing about. It's common issue with modern cinyeemathigifagic experiences, though, that doing stuff roots your character in place, that power attacks make him dash in random direction for a second or two before even executing them (hi, oblivious) and so on. Consolefag shit.
Of course, better FoV would be useful, I wouldn't have much problems if it was wide enough to allow simultaneous tracking of both tentacles and ground. I probably would also fight the battle relatively easy in FPP, with standard responsive FPP control scheme, but to sum it up, the reloads in Keiran battle were frustrating because you either knew perfectly well what to do exactly and how (yrden, now dodge you faggot! dodge I say! FFFFUUUUU-) while dying constantly, or you were killed for not reading devs' minds and figuring out what you're expected to do.
Remove those reasons, you'll remove the reloads. The battle will also turn rather easy, because other than unresponsive controls and not being given any clue regarding what to do there is no challenge but being fast with your keys to it and being fast with your keys is what you should have perfected halfway into prologue.
The threat of death needs to be real, and actually carried out at times.
The threat of death should be real, but death shouldn't be used to explain the challenge to the player, only to punish him for failing. Also, as many failures as logically plausible should not lead to game over, because while death may be an ultimate punishment for the character, player will just reload, making it rank low on player's punishment scale.
Being advised (by NPCs, or sources you obviously seek and read at every opportunity) to avoid melee with zombies, then after the fight, good night sleep and several insignificant encounters you nevertheless saved after learning that you feel sick, that it's because you've caught something nasty, that it WILL kill you, and that only way to avoid this is finding a really good healer, who will demand most of your precious stuff in exchange is better than having "you contract a deadly disease and die shortly after, reload (y/n)?" pop up during the combat, because while the latter is technically harsher on character it's much softer and very forgettable for the player, who can just reload.
So naturally, it gets harder over all. I personally found Fallout always too easy, and I'm not even that much of a Min-Maxer or "pro" at gaming. Making everything a wee bit harder and adding more possibilities in combat is what I want.
