Gameplay feedback (may contain spoilers)

A sci-fi RPG game
Post Reply
Young scout
Posts: 43
Joined: Sat Jan 19, 2013 9:37 pm

Gameplay feedback (may contain spoilers)

Post by Seloun » Sat Jan 19, 2013 10:49 pm

I've just played through Planet Stronghold, and I enjoyed it immensely. Lots of fun and entertaining gameplay. I do have some criticisms regarding some of the gameplay design elements, however.

1) Psionics
The psionics skill seems too dominant by scaling too well. It controls the effectiveness of the powers as well as how frequently you can use them (via resting regen) and how effectively you can do other things while using powers (by controlling turn based regen), and damage you do with a psionic weapon. In particular the non-resting regen should probably have been associated elsewhere (possibly the highest psionics skill) to let non-psionicist classes with a tagged power use them more freely. This would still keep psionics as a quadradically scaling ability (better powers and more use of powers); moving resting regen out as well would help in that regard.
Conversely, the individual psionic power skills seem too specific, though allowing non-resting regen to be available more readily to non-psionicists might help with that perception. This would also help differentiate the classes/characters a bit more.
Also, stacking Protect is probably overpowered; combined with aggro control, this seems to basically trivialize battles without mixed damage ranges/types. Disrupt is similar, but is much shorter duration, only affects one enemy and probably better balanced of the two.

2) Burst/Aimed, accuracy and weapon types
It seems somewhat counter-intuitive that the weapons that perform best with burst are weapons that fire fewer shots. Meanwhile, aimed shots seem mostly pointless except for single shot weapons (since you basically lose nothing in that case). Likewise, it seems odd that Accuracy, a scout skill, benefits multi-firing weapons more than sniper rifles.
I'm not sure if there would be a simple solution to the burst issue; in practice the mode seems to work well (it's just counter-intuitive) as an aggro control mechanism. One possible solution would be to have the extra shot hit a random target, which would make it less valuable in some cases for low fire rate weapons.
Accuracy probably should have added damage on a crit, while Aimed probably should have increased crit rate much more dramatically - maybe always one shot, always crit, with base crit value being much smaller than it is now, and Accuracy adding to crit values. This would make sniper weapons more valuable for high armor targets (which seems reasonable) though that weapon niche is also somewhat shared with heavy weapons. In practice though against any significant target disrupt + fast firing weapon seemed to be the better option, partially because strong general purpose psionicists meant disrupt was available as a side effect of everything else. Also, anything that makes highly unreliable things a little bit more reliable tends to feel useless (especially if the outcome of the unreliable thing is very significant) - as addressed in the next point.

3) Most important: Random binary skill checks with immediate feedback
This is really a subcategory of a more general pet peeve of mine: games should not make save/load feel like the strongest available power. Inevitably and in truth save/load is probably always the biggest advantage the player has, but it shouldn't feel that way. Random binary skill checks with immediate feedback strongly exacerbates that feeling; your choices in character development is trumped by a standard, basic game feature.
In this game, these show up in two primary ways:

* Check to avoid an encounter/make encounter easier
These can be dealt with by a fixed skill check rather than a random check. Simple solution, and accomplishes the same basic function. This works because the check is one-time and generally not threatening to the player continuing the game. This requires slightly more care in designing the skill threshold; if done poorly a player could feel like they're 'chasing' the goal threshold without ever hitting it (which would then feel like their investment is wasted); this is usually easy to counteract by not making the thresholds always increasing (sometimes you run into a hard check, sometimes an easy one). More coarse-grained skill values also make it easier for the player to gauge how good a character is at something; at the most basic level it would be a 'Science' tag/feat/attribute instead of a science skill value, or a 'science' 'adv. science', etc. to whatever level of enumeration.
* Check that must be passed to continue
This shows up in the game as e.g. locked doors which must be picked to progress, and each failure triggers a new encounter. Simple threshold obviously would not work, though a threshold with a stacking bonus per attempt would, or a cumulative score (you need 200 lockpick points to continue, and each encounter adds your lockpick skill - maybe with a _small_ variance - to the total). Those options also provide a better way for the player to determine how much they need to ration resources (since they can estimate the number of battles required to succeed).

In both cases, the other option is to delay feedback. In the random lockpicking example, the player might always fight an encounter after the choice (without seeing the result of the choice); it's only after the battle the result is displayed. Typically, longer the delay the better, though this approach mostly works by increasing the cost of reloading (player time) rather than reducing the benefit, though more accurately most solutions to this works by increasing the cost of reloading (as in principle you could always restart to match e.g. a threshold check) - the only exception is if there's another in-game resource constraint (e.g. there are two checks, and you cannot get enough points to pass both) but those solutions tend to be very difficult to implement and often very punishing to a new player.

Random results in battle tend to be more forgivable because each roll is much less binary in its result. You might take 12 damage instead of 11 damage, and you might save/load to optimize for that, but in principle any single instance (or even small number of instances) of those results are not likely to make a difference in the ultimate outcome (and even if they did, you often can't tell they will until much later). There are exceptions - in this game, a 3-point Harm in a lot of cases is a Random Binary Skill Check With Immediate Feedback; it would have been far better if the damage was a small range or otherwise not so dependent on one roll. The same is true for most of the powers, e.g. Protect and disrupt should be modelled as something that always works, but with the effect somewhat randomized instead of a full power thing that hits or misses (as a side note, randomized value on protect would make it considerably more situational instead of being able to dial up just enough armor to put you 1 point below the maximum enemy damage range).

I would like to reiterate that I had a lot of fun playing the game (I would purchase it again knowing what I know now), and to note that the lengthy criticism is mostly a reflection of my own interests.

User avatar
Pack leader
Posts: 13745
Joined: Thu Jun 16, 2005 4:43 pm

Re: Gameplay feedback (may contain spoilers)

Post by jack1974 » Sun Jan 20, 2013 8:38 am

Thanks for the long feedback. If you browsed the forums you have noticed that we're working on the sequel right now. It will use a very different (improved) engine, that follows already many of your suggestions, like Psionics (or even weapons) that do a range of damage rather than being hit or miss (there's a detailed explanation by the coder in PS2 subforum, I confess I am not much good at those math things myself :oops:).
Is interesting the idea of displaying the exact skill value needed to pass a check, though sometimes to make the game feel more realistic, there will be checks that needs to be picked immediately. I mean for example a locked door, the player might be given the option to fall back and visit that location later. But in other cases, like the team is trapped in a underground facility that is about to collapse by an earthquake and they need to hack a terminal to open the door, that check must be done immediately. However knowing the skill value needed might help others to make walkthrough for the game in any case :)
As for the random encounters, in the sequel there will be less grinding for XP, more for resources (since the game is based on surviving and there's also a simple colony sim building minigame) so I should be able to design better the skill checks for each situation in the game. Of course what you suggest to have a base skill and increase it proportionally to the party average level is also a good idea.

Post Reply