Your opinion on "coinflip" or "gamble" type skills

[Andar]

If you believe a deal double damage with 50% chance to miss is OK, I don't even know what to tell you.

no - but that is a bad randomizer no matter what way you look at.
no one will take any gamble if there is no chance of winning something, and statistically double damage at 50% is absolutely the same.
if you make it a double damage with a 40% miss chance however, a lot of people will take that roll.

5% chance to insta-kill the boss. Do I pass?

again, that is a bad use of randomizer no matter what.
I have designed several PnP-rulesystems from scratch just for the fun of it, and really considered the good and bad aspects of rolling the dice in them (and there are dice systems out there that are really bad for fun, no matter how a DM handles them).
I don't have time for it now, but I might give a description and ask for opinions in a different topic later this week or month.

and high numbers are not inherently more "grindy".

not inherently, but subjectively.
the human mind simply "blinds out" at high numbers. It is very rare that a player considers more than the first two or three digits on any high-digit-number and tries to process those numbers.
but if you keep numbers below 100 or so, people are much more aware of them, and that has effects on how they handle a situation.

It's the same reason why you hear sentences like "a hundred deaths are a tragedy, a million deaths are a statistic"

 

[HexMozart88]

I think it's all right if you have a good risk-to-reward ratio. One of my games has all of its skills operate on a dice roll, and I haven't received too many complaints about it so far. The only thing you have to watch for and that I've had to watch for is having "sinkholes" where the player keeps missing every time or keeps doing 1 damage and the enemy keeps doing 20.

 

[GHNeko]

As someone who believes there isn't simply just bad game design, but rather bad execution. I think gamble/rng style skills have their place and can work well and not cause player frustration but the barrier for proper execution is higher than most people guess; especially if they're not aware of how dated standard rng-based design in JRPGs are.

Without going into too much detail, I have a "Gambler" class, and RNG is built into the core design of their kit and they're balanced around the idea that turns can vary wildly, however; rather than the traditional "do more damage or miss", I've opted to bake pros and cons into "good" rng and "bad" rng as a core design. I say "good" and "bad" rng because the class has to roll for 2 states it self inflicts and depending on the roll, the magical stats and physical stats of the character will go up or down based on the opening roll. Each style of roll unlock different skills that are either risky but strong (for bad rolls with stat reductions), or safe and reliable (for good rolls with stat increases) for the player to use. No matter the roll, you'll get something that you can take advantage of for a period of time before you need to reroll for new states. Since the class is something you opt into and aren't forced into; this also means that those who don't like the concept of RNG at all in their game can still avoid the core concept of gambling.

I don't think that RNG/Luck heavy skills or even classes are bad, but I do think a lot of the time from my own experiences, is that people use this type of design in a more "traditional" format that is unchanging and unadapting to modern design and sensibilities.

 

[M.I.A.]

My answer as a Player: I almost never use a risk skill unless completely desperate or I am so safe and have no chance of failure just for fun.

My answer as a Developer: I like to ask the questions "how can I make this better? How can I improve upon this for a player who may like it?" For chance skills, I lie. I may say that it has a 50% chance to miss or a 50% chance to crit.. but it's a lie. For a coin-flip skill like that, I will actually program it as having a 25% chance to miss and a 75% chance to crit. This can encourage the player to use the skill more often since the reward is more perceivable. This is not always how I make a coin-flip skill though. I may also make it to where each time the player uses the skill and it misses, the odds of firing off a crit are increased, but I will never decrease the odds of a crit below 50%.

Just my personal philosophy around gamble skills. There are several ways to make them in the players better interest and favor than just leaving it as 50/50.

Hope this helps!
-MIA

 

[Basileus]

@Sword_of_Dusk brought up some good examples from Dragon Quest already, but there is another example in the form of the Jester class from Dragon Quest III. Jesters start out pretty normal but start to ignore the players orders sometimes to do random things, and they learn more and more tricks as they level up. This can range from attempting to do a party trick and breathing fire on all enemies, to telling a joke that causes a monster to start laughing and miss their next turn, to stunning themself.

Naturally, more serious players will never even consider using this class, but some players will enjoy the quirky randomness to break up the combat loop. They also have a secret reward for putting up with them - unlike all other classes, Jesters can upgrade to the Sage class without needing an extremely rare Scroll of Enlightenment.

So the randomness can be a source of fun for some players and a kind of "Magikarp Power" for others. In fact Pokemon borrows the other kinds of RNG too, like with the move Metronome and the disobedience mechanic. While not every player will like these mechanics, they are kept optional and can even lead to some fun challenge runs. I think there can be some value to creating high-risk/high-reward or gambling moves as long as you don't force them on the player.

 

[Milennin]

no - but that is a bad randomizer no matter what way you look at.
no one will take any gamble if there is no chance of winning something, and statistically double damage at 50% is absolutely the same.
if you make it a double damage with a 40% miss chance however, a lot of people will take that roll.

That's just people taking advantage of the odds, since a 40% miss on double damage means you're on average hitting better than you would with a regular attack. But just because people would take the gamble doesn't mean it's good or that it's fun. I take the gamble on attacks with a 95% hit chance because it's the best thing I get in some games, and I still hate that it will miss one in 20 times.

I find RNG acceptable if missing the dice roll doesn't make or break my turn in combat. Or if the gamble is on a secondary effect of a skill, as long as the primary effect goes off I'm OK with that. I challenged myself to make a big RNG swing in RTP with an unlockable passive that adds a chance to insta-kill to the regular attack which I find myself using quite a lot when I get it. Even if it doesn't go off, at least it gets the regular attack damage in, so the turn isn't completely wasted. Sometimes, it'll go off a few turns in a row making easy work of encounters, other times it will not go off at all and I find myself taking a lot longer in an encounter than I otherwise would. On rare occasions, it'll even wipe the party because that one party member continues trying and failing to insta-kill enemies with an attack instead of contributing by using their skills. I've ended boss fights with my party on the brink of death by landing the lucky insta-kill on them with that passive, which was pretty awesome too. The RNG on the boss is also determined by the boss's HP. The lower its HP is, the higher the chance for the insta-kill to trigger becomes.
Still, RNG will never be perfect because the variance on outcome is too big. Which is why I'll never be a fan of these abilities.