If we turn it into a two card game, the results get a bit more interesting. If we turn it into a 2 card game with player control to be dealt more cards, split, double, etc. we get blackjack which is amusing enough on its own (at least until we find out that there is only one mathematically optimal way to play.

If we turn it into a 2 card game with 5 community cards and massive combinational variety and player control, we get Texas Holdem which has kept players entertained for years and years, and the results can vary from a 7 high to a royal straight flush.

Of course with RPGs, we want a good deal of determinism. If the player felt like he was completely the victim of chance, he might as well play card games. However, if the level of variation in the results is interesting enough, I think that can add a lot to the fun factor. When we aim for a character’s eye in Fallout and manage to blind him,it never really gets old. When we shoot at a character’s arm and cripple him and make him drop his weapon, there’s a little bit of delight and excitement. If we blow his entire arm off, there’s a little jump for joy each time. And however annoying it is, when that happens to one of our characters, I think that level of unpredictability keeps people playing. Then again if it’s hit/miss and our 80% hit misses 3 times in a roll, we can’t help but be annoyed by the random number generator, not just because it’s screwing us over, but in a really uninteresting way.

I just want to start by saying I recently found out
about Telepath Tactics through Kickstarter, I think it will be great
when it’s finished, and I’m glad you are giving it a second go. I’m also
really glad that you are making posts like this and having this
conversation with the TRPG community.

Now, about the randomized results.

The
post is very well done and I’m very glad you came over to the design
position you have arrived at. I think the specific instances of
randomized results you mention are great. However, I would like to make a
small case for the intrinsic value of randomized results as a baseline
for attack success or failure.

I hate the game Risk because it
is too random. I love Fire Emblem because it is regularly random
(absolute misses and absolute hits are possible, but most attacks fall
in the range of probabilities between).

It would seem that aside
from the Assasin’s dodge ability attacks success is almost always a
guarantee in Telepath Tactics. I do think it will be loads of fun, but
was I took a bit of a pause when I initially read that about the game
design.

At first I thought it was simply a matter of familiarity
(Fire Emblem, FF Tactics, Dark Wizard, Vandal Hearts, etc, etc,) all
have regular use of randomized results in the combat engine. I kept
thinking about your points and weighing whether I my initial reaction
was out of apprehension of change or a legitimate merit of randomized
results and their use in games.

In general I don’t think that
they remove control from the gamer, they just give the gamer one more
uncertainty to consider. They add uncertainty to outcomes, but that
doesn’t mean they are not tactical decisions or not in the player’s
control.

The example you used with the troll is
not bad because of randomized results, but what sounds like either bad
game design (designing a game in which there are really good odds that
the player will lose a character in that fight, stacked against the
player party) or a failure on the player’s part to effectively up-level
his characters, equip them, or master their uses before they arrived at
that fight (maybe he’s using his b-squad entirely). Other possibilities
and questions would obviously linger; why did the player move all of his
five characters within striking range of the troll? Is it impossible to
have avoided this scenario in the fictional example? Are they all
injured characters or is the Troll just so strong that he can kill any
character “almost certainly” one hit? Again it seems like either bad
game design (too powerful troll) or bad long term player choices
(inadequate leveling decisions or positional choices) would have led to
this circumstance.

Conversely I lost a character to permadeath
very recently in my copy of Path of Radiance. The character was at full
health and being used by me to assist in accomplishing a secondary
objective (clearing a path for another character to be used in stealing
an item I wanted). The character in question had a very good chance of
being hit by the enemy but only a moderately decent chance of receiving a
critical hit and thus sustaining enough damage to be killed. It happened though.

It
would have been absurd for me to feel that I wasn’t in control of it
however. I chose to put that character in that situation, while even
though it was not the “most likely” outcome death was a reasonably
likely outcome. I made the tactical choice however that the uncertain
possibility of death to that particular character (whom was a b-squader I
chose for a special purpose in this mission), was a tactical gamble
that was justified by my desire to acquire a stolen item. I played a
gamble, lost something, and gained something more. While I would have
been more excited to pull it off without the loss, the loss makes my
tactical choice in play feel more meaningful -there was a cost from my
choice, which could have been to never risk anything.

So then,
why is that good? For starters I think it’s much more realistic. Yes, I
know you are talking about play dynamics, specifically how they make a
game fun. For me, and many players tactical realism in a TRPG is very
fun. Comparing TRPGs to chess loses something, I think. chess is 100%
deterministic because it is predominantly or entirely abstract strategy.
Yes symbolically the pieces “represent” knights, and kings, and
bishops, but tactically it is pure abstraction. There is no tactical
legitimacy as to why the queen is a super badass and the king can barely
move, nor ever defend himself. for that matter why would a bishop ever
only move in a zig-zag pattern or knights an L.

The entire game
is abstraction of strategy. It is a finite possibility of circumstances,
built to be expressed in a format and world that doesn’t even feign
realism, and abides absolutely and unflinching to a small set of
immutable laws and rules. All that can be gained from it in terms of
real world strategy is principles because of the highly abstract nature:
Plan ahead, deceive and distract, be willing to sacrifice to win, etc.

TRPGs
while having some abstraction (because they are a game) and some
obvious breaks with realism (because they are usually fantasy based),
tend much more to be an avenue for more realistic and less purely
abstract strategy. Realistic tactical details are usually considered a
good thing in the genre, like mounted units traveling farther, arrows
striking from a distance, mud or swamp slowing character movement, and
certain weapons having varied effects. While still being partly abstract
(and unrealistic) they tend to try to represent more of the real world
of tactical considerations through environmental factors, strength/speed
trade offs, and well, attacking with skill and weapons instead of
jumping over one another for the kill.

The particular role that
the dice (or random number generator play in this is a continuation of
realism in tactical decision making). In my younger years I was a
military intelligence analyst in the US Army and uncertainty of outcome,
even when the odds are in your strategic favor, do not guarantee
knowledge of the outcome. For me the fun in many games is tied to
weighing differing potential outcomes, contingency plans, and what ifs.

You
see it in life everywhere. No matter how certain one is that boxer A is
a better fighter than boxer B, a win by boxer A cannot be guaranteed
beforehand(aside from match fixing). The huge array of potential
variables (mistakes, miscalculations of lefts or rights, reacting
slightly too slow, underestimating, etc) all persist. The roll is the
representation to account for this. You can get odds, and in games like
fire emblem is is possible to get a guarantee, but most fights have a
least some degree of chance.

That chance while being more
realistic, also provides for added thrill when you gamble with a
strategy and come out on top. Such as getting a high critical hit
possibility in the right place at the right time. Conversely when things
that have odds of going right go wrong, you are forced (as sometimes
happens in real life strategy) to recover a bad situation that odds said
were not likely to happen – but were in fact possible.

This
causes players to need to be improvisational and is another area where I
see randomization directly benefiting the player’s options of
situational strategies innately. One reason that computers can beat
people in chess is because it is so deterministic that computers benefit
from being able to “memorize” the likelihood of various outcomes for
every possible configuration and recall all of those outcomes
immediately. Human minds are not as ready made for that and the greatest
chess players on earth have vastly different neurological development
from all that playing that optimizes them for such thinking. However,
studies have shown that if you inject even small amounts of rule changes
in the game great players become just like everyone else because all
their strategy was highly specialized in the immutable rules of
tournament play. They actually suffer greatly from having to face small
“unexpected” changes.

Improvisation is something that, unlike
humans, computers cannot due. The best they can due is give the illusion
of it with clever and deceptive programing. That realistic injection of
randomized outcomes (going this way when things should have gone that
way) gives players a way to exercise that humanly unique aspect of
strategic decision-making, improvisation.

“Crap, that character should have been dead by now, guess I’ll have to come up with a plan that changes everything.”

1) “Vertical” randomness, where the result of the randomness determines whether the outcome is good or bad and to what degree (eg, dodge chance, chance to hit, variable damage).

2)”Horizontal” randomness, where the result of the randomness determines which of a selection of equally “good” or “bad” options is selected. Eg, A character in an RPG gives you a random magic item as a reward. The result is always “good” and to approximately the same degree, but the random factor introduces variety; there are many possible items with different properties you could receive. Random enemies, random loot (although that tends to use vertical randomness as well), procedurally generated content in general are all random factors which increase variety and possibility space while not arbitrarily giving good or bad outcomes to the player.

So my principle would be to always aim for horizontal randomization to increase possibility space and minimize vertical randomization for the reasons described in the article.