Running away

[skjarl]

What's the rationale behind not being able to run away from monsters anymore? We're the same speed, but somehow they can keep up and hit me while moving. I don't understand why this was added.

[Sar]

https://crawl.develz.org/tavern/viewtopic.php?f=5&t=9446

Look for "energy randomization".

[TheDefiniteArticle]

basically the devs consider randomly killing a player who made no mistake to be preferable to "kiting"

also random energy doesn't actually fix kiting in any real way: you still kite like the old days, it just gets cut short now, because the ogre randomly gets an extra move and kills you halfway through the process

[KoboldLord]

If you wait until you're one hit from death before you start trying to run away, you've already made your mistake. Next time, if you can't handle that ogre try running away at full health.

[Lasty]

. . . or you randomly get one extra step away and escape up the staircase/through the door when you thought you were toast. Whichever.

[TheDefiniteArticle]

When you are on D2, "one hit from death" is 100% hp.

More importantly, you don't even need to be "one hit from death" for random energy to kill you. You're exploring normally, and you come upon this situation:

  Code:

################
.............@.#
##############.#
##############.#
##############.O

You are a smart player, and you know that you shouldn't try to fight that ogre in your current state. You know the correct move is to run away and take some stairs.

Except thanks to the fact that random energy exists, that ogre can close the gap. And if he does, you as the player are just shit out of luck. It is completely beyond my understanding how anyone can think this is a reasonable situation.

Another situation that I've encountered:

  Code:

..........
.......@D.
..........

In this scenario I already have Swiftness active (it is old Swiftness, so my movement delay is 0.8), and the hydra has no active effects, and there is no water anywhere in LOS, so the hydra's movement delay is 1.0. The up staircase is a ways to left, just off-screen. There are no other monsters around; in fact I wish there WERE other monsters around, because a couple nearby hallways would allow me to put the monster between me and the hydra.

Now, in a sane and rational world, I should be totally safe. I move faster than the hydra, so I should be able to create a gap. There are no outside factors to prevent me from making that gap. I am at full HP, and even on the outside chance that the hydra rolls max damage (which is extremely unlikely), it can't kill me in one attack.

Here's what happened:

  Code:

........
...<@D..
........

Also instead of being at full HP, I am now at 25 HP.

. . . or you randomly get one extra step away and escape up the staircase/through the door when you thought you were toast. Whichever.

The fact that you can make a horrible mistake like try to fight that ogre, and then get reduced to 5 HP in one melee attack, and then walk away and survive in the same situation that killed the smarter player, is NOT any better. It is equally stupid.

[rchandra]

also random energy doesn't actually fix kiting in any real way: you still kite like the old days, it just gets cut short now, because the ogre randomly gets an extra move and kills you halfway through the process

Sure it does, it makes it so kiting isn't always the best choice all the time. It's easily possible that the best way out becomes through the enemy, or using a consumable, instead of regenerating health at no cost.

[Lasty]

examples of random energy screwing me

I question these examples. In the ogre example, the ogre cannot catch up to you and swing without the two of you taking at least 20 steps (more likely 25ish), and on every single one of those steps rolling the 1/3 chance of gaining energy instead of the 1/3 each chances of losing energy or having normal energy. The odds of this ogre catching you over an arbitrarily long distance are worse than flipping heads 25 times in a row on a fair coin; the odds of it happening on a reasonable timescale are much lower.

In the hydra example, energy randomization can only add 1 AUT per step, and you gain 2 AUT of space each time you take a step w/ oldswift. This example couldn't have happened unless either there's a detail you're forgetting or there was a bug w/ movespeed at the time.

[monty]

Was energy randomization changed recently in trunk? It's existed to some degree as long as I've been playing Crawl.

[moocowmoocow]

https://www.youtube.com/watch?v=BxoFScgRFIY

[duvessa]

thedefinite's examples are weird, but randomized energy does absolutely nothing to stop kiting/pillar dancing except for ONE monster (orc priest) and actually makes it worse with every other monster: it is much slower to wait for randomized energy to make a gap than it is to just weapon swap - unfortunately, the former is optimal. and in the cases where you do intend to fight a speed 10 monster, it is optimal to move away from it until it is next to you, because that minimizes the number of actions it gets; previously you could just wait.

[tasonir]

People in this thread are acting like randomized energy gives monsters a predictable gain in speed over time. My understanding is they have a 1/3 chance of losing an aut, gaining an aut, or no change. I've had monsters fall behind a square while I was kiting (and often got up stairs only because of this) many times. in TheDefiniteArticle's first diagram, the ogre has to move twice to be in melee range of him - there's no way in hell he could catch up if he runs immediately for a stair, even if the stair is 30+ spaces away, he'll make it easily.

Likewise, it is mathmatically impossible for a 10 aut-per-move hydra to catch up to an 8 aut per move player. Even if they hydra rolled "gain an aut" every turn, he's still falling behind by an aut to the swifted player. Either I'm wildly offbase or something is very wrong here.

[TheDefiniteArticle]

People in this thread are acting like randomized energy gives monsters a predictable gain in speed over time.

When I am adjacent to a titanic slime creature, I act like it will hit for the maximum possible damage every time it attacks. If I bet my life on "it'll probably be okay", I will die hundreds of times before I make it to Zot. And as you know, dying once ends the game.

In the ogre example, the ogre cannot catch up to you and swing without the two of you taking at least 20 steps (more likely 25ish), and on every single one of those steps rolling the 1/3 chance of gaining energy instead of the 1/3 each chances of losing energy or having normal energy. The odds of this ogre catching you over an arbitrarily long distance are worse than flipping heads 25 times in a row on a fair coin; the odds of it happening on a reasonable timescale are much lower.

20 steps is easily a low end estimate, since levels in this game are gigantic and autoexplore doesn't care about stairs. Obviously the big scary thing doesn't catch up like that every time or even commonly (but it shouldn't be possible for it to happen even once; a sink that only spontaneously combusts once every ten years is not a sink I want in my kitchen). What normally happens is that over a long walk, the monster will move a bit closer and then a bit farther away, back and forth. Which is also horrible because it encourages me to run around the same circle a dozen times until the fucking planets align.

This example couldn't have happened unless either there's a detail you're forgetting or there was a bug w/ movespeed at the time.

I'm sure you're right; the fact that this whole thing is complicated, opaque, and unintuitive is the main reason it gets on my nerves so much. Walking is the most fundamental and basic action, so why does it have to be obscured by crazy rules?! Other awful tedious mechanics are simple enough or obvious enough that I can take steps to minimize/avoid the tedium (such as item destruction), or else they are unimportant enough (such as item destruction in Zot) that I don't consider the advantage to be worth the expended effort.

But running away from enemies is a matter of life and death that can easily occur many times on even a single floor, and after achieving basic competency, getting better at running away has utterly dwarfed every other source of improvement for me. So I have two choices: I can read the source and figure out whatever those rules you just rattled off were, and then count the distance to every staircase and door every time I need to run from a scary thing to make sure it can't possibly catch me. Or, I can avoid all that by just assuming that any monster which gets closer than 5 spaces away is already guaranteed to reach melee range eventually and behave accordingly. I'm assuming the design intention is for me to pick the second option, since every design decision in Crawl works towards the goal of extreme-yet-vague-paranoia.

[Igxfl]

TheDefiniteArticle, have you ever played ChessRogue? It's completely deterministic. It's also fiddly and horrible because of that.

If Crawl's current design is encouraging *too much* paranoia and running, which it very well might, the solution is not allowing deterministic ways to carefully stay one move from being damaged. The solution is to penalize overly paranoid behavior a little more, because forcing risky decisions makes for more interesting gameplay.

[Patashu]

In the interest of informed public discourse, I have created a 'Will the ogre bonk you on the head before you can escape?' calculator. It is available as a jsfiddle: http://jsfiddle.net/BfET5/ (Press the 'run' button to make it appear, I don't know why it doesn't show up right away)

It produces scientifically rigorous and peer checked results automatically such as:

  Code:

In 50000 universes, with the ogre 20 auts away and the stairs 100 tiles away:
The ogre bonked you 1.318% of the time!
You escaped up the stairs 98.682% of the time, without a single bonk on your head!

Let me know if the algorithm is wrong.

UPDATE! I believe this updated version gives more relevant-to-playing-Crawl information: http://jsfiddle.net/BfET5/1/

[tedric]

In the interest of informed public discourse, I have created a 'Will the ogre bonk you on the head before you can escape?' calculator. It is available as a jsfiddle: http://jsfiddle.net/BfET5/ (Press the 'run' button to make it appear, I don't know why it doesn't show up right away)

It produces scientifically rigorous and peer checked results automatically such as:

  Code:

In 50000 universes, with the ogre 20 auts away and the stairs 100 tiles away:
The ogre bonked you 1.318% of the time!
You escaped up the stairs 98.682% of the time, without a single bonk on your head!

Let me know if the algorithm is wrong.

I don't know what's wrong, but I feel like it can't be right. For very reasonable real-world values like 20 auts/50 tiles, I'm getting an average of less than 0.1% bonks (to avoid confusion with decimals and percentages: that's less than one tenth of one percent, or less that 1 bonk in a thousand trials). Anecdotal evidence from my own experience says I get bonked much more often than that.

EDIT: Hmm, maybe that's just user error. I was thinking 20 auts was 2 tiles away (me, empty tile, ogre) but I guess that's actually 10 auts. 10 auts returns much more intuitive results.

[Patashu]

In the interest of informed public discourse, I have created a 'Will the ogre bonk you on the head before you can escape?' calculator. It is available as a jsfiddle: http://jsfiddle.net/BfET5/ (Press the 'run' button to make it appear, I don't know why it doesn't show up right away)

It produces scientifically rigorous and peer checked results automatically such as:

  Code:

In 50000 universes, with the ogre 20 auts away and the stairs 100 tiles away:
The ogre bonked you 1.318% of the time!
You escaped up the stairs 98.682% of the time, without a single bonk on your head!

Let me know if the algorithm is wrong.

I don't know what's wrong, but I feel like it can't be right. For very reasonable real-world values like 20 auts/50 tiles, I'm getting an average of less than 0.1% bonks (to avoid confusion with decimals and percentages: that's less than one tenth of one percent, or less that 1 bonk in a thousand trials). Anecdotal evidence from my own experience says I get bonked much more often than that.

Keep in mind that an ogre that is two tiles away from you (visually 20 auts) could actually be as close as 11 auts away (one more aut gain and he'll double move to be one tile away from you), and 11 auts/50 tiles gives me 5.8%. Unless you saw his most recent doublemove or move loss you can't say very accurately how many auts away he is.

I also think that going up stairs takes more than 10 auts - but I'm not confident enough in my Crawl knowledge atm to say if that matters or not.

[tedric]

Knowledge bots say stairs take 13-14 auts.

[Patashu]

Are the assumptions in this version 2 correct? (again, press run once to make it show up) http://jsfiddle.net/BfET5/1/

I haven't accounted for stair delay yet - does it mean that, for example, if I reach the stairs with the ogre 13 auts away and hit <, the ogre will bonk me but not follow me? If so, I need to make a perfect trial > 13, not > 10.

Also, what does 'stairs take 13-14 auts' mean? Does it mean 50/50 chance? 33/66 chance? Is it worth a sourcedive?

[TheDefiniteArticle]

It's completely deterministic. It's also fiddly and horrible because of that.

Firstly, you'll have to defend that statement before I can take it seriously. Secondly, my opposition to randomness is confined solely to movement. Just like how in Street Fighter 2, it's perfectly fine that damage and stun and even input windows are randomized. Because when you push past all the fluff and flavor and side-options and look at the core rules that define what the game is, it is ALL about positioning. Everything else ultimately only matters by virtue of having repercussions that effect positioning.

If Crawl's current design is encouraging *too much* paranoia and running, which it very well might, the solution is not allowing deterministic ways to carefully stay one move from being damaged. The solution is to penalize overly paranoid behavior a little more, because forcing risky decisions makes for more interesting gameplay.

What I'm saying has nothing to do with Crawl encouraging too much or too little paranoia and running. Crawl encourages this amount of paranoia and running, it's just how the game is, and it's so deeply engrained that it will never ever change, and if you think risky decisions make for more interesting gameplay then Crawl is not the game for you. In fact pretty much every change to the game that encourages risky behavior is widely considered awful and if not removed, maintains a quiet hatred from long-time players. The current incarnation of curse skulls is a textbook example.

My complaints are that random energy:
>Meaningfully lessens the effect of player skill
>Encourages kiting
>Is unintuitive and opaque and for the core mechanic of a game, that is a mortal sin

The crux, though, is this. As horrible as I really believe all three of those things are, they could all potentially be acceptable if random energy provided some benefit which made up for the loss. But I am unable to see ANY benefit! And the only answer one gets when asking about the benefit of random energy is "it discourages kiting." To my eyes, that statement is not just false but self-evidently false! The whole reason kiting is bad is that it takes a lot of time, so obviously randomization, which levels out over a long time period, will not have an appreciable effect. In the short bursts of activity that make up normal play, random energy can have drastic effects on the outcome of engagements, but when kiting the grand total effect of random energy amounts to a few bits of trivia and really really dumb spoilery optimal behavior.

This is the sort of thing that keeps me up at night. If I at least understood the train of thought, I could say, "Oh, well they're just fucking lunatics" and forget about it and return to the business of playing the game (which is how I handle curse skulls).

[tedric]

Are the assumptions in this version 2 correct? (again, press run once to make it show up) http://jsfiddle.net/BfET5/1/

I haven't accounted for stair delay yet - does it mean that, for example, if I reach the stairs with the ogre 13 auts away and hit <, the ogre will bonk me but not follow me? If so, I need to make a perfect trial > 13, not > 10.

Also, what does 'stairs take 13-14 auts' mean? Does it mean 50/50 chance? 33/66 chance? Is it worth a sourcedive?

I know that I've had enemies move adjacent and hit me when taking stairs, but they only follow if I start up the staircase while they're already adjacent. Beyond that... I don't know if it's truly worth a sourcedive, that's up to you and what bigger, better things you might spend your time doing

[Patashu]

1)

Hmm... From my tests, stairs work like this:
-When you take a staircase, you spend 10 aut on the floor you took it on (20 if encumbered - Just say no to burdens, kids). (Source: main.cc, search for _take_stairs)
-On the floor you arrive on, you spend 6.666 (randomly rounded) aut if the floor was just created now, 13.333 (randomly rounded) aut if the floor already existed. (Source for this part is files.cc, search for you.time_taken ) (Didn't test to see if this part is also longer when burdened, but it's possible)

For example:
You're standing on a down staircase leading to D:3, which you have loaded before. Ijyb is next to you.
You press >, spending 10 aut on D:2 during which Ijyb hits you, then you're sent to D:3.
A further 13.333 aut (randomly rounded, so let's say 13 in this case) auts are spent on D:3. There also happens to be a giant cockroach next to the upstairs on D:3, and depending on his energy, he'll get one or two hits off on you.

(Also, haste seems to speed up both parts of it - I just saw someone take 1.5 turns in total to take stairs when it should have been 2.3-2.4, which is about 2/3rds.)

So that's set straight.

2)

tag_followers(), which tags all monsters adjacent to you MF_TAKING_STAIRS BEFORE the 10 aut delay on the level you are leaving, means that a monster that is not adjacent to you when you physically press '>' or '<' cannot follow you on that staircase, period, even if you're burdened or whatever it still won't happen. Cool! (Source: main.cc, _take_stairs)

3)

The next claim I want to look at is 'while taking stairs, your EV is halved' which I have heard from I think crate.
So far I haven't found a line of code that is responsible for such a thing. Does anyone know where it would be?
And is the EV halving meant to be on the level you took the stairs, or on the level the stairs lead to, or both?

[ontoclasm]

This is the sort of thing that keeps me up at night.

Maybe some relaxing music might help.

I wasn't around when energy randomization was implemented, but I'm pretty sure "discouraging kiting" is not the intent. The impression I've gotten is that energy randomization is meant to prevent states where the player is somehow trapped and forced to pillar dance. For instance, suppose you are a deep dwarf, you're at 1 HP, and an adjacent hobgoblin is chasing you around a pillar. For whatever reason, all exits from the little circular hallway you're in are dangerous and you don't have any relevant items. Obviously this is a pedantically specific case, but less extreme examples probably happen pretty often. Well, you can't safely take any action that isn't moving away, and you can't safely leave the hallway, so the "correct" action on your part is to just walk around in a circle until you starve or a randomly generated monster comes and kills you, I guess? I mean, I would personally just roll the dice, but maybe you're on a streak or are the sort of person who uses "optimal" multiple times in one sentence.

In this case, energy randomization sort of nudges the situation towards a conclusion: yes, you might get hit and die, but eventually you will gain a step, and can take an action to maybe change things (for instance, by taking a staircase). Sure, I've been "screwed" by an ogre getting a free hit (this happened to me occasionally when I sucked even more at Crawl than I do now, but notably I can't recall a single time in the past, say, 600 games), but I frequently make use of energy randomization to get away from monsters.

It's maybe not a good system, but I'm pretty sure it's not meant for what you think it's meant for, and doesn't work the way you think it does. By the way, note that the only person in this thread who's said "it discourages kiting" is you yourself, so. Uh.

[and into]

Deep Dwarf is actually a really bad example because it can't regenerate. Without energy randomization you just attack and hope for best, with energy randomization you move and hope for the best. The former is actually more fun and less annoying than the latter, so it is a point against energy randomization.

My understanding is that energy randomization is/was intended primarily to limit one's ability to take advantage of HP/MP regeneration during indefinitely long retreats (pillar dancing). It may succeed somewhat in that, but I think it introduces too much weirdness in terms of how one positions in many situations where one is not trying to walk around for unreasonably long periods of time. It seems a bit weird to me that energy randomization was adopted here while leaving in a number of enemies that do not naturally regenerate HP, for example.

Plus—and this is something that people always forget—there is an inbuilt risk to kiting or really just to moving for extended periods of time when a non-trivially strong enemy is in LOS. It is very seldom a good idea, actually, and would continue to very seldom be a good idea even if energy randomization were removed (or if it had never been added).

[duvessa]

I am sure everyone arguing against randomized energy is aware it was introduced to combat pillar dancing. This is why, in my first post in this thread, I stated that it does nothing to combat pillar dancing, except in the single case of orc priests. Other monsters either:
1. are speed 10, do not have ranged attacks or a means of becoming not (effectively) speed 10. Can be "pillar danced" in almost exactly the same way with or without energy randomization, you just use 1 more wall tile as part of the "pillar".
2. are speed 10, have means of becoming effectively not speed 10 (haste, slow, confuse). Could never be pillar danced in the first place because they can haste/slow/confuse you after you attack.
3. are slower than speed 10, same result as category 1 except you don't even need walls.
4. are faster than speed 10, same as category 2.
5. do not appear in the first few dungeon levels, after which pillar dancing is useless against any monster.

Now if every monster in the early game were the same as an orc priest (speed 10, ranged attack that does damage and damage only, nothing else) then energy randomization would seem a lot more reasonable to me. But the vast majority of early-game monsters are in category 1, and pillar dancing/kiting them takes a much larger amount of real-life time with energy randomization. Furthermore, randomized energy introduces

in the cases where you do intend to fight a speed 10 monster, it is optimal to move away from it until it is next to you, because that minimizes the number of actions it gets; previously you could just wait.

which is really horribly incredibly awful and I consider it far, far more tedious than pillar dancing orc priests.

[skjarl]

Nobody has called anyone's mother a whore yet, so I consider this thread a rousing success. That said, I dislike this mechanic enough that I'm going to pull down the source, disable it, and compile a new version to play with locally. I don't always agree with the decisions the developers of this game make, but I can live with most of them because DCSS *IS* a fantastic game.

[Sandman25]

This is why, in my first post in this thread, I stated that it does nothing to combat pillar dancing, except in the single case of orc priests. Other monsters either:
1. are speed 10, do not have ranged attacks or a means of becoming not (effectively) speed 10. Can be "pillar danced" in almost exactly the same way with or without energy randomization, you just use 1 more wall tile as part of the "pillar".

I don't understand. I am a caster on D1, there is a lucky goblin who refused to die and made me have 1 HP and 0 MP.
1) Without energy randomization: I safely pillar dance for full HP and MP, then try to kill the goblin again. If I fail, I can repeat unlimited number of times (assuming no other monsters generate).
2) With energy randomization: I can be killed every turn. So I am forced to avoid such situation because I cannot safely pillar dance.

Another example - I can kill a monster in 2 hits, it can kill me in 2 hits too. I can attack once and see if monster missed me. If it did, then I pillar dance until max HP and try again, effectively killing the monster without getting in danger of being killed. It is safer than attacking all the time where I could die in 2 turns. Again energy randomization prevents this tactics because I still can die in 2 turns.

[basil]

I don't understand. I am a caster on D1, there is a lucky goblin who refused to die and made me have 1 HP and 0 MP.
1) Without energy randomization: I safely pillar dance for full HP and MP, then try to kill the goblin again. If I fail, I can repeat unlimited number of times (assuming no other monsters generate).

And if a monster generates or some other monster happens to walk by, you're likely to die.

I see pillar dancing as a last resort instead of some turbopowerful strategy; usually, I don't have any alternative since I don't have any identified consumables that would help. So in that circumstance, I would pillar dance and cross my fingers (or pillar dance before the hobgoblin knocked me down to 1hp, which isn't really an argument for randomized energy).

[mikee]

You're still discouraged to get into that situation. (Don't forget that monsters do not spawn adjacent to you.)

[basil]

(Don't forget that monsters do not spawn adjacent to you.)

Neat.

You're still discouraged to get into that situation.

That's what I was driving at, yes. Even without energy randomization, the longer you try to shenanigan dangerous things to death via pillar dancing, the more likely something unexpected will happen and kill you: something fast notices you, something blocks off your pillar dancing route, a gnoll flings a net, OOD if you do it way too long, whatever. So energy randomization just serves to penalize things that are their own punishment.

[TheDefiniteArticle]

You obviously start pillar dancing before you are at 1 HP to begin with. Why on earth are you waiting to use optimal behavior until after you get hit in the face? Or are you telling me that the purpose of random energy is to stop BAD players from pillar dancing? (that explanation actually kinda-sorta makes sense if I squint real hard)

Your other example does not even TRY to address my concerns. All you've done is restate the problem another way: In the best possible scenario for random energy, all it is doing is raising the minimum amount of HP necessary to pillar dance, not actually deterring the behavior. And obviously your scenario is a wacky ideal world of frictionless surfaces that no doubt made a physics teacher somewhere quiver.

[Sandman25]

You obviously start pillar dancing before you are at 1 HP to begin with. Why on earth are you waiting to use optimal behavior until after you get hit in the face? Or are you telling me that the purpose of random energy is to stop BAD players from pillar dancing? (that explanation actually kinda-sorta makes sense if I squint real hard)

Your other example does not even TRY to address my concerns. All you've done is restate the problem another way: In the best possible scenario for random energy, all it is doing is raising the minimum amount of HP necessary to pillar dance, not actually deterring the behavior. And obviously your scenario is a wacky ideal world of frictionless surfaces that no doubt made a physics teacher somewhere quiver.

Yes, I obviously start pillar dancing before I am at 1 HP, that's why I almost never die on D1 to normal speed monsters (or never?, I can't recall any deaths now). It's obvious (let me know if I am wrong here) that energy randomization increases danger.

In the best possible scenario for random energy, all it is doing is disabling pillar dancing altogether.

[Jeff]

Just came back to Crawl after a 5-year layoff. I really like most of the changes, but random energy is one of the worst dcss design decisions ever imo. All it does is create random early game deaths and force even more conservative behaviour because you can't run away reliably.

[Patashu]

Just came back to Crawl after a 5-year layoff. I really like most of the changes, but random energy is one of the worst dcss design decisions ever imo. All it does is create random early game deaths and force even more conservative behaviour because you can't run away reliably.

You can still run away fine.

Check out the calculator I made in an earlier post in this topic - unless you play extremely unconservatively and the next doublemove will kill you, you can run away from things just fine.

Technically it's possible that a monster could doublemove you every 10 turns and kill you guaranteed - but technically it's possible that every orc priest in the game will smite you for 17 damage every turn ever, too. It's a roguelike, not a puzzle - everything you do is a percentage play to some extent or another.

(You can be against randomized energy behaviour for other reasons, like the ones duvessa and crate usually bring up in these kinds of threads - but it doesn't make the game unfairly risky.)

[Jeff]

Just came back to Crawl after a 5-year layoff. I really like most of the changes, but random energy is one of the worst dcss design decisions ever imo. All it does is create random early game deaths and force even more conservative behaviour because you can't run away reliably.

Check out the calculator I made in an earlier post in this topic - unless you play extremely unconservatively and the next doublemove will kill you, you can run away from things just fine.

In the early game it's very easy to be in situations where a double move can kill you.

While I'm not one of the best players around these days, I'm far from "unconservative." I have a decent resume with 35.5% winrate and 10-game and 6-game win streaks.

[Eyesburn]

Make it turn-based with action points, like Fallout 1/2

[Wahaha]

In the early game it's very easy to be in situations where a double move can kill you.

Can you give an example? A situation I can think of is if you're fighting something and you're 1 hit away from death and try to walk away to regen health. That's not unfair.
A situation that is unfair is if an ogre appears 1 tile away from you and as you move to the stairs it moves adjacent to you. But this is rare, and doesn't always lead to a death because you can just keep walking away until the ogre loses energy or hits you once and falls 1 tile behind. So I'm wondering what kind of situation you're thinking of. I think there is very little difference between energy randomization and no ER either way.

[Jeff]

A situation that is unfair is if an ogre appears 1 tile away from you and as you move to the stairs it moves adjacent to you. But this is rare, and doesn't always lead to a death because you can just keep walking away until the ogre loses energy or hits you once and falls 1 tile behind. So I'm wondering what kind of situation you're thinking of. I think there is very little difference between energy randomization and no ER either way.

I noticed it most with groups of enemies like gnolls, orcs and wights. If any of the priests/wizards or a gnoll/orc warrior/wight with polearm get a double move you can die easily at low levels, whereas in the past you could usually run away from them more reliably.

The early game is dangerous enough without random energy.

[fisheye]

As a new player who has never played crawl without energy randomization, one of the benefits of energy randomization is pedagogical - you get to learn really early on that planning on doing extended pillardancing simply doesn't work if you have the same speed as the enemy and they're too close. Once you get that beaten into you by losing characters to randomization, it changes your tactical perception somewhat.

Without randomization, pillardancing would actually be quite effective in the early game, and then you'd have to break that bad habit in the mid or late game if you ever want to ascend a character, which would make the game harder to learn (although easier to play).

[mikee]

Hi, Jeff. Nice to see you again.

[mikee]

As a new player who has never played crawl without energy randomization, one of the benefits of energy randomization is pedagogical - you get to learn really early on that planning on doing extended pillardancing simply doesn't work if you have the same speed as the enemy and they're too close. Once you get that beaten into you by losing characters to randomization, it changes your tactical perception somewhat.

Without randomization, pillardancing would actually be quite effective in the early game, and then you'd have to break that bad habit in the mid or late game if you ever want to ascend a character, which would make the game harder to learn (although easier to play).

There are some good posts in this thread about why that's not really the result of randomized energy that you should read if you haven't already.

But also I don't think that it's possible to effectively discuss randomized energy on tavern. As can be seen in this and every other thread on it, it has a lot of effects that many people don't realize and aren't aware that they don't realize. And discussions like this one tend to get bogged down in different details that different people think are essential to randomized energy.

[Jeff]

Hi, Jeff. Nice to see you again.

Hey, thanks. It's been a long time, haha.

[Sandman25]

But also I don't think that it's possible to effectively discuss randomized energy on tavern. As can be seen in this and every other thread on it, it has a lot of effects that many people don't realize and aren't aware that they don't realize. And discussions like this one tend to get bogged down in different details that different people think are essential to randomized energy.

Could you please list them for clarity?
1) Retreating from 1 tile away monsters until they get double move and become adjacent (not very important IMHO, most players avoid extremely dangerous fights and it has a drawback of randomly getting another dangerous monsters in LoS, too boring vs trivial monsters)
2) Retreating from adjacent monster while waiting for 0.9 move to be able to escape via stairs (without ER we get another issue - kiting monster until it misses player and player hits the monster or kiting monster when we regenerate 1 hit damage faster than the monster does)
3) When 2 monsters pursue player who entered a corridor 2 turns ago one of them randomly can overrun other and enter the corridor instead (it is good as it decreases predictability)
Before

  Code:

..###
BA.@#
..###

After

  Code:

..###
.AB@#
..###

4) Famous kill when monster moves and attacks (this is arguably bad)
5) The same as 4 but the player was on stairs and ordered to use them (I died this way to a hydra who was 1 tile away) (this is arguably good despite similar to 4 as it decreases importance of stairs which many players dislike)
More?

[Jeff]

There is a lot of discussion on energy randomization here:

https://crawl.develz.org/wiki/doku.php? ... domization

I'm not sure where the best venue to discuss reverting this change or if the cause is completely hopeless, but I agree with Stabwound that this was the worst change in the history of DCSS development.

[starhawk]

Energy randomisation definitely doesn't make getting Abyssed any easier. Aaargh.

[McGeekster]

Energy randomisation definitely doesn't make getting Abyssed any easier. Aaargh.

I agree. I've had several characters banished after walking into a room with a Deep Elf Priest and getting banished before you can say "Lugonu".

[Sar]

Monsters acting first when you move in their line of sight is not a result of energy randomization. Also, deep elf priests don't banish, sorcerers and demonologists do.