Some observations from a DM returnee

[pjs]

There don't appear to be too many new topics, so here goes a bit a bit longer one.

First a bit of nostalgia. As a small kid we got Atari ST 1040 in 1986. This was disappointing in the first half a dozen years as the friends had mostly C64's (and in 1990 or so a few Amigas) and there were not too many games for ST. But we got the original Dungeon Master from the store in 1988 or 1989. For a 10-year kid, the game was a bit scary and challenging. I think I got stuck the first time on the magenta worms level because those were so difficult and kept regenerating or couldn't figure out how to deal with the ghost. There wasn't access to maps, walkthroughs etc. back then. I don't recall if I got past that. At the most a friend went as far as the knights level and got stuck there (didn't find the way forward, too difficult opponents). Later on PC, I played the game through, with maps and walkthrough, a couple of times in late 1990's and early 2000's.

I just replayed Eye of the Beholder games and got the inspiration to play the grand master again with CSBwin (Linux 64-bit). I can't help wondering how advanced the game mechanics of DM were compared to EOB or other AD&D games - or any other computer RPG games such as Ultima for that matter. In DM, all the different skills and various ways stats, skills and equipment affected the game in various ways - you didn't really need to understand any of it to play the game (though some understanding of visible stats is beneficial), but as a former RPG player, it's quite breathtaking.

I tried to find resources where the different qualities of champions would be evaluated, and listings for "best" parties. The closest I got was a thread here five years ago about a community vote on best and worst characters to have (https://www.playbuzz.com/davegb10/dunge ... -selection). Obviously you can manage with all characters if you want. But for me I wanted to use 'best' characters (for some definition of 'best'). It would have also been interesting to see different good party compositions for different playing styles (e.g. minimal training, dedicated front line fighters, all-around good characters, max strength for lugging along all the loot, etc.).

I ended up choosing Chani, Hisssa, Gothmog and Linflas. Chani worked very well with Hissa, because giving the moonstone to Hisssa, after a few tries of war cries and wizard/priest level ups, Hisssa had a decent amount of mana already and over 100 at the end of the game. Actually Hisssa got to 60 mana before Gothmog, due to me focusing so much on casting wizard and priest spells with Hissa(!!).

I ended up training along the way and died a couple of times from trying to train also with mummies or trolins, which can be very deadly, instead of just rockpiles and screamers. Kids, don't try it in the beginning at home. The main training happened in the screamer room (and walking back and forth a couple of times on the same level) until everyone adept at fighter and ninja, expert in priest; the wizard was whatever had come by by then (about adept). Less would have also been sufficient, but I would definitely suggest training at least 3x craftsman (which takes only very little time) or artisan, otherwise going forward in the dungeon could be challenging. I think getting expert at priest took about 15 minutes of war cries per character by pressing down the keys while doing something else. War cries, punching and jabbing with rapier. There is good general information about skills and training at http://dmweb.free.fr/?q=node/691, but I don't recall coming up with a comprehensive summary or a table of the most time-efficient ways of training various skills. I suppose those are the best and easiest in the beginning unless you have the rope.

With this kind of training strategy, I could have also chosen a different party, maybe characters with more strength than Gothmog and Linflas, because everyone essentially got to the similar HP/stamina/mana stats and the key thing was other statistics of the characters. But at least so far it is not yet obvious to me what would be the best compositions of stats and skills. Strenght might also be overrated if you get comfortable with the fact that you really don't need to lug around 4 chests full of food when you descend the dungeon and you can leave behind most of the items you are not longer going to need.

I also read up the threads on various training strategies. I think especially with CSBwin and keyboard mappings, the screamer room is far superior compared to the rat nest. While you get 50% more experience (the multiplier 3 vs 2) in the rat level, the difficulty and simplicity is very different. There was also some discussion of training with a single black flame at level 12, which would be multiplier 6 so 3x more efficient than the screamer room. You could get there before battling knights and oitus at the end. So if you wanted to train less in the screamer room but want to beef up before then end, it could be a good choice, though you may need to be more careful than with screamers. The unfortunate thing is that the flame didn't seem to regenerate as easily as the screamers, if you ended up killing it with vorpal blade's disrupt (walking back and forth over it, even going upstairs and coming back again didn't immediately seem to regenerate it though I thought it should).

At the finish, 3/4 of characters were at the first level of wizard mastery, 1/4 at expert. Two typical frontline guys, Hissa and Linflas, had obtained expert fighter. Otherwise there hadn't been level-ups since the screamer room. This suggests that training up to adept (or expert in priest) was a bit excessive..

I wonder if I will try again at some point with a somewhat different party and maybe train less with the screamers and see how it affects the difficulty in the rest of the game. But maybe I will try Chaos Strikes Back first. There were some issues with it, but I think I'll post a separate thread on it on the CSBwin subforum.

CSBwin apparently tries to rather faithfully follow Atari ST version. I noticed some differences to the PC version, e.g. no music at the start screen and you could not drink directly from the fountains (also documented in http://dmweb.free.fr/?q=node/247). The one feature that is completely missing in all versions(?) that I would have found useful is a way to rearrange your champions so your current front line guys would be in slots 1 and 2. Because I ended up choosing the champions in wrong order, this would end up in a bit of hassle when attacking and doing actions.

[ChristopheF]

the flame didn't seem to regenerate as easily as the screamers

Each creature generator is disabled for a configurable delay after being activated. The black flames at the entrance of the large hall on level 12 have a delay of 1216 ticks (about 200 seconds). You can see that when editing the dungeon file with CSBuild.

a way to rearrange your champions

That is the purpose of the four champion icons in the top right corner. To swap two champions, click on the first one, then click on the second one.

[Prince of Elves]

a way to rearrange your champions

That is the purpose of the four champion icons in the top right corner. To swap two champions, click on the first one, then click on the second one.

I think it was about rearranging their "number" (and thus their fighting slots) rather than the battle position. It can be very annoying if you end up with character 1 and 4 on the front line, it's way more convenient to have 1 and 2 in the front and 3 and 4 in the back or vice versa (taking care to always swap them in doubles if I want to give the back row more fighter training).

[Estoroth]

Yes I think it is important to anticipate who will be the front line fighters inside the party before ressurect/reincarne champions. I'm agree with Prince of Elves, it can be very annoying to click on action button 1 and 4 compared to 1/2 or 3/4. You can move champions but not the action button.

I have replayed Dungeon Master recently with CSBwin and honestly I don't see difference with the original DM, for me it is the better way for play DM or CSB in modern computer. You get some nice little additionals feature not available in original game as logs with each type of monsters you have killed.
For my party I have choosen
1_Hissssa Lizar Of Makan
2_Halk The Barbarian
3_Sonja She Devil
4_ Stamm Bladecaster

Starting game with character with no or very low mana is really not an problem as soon you gain some mana point in Priest skill unsing War Cry when a monster hit you. I have no used a lot of magic during my game instead maybe doing magic torch !
I have the feeling that characters with good physical ability in the beginning of the game can become powerful in magic with the appropriate training while the opposite is not necessarily true.

I had no difficulty completing the game with my group mentioned above.

My recommandation if you want to restart DM with a new party, pickup minimum two of the following champions I have already taken in my group for your couple of slot 1/2 or 3/4 and use reincarnate option.
You will be guaranteed to have powerful melee characters able to wear the complete set of heavy armor at the end of the game.

[Ameena]

My usual "group" these days is just a pair - reincarnated (and therefore renamed) Hissssa and Wuuf. After I've picked them up and headed down to Level Two, I refuse to move from the bottom of the stairs until I've maxed the light level (and I also refuse to use physical torches since they burn out and using them means I'm not practising Wizzy skills to create magical light). Usually takes a minute or two. Hissssa can cast Lo and then after regenning the mana point back can also cast Ful, so it is possible to get off a light spell or two there. And Ya potions are the cheapest spell in the game so I can conjure a few of those here and there too as I go, pelting around with the keyboard and wearing out both their stamina super fast . But I don't really bother with stopping to train/grind. Not any more - bored of it and can manage well enough without. I take these two into CSB and have completed it with them at least once so that's fine .

[ChristopheF]

Sorry I misunderstood the issue. If you want to fight really fast, you should avoid using the mouse and use CSBwin keyboard shortcuts.

On a QWERTY keyboard, you may press Q, A or Z (first column of letters on the keyboard) to open the action menu of the first champion. Then press Q for the first action, A for the second, Z for the third. So if you press the A key twice on the keyboard, this will trigger the second action of champion 1.
Same rules apply for the other 3 champions with keys in the second, third and fourth columns on the keyboard.

You can also remap all keyboard shortcuts to your convenience in the configuration file you prefer anything else.

Magic menu opens with F1, F2, F3 and F4 and then use numbers 1-6 to select magic symbols and spacebar to cast the spell.

[jayrshaw]

Hi pjs / all,

I believe I first played Dungeon Master on the SNES when I was a teenager, but I never got very far into the game until I replayed it during graduate school in the early 2000s. I didn't actually complete the game (or get any further than the floor with the Scorpions) until I finally made a real effort to do so last year. I always thought the skill/experience system in this game was interesting, though, and I also thought it was cool that the sudden appearance of a monster or group of monsters in the dungeon halls could always actually surprise me a bit - I still recall thinking the first Screamers I encountered in the dungeon were really scary!

During my 2020 playthrough, I used the same party that Estoroth did (Hissa, Halk, Sonja, and Stamm), and I am still generally of the opinion that Strength is the most important of the character statistics (although I do agree with your point about high Strength levels being far less necessary if you decide not to wear heavier armor and stick to grabbing only the most essential items in the dungeon). Choosing particular Champions is generally going to matter a lot more for a playthrough where you don't intend to train your Champions a lot and/or don't intend to use stat-buffing potions to artificially raise your Champions' Wisdom, Strength, Dexterity, and/or Vitality than it is for a playthrough where you do intend to do these things.

If you intend to train your Champions a lot, it really doesn't matter a great deal which ones you choose because (a) all of the Champions will end up with high (and, with the exception of Stamina, relatively similar) Maximum Health, Stamina, and Mana regardless [starting Maximum Stamina actually matters much more for determining characters' Maximum Stamina later in the game than starting Maximum Health and Mana do for determining Maximum Health and Mana later in the game, unless you intend to train everyone's Maximum Stamina all the way to 999]; (b) because you can buff most of the other character statistics to very high levels using multiple layers of Lo level Dane, Ku, Ros, and Neta potions; and (c) because you can generally make your characters completely immune to enemy fireballs using the Fireshield spell (making the Anti-Fire statistic largely irrelevant as long as you are willing to keep a Fireshield spell up in situations where you believe your Champions might be hit by fireballs).

I'm guessing there are probably certain Champions which would work best for playthroughs where you don't intend to train your characters very much, but I don't know which ones they are since that isn't the way I played through the game - I guess two stronger characters in the front and two characters with high Mana and Wisdom in the back would probably work as well as anything. Regardless, you would have to be insane not to train all of your characters in all four of their skills to at least some extent since the lower skill levels are very easy to obtain and because they really improve your characters' capabilities and survivability (the only real exception to this would be if you are attempting a low-level speedrun). The most experienced players performing low-level speedruns of this game seem to prefer going through most of the game using Stamm by himself (I believe this preference is mainly because of Stamm's high stamina and stamina recovery).

For training, I usually have my characters with no / very low mana use War Cry against the first Screamers I encounter in the game until they have earned enough Healer levels and Mana to begin practicing spells. I then gain several levels for all of my characters in all of their skills using Screamers at the point in the game where you first obtain an Empty Flask (the Empty Flask is crucial for healing your characters periodically). The higher Health, Stamina, and Mana from this training at the very beginning of the game really help with the levels of the dungeon you need to get through before gaining access to the Screamer Room on the Worm level of the dungeon. I did all of the rest of my training in the Screamer Room; while I believe the Black Flame room is technically the fastest way to gain experience, you can train in the Screamer Room with far less hassle than you can in that location or the Rat Room - the Screamer Room also has the advantage of being accessible much earlier in the game, which means the training you do there will be useful even on relatively early floors of the dungeon you have to go through prior to gaining access to the Rat Room (and especially the Black Flame room).

Anyway, thanks for the thoughtful post and hope you and everyone else here has a Happy New Year!


--Jay

[pjs]

I wonder if I will try again at some point with a somewhat different party and maybe train less with the screamers and see how it affects the difficulty in the rest of the game. But maybe I will try Chaos Strikes Back first. There were some issues with it, but I think I'll post a separate thread on it on the CSBwin subforum.

FWIW, I just played through CSB for the first time using the dmweb maps and walkthrough (no patience to try to explore it blind - certainly could be a very big challenge if that's your cup of tea); about 15 hours of play time. The front line guys got one fighter levelup, everyone ended up 2nd level master wizard and one got first master priest; so at this point the level-ups are rare unless you grind it out. CSB seems to have a higher replayability due to all the options available. Already thinking of retrying for a better strategy, but I'll try to hold back for a while

[Ameena]

It's perfectly possible to finish both games without bothering to really stop and train anywhere, also using no potions at all - I can't be arsed with that . I use just two characters and I picked them mainly because they look cool (no boring-looking human types ) and also Hissssa at least has the mana to cast stuff without needing an external boost from holding a wand or whatever. Also, as a bonus, Wuuf starts with a flask, not that those are hard to come by on the next level anyway, and I just get them one each which they keep in their off-hands the whole game. It's not as if shields do much .

[Prince of Elves]

What do you need a flask in hand for if you don't do potions "at all"?

[Ameena]

Oh sorry, I meant the buff potions. I use health and cure poison potions, and early on stamina potions but that's it .

[pjs]

Just for the fun of it, I played DM through again, this time with two characters, Hissssa and Chani, in order to try with just two in CSB (next step after that: reincarnated characters in CSB). I did train, however, this time also with the black flame at 12.

Punching the black flame to get ninja from expert to master 1 took about 9-10 minutes, rapier jabbing the black flame to get from fighter master 1 to master 2 at fighter took 12 and 14 minutes, war crying for priest 1->2 about 31 minutes. I wouldn't have bothered with training to master 2, but I wanted to see how long it would take "on the clean slate".

You can't really compare training different skills, but you get the picture. All other things being equal, you can calculate (because getting to the next level requires double the amount of training than previous level) that getting ninja archmaester from master 1 with this method would take 10*126 minutes = 21 hours, fighter 13.7 hours and priest 33 hours. Good luck!

A few other observations on this training: after done with punching the black flame, once finished with training I had to kill it about 10 times with vorpal blades, it immediately resurrected. I wonder what that was all about. War crying seemed to make the flame afraid and it did not attack at all.

That also made me wonder about the simplified formula (http://dmweb.free.fr/?q=node/692), according to which you get 2x (or ninja/fighter: 4x) experience when a monster has attacked you within a certain time. I wonder if this is correct. I could not find this in the source code and in the very thorough mechanics thread (viewtopic.php?f=27&t=31345) it would appear that those bonuses only apply for the experience you get when monsters hit you - not the experience you get when you attack the monster. (You also get extra xp if you hit monsters.) If this is true, there may not be such a big difference whether you are training the skills with monsters or not.

[ChristopheF]

The formula described at http://dmweb.free.fr/?q=node/692 is implemented in the function named 'F0304_CHAMPION_AddSkillExperience' in CHAMPION.C in the ReDMCSB source code here: http://dmweb.free.fr/Stuff/ReDMCSB_WIP20210206.7z

[Paul Stevens]

A few other observations on this training: after done with punching the black flame, once finished with training I had to kill it about 10 times with vorpal blades, it immediately resurrected. I wonder what that was all about.

It sounds as if several 'regenerates' were queued.

When a regenerate is triggered, if the cell is already occupied, then the
action is queued and occurs as soon as the cell becomes empty.
Since Black Flames do not move, this can happen quite easily.

[pjs]

The formula described at http://dmweb.free.fr/?q=node/692 is implemented in the function named 'F0304_CHAMPION_AddSkillExperience' in CHAMPION.C in the ReDMCSB source code here: http://dmweb.free.fr/Stuff/ReDMCSB_WIP20210206.7z

Thanks, found it now. If I understand correctly, the monster attacking condition is satisfied (at least) when the monster is in the next square and is "chasing you" in order to attack. In other words, dodging around the monster (and scoring a hit yourself now and then) satisfies the criteria of "monster attacking" all the time, even though the monster does not "swing" at you or score a hit through protections (e.g. a shield). On the other hand, if the monster is afraid or tries to run away (e.g. when war crying the black flame in the previous example), I suspect the monster is not attacking then - so it is not be sufficient to be on the next square if you want to get 2x/4x experience.

[Paul Stevens]

If I understand correctly, the monster attacking condition is satisfied (at least) when the monster is in the next square and is "chasing you" in order to attack

You certainly may be correct. But if I had to guess, I would say that
the bonus to experience depends on a monster having actually attacked
you in the previous few seconds (whatever that time period is).

[ChristopheF]

If I understand correctly, the monster attacking condition is satisfied (at least) when the monster is in the next square and is "chasing you" in order to attack

No. Where did you get that? The criteria in the source code is only about the time elapsed since the last time a creature attacked a champion.

Where did you see 4x experience ?

[jayrshaw]

Hi ChristopheF,

I think pjs is referring to the following two sentences on the "Experience and Training" page of the Dungeon Master Encyclopaedia:

1. "If the Improved Skill associated to the action you perform is between 04 and 11 (hidden Fighter and Ninja skills) and if no creature has attacked the party for at least 150 clock ticks (25 seconds), then Experience is divided by 2."

2. "If the Improved Skill associated to the action you perform is above 04 (any hidden skill) and if a creature has attacked the party less then 25 clock ticks ago (about 4.17 seconds), then Experience is multiplied by 2."


If I'm reading this correctly, it sounds like if you fail both of the two tests described above when training a Fighter or Ninja skill (e.g., you have not been near an enemy for more than a minute), the amount of experience your character earns will be 1/4 amount that it would be if you passed both tests (e.g., if an enemy just attacked you within the past second), which is probably where pjs' reference to the 4x experience multiplier comes from.

By the way, this is anecdotal, but my observation is that the penalty for being out of combat is even more severe than this for Fighter and Ninja skills in the SNES version of Dungeon Master. I specifically recall a situation during my playthrough last year when I saved the game while fighting two Magenta Worms right before one of my characters was about to gain a Fighter level; I ended up finding myself caught in a bad RNG loop where I couldn't maximize the character's stat gains when earning the level while fighting the Worms - I ultimately retreated as far as I could from the Worms and then attempted to earn the level simply by using Fighter moves into the air when no enemies were in the immediate vicinity; I quickly discovered that it took *far* more repetitions to gain the level in this manner than it did when I was facing the Worms. This is consistent with other informal experiments I have conducted in the SNES version of the game where I attempted to train Fighter and/or Ninja skills without an enemy present, but I would have to test it in a more disciplined manner to confirm the phenomenon.


--Jay

[pjs]

If I understand correctly, the monster attacking condition is satisfied (at least) when the monster is in the next square and is "chasing you" in order to attack

No. Where did you get that? The criteria in the source code is only about the time elapsed since the last time a creature attacked a champion.

Where did you see 4x experience ?

My bad. I was looking at where G0361_l_LastCreatureAttackTime is updated (GROUP.C: G0361_l_LastCreatureAttackTime()). But apparently it is only called in the code branch where the creature actually attacks (and an attack sound is played). However, if I understand correctly, the timestamp update is not conditional on whether the attack hits or misses (or whether it causes damage or not, e.g. due to a fire shield).

As Jay mentioned, the cumulative effect of halving and doubling gained experience means that ninja/fighter experience is 4x in total if there has been attack within 4 seconds compared to the case when there hasn't been an attack for 25 seconds.

This suggests you get lousy xp, especially ninja/fighter xp, if you just sidestep monsters because they don't get to attack you. In contrast, attacking screamers, rock piles, flames, etc. is much more efficient for gaining xp, because you actually want the monsters to try to score weak hits at you periodically to beef up the experience you will get from your own actions.

But this may also suggest why war crying the black flame at L12 was inefficient: the black flame didn't try to attack at all. In contrast, screamers and many other monsters do try to attack while war crying. The cumulative effect of this is that war crying black flame at L12 is only 50% more efficient than training in the screamer room even though the difference in experience multipliers is 3x. Likewise, training in the rat room is very likely in total far less efficient than screamer room unless you allow the rats to attack you all the time. So the screamer room is actually rather good for training.

[jayrshaw]

Despite the fact that it isn't my preferred method (due to inconvenience and lack of availability until very late in the dungeon), I would still guess that the Black Flame room is going to be the fastest way of training all four skills. For Fighter levels, the easiest way is probably to repeatedly Jab at the Black Flame with the Rapier or the Bolt Blade; for Ninja levels, you would probably just want to repeatedly Stab at the Black Flame with a Dagger or similar weapon; for Healer levels, you would just want to repeatedly create a ton of Ma (Stamina) potions while the Black Flame is attacking you; and for Wizard levels, the fastest way is probably to repeatedly cast the Poison Cloud spell at the Black Flame or towards another safe area while the Black Flame is attacking you (although I would personally use Magic Torch instead, since it gives nearly as much experience per mana point as Poison Cloud and also teaches the hidden Fire subskill, which is more useful in most situations than the hidden Water subskill the Poison Cloud spell teaches).


--Jay

[pjs]

An optimal training method also may also depend a bit on whether you can just press down a key (unless you automate combinations). From that point of view, war crying is very straightforward. But if this is not a factor, based on http://dmweb.free.fr/?q=node/195 you can get 48 xp for MON YA spell with fatigue of 15 (ticks i.e. 1/6 seconds) in contrast to war cry, which gives 7 xp with 3 fatigue. So 48 vs 35 xp in the same amount of time, plus possibly the double if the monster doesn't get afraid and attacks within 4 seconds. For significantly weaker YA spells the xp value per time used is worse.

Looking at http://dmweb.free.fr/?q=node/690, punching gives 8 xp with fatigue of 1 while stabbing with a dagger gives 15 xp with fatigue of 5. This means that punching is far more time efficient for gaining xp, because you can do 5 punches with total of 40 xp in the same time you can do one stab with a dagger. Therefore, punching is by far the best way for training ninja skills.

I just noticed that fluxcage gives 12 xp with fatigue 2. Spamming fluxcage at a black flame occurred to me as an idea for training wizard skills, but this can only be done at the very end. Contrast to max-power torch spell gives 28 xp with fatigue of 15, so fluxcaging is on the face value over 3x more efficient than casting torch spells. However, the firestaff gives you 2 skill levels, and for this reason I have never worn skills enhancing items when you training. It makes it more difficult to get to new levels. But now a thought occurred to me. Don't you still get the same xp even if you don't level up, so do you get the level immediately after obtaining next XP once you remove it (ie. you reach a new "original" level with the XP you already gained with the item). Anyone know?

As a concrete example, consider an expert wizard sitting at 100 000 xp (next level is LO master at 128 000 xp, so 28 000 xp away). When you have firestaff at hand, you are are UM master with the next level up at 512 000 xp. If you train 50 000 xp with the firestaff and remove it, once you gain 1 additional xp (the level ups are checked only when you get additional experience I think), I suppose you get LO master immediately? Or if you trained up to 256 000 xp with the firestaff, you would get two levels immediately after getting xp two times after you removed it?

[ChristopheF]

Check function F0303_CHAMPION_GetSkillLevel in CHAMPION.C.
It is called twice in F0304_CHAMPION_AddSkillExperience (before and after updating the experience value of the champion) with a flag to ignore object modifiers. This means the Firestaff and other items have no effect on skill level up.
In your example, as soon as you get 28000 xp, your champion will level up no matter what object is in hand.

[jayrshaw]

Hey pjs,

Not sure where you are getting the 48 experience point amount for a full power MA (Stamina) Potion spell or the 28 experience point amount for a full power Magic Torch spell from the "Dungeon Master and Chaos Strikes Back Spells" page of the Dungeon Master Encylopaedia you linked to, but I don't think those are the correct numbers. Check out the table Maven posts in the following thread:

viewtopic.php?f=27&t=22395&p=96760&hili ... nce#p96760


I experimented with this extensively during 2020 in the SNES version of the game, and I can confirm that you can gain Healer levels *far* faster by repeatedly creating full power MA (Stamina) Potions while being attacked by a Screamer than you can by repeatedly using War Cry against a Screamer in the exact same location within the dungeon. I was gaining Healer levels exclusively by using War Cry for a long period of time before reading Maven's post and trying out the MA potions (I had been saving all of my mana to train Wizard levels using Magic Torch) - I was shocked at how much faster my characters could gain Healer levels by creating the MA potions than they could by using War Cry. I actually trained my characters all the way to PAL Master Healer (d Master in the SNES version of the game) using the potions, but I only had the tolerance to train them to ON Master Fighter (b Master in the SNES version of the game) using War Cry, despite the fact that you earn a bit *more* Fighter experience using War Cry than you do Healer experience.

I also tested experience earned using Punch, Kick, and Stab over time in the SNES version of the game. I would not take the "fatigue" value you see on the Dungeon Master Encyclopaedia at face value, because it isn't always an exact indicator of what a character's recovery time will be in various particular situations he/she might be in. If I recall correctly, the recovery time is significantly faster for performing an action into the air or a wall than it is for successfully performing the same action against an actual enemy. In addition, in the SNES version of the game, all actions are performed much faster when you are in the character inventory menu (and slightly faster than that if you are in the inventory menu of a character who isn't the character executing the action). I found that I could execute roughly 72 Stabs per minute when attacking a wall from a second character's inventory menu with a Screamer flanking me in the SNES version of the game, which is not significantly slower than Punch in the circumstances I described (assuming you are pressing the controller button manually and not using a turbo controller). Also note that an action can be quite a bit slower than an alternative action and still be the fastest way of gaining experience if you are willing to alternate between two characters performing the action (i.e., have Character B perform the action while Character A is recovering and then switch back to Character A). Some of this will probably vary if you have and use a turbo controller (I was not doing this during my playthrough, but I could see how this would definitely be the most convenient way of training if you happen to have a turbo controller on hand).

Now that I'm looking more closely, I see the Encylopaedia states that a "Fatigue" point is supposed to equate to roughly 1/6 of one second. This can't be the right amount for when you are actually attacking an enemy directly, because that would mean that you would recover from attacking an enemy with Chop (8 Fatigue) in ~1.33 seconds - the recovery was definitely not that fast for me in the SNES version of the game when I was using Chop against enemies using an Axe. Strangely enough, the fatigue formula does sound like it generally matches the speed at which I was executing Stab against a wall from another character's inventory menu. Regardless, I don't think it was technically possible to execute anywhere near 6 Punches in one second in the SNES version of Dungeon Master I was playing during 2020 (despite the fact that Punch's "fatigue" value of 1 theoretically equates to a recovery time of 1/6 of one second); perhaps this varies for different versions of the game, but in the SNES version of the game, I don't think the game's engine will process actions that quickly even if it should theoretically be technically possible based on Punch's "fatigue" value - I guess you would need to try testing with a turbo controller to see how many Punches a character can actually execute in practice over a given period of time to have a decent feel for it, but I can say that the 72 Stabs per minute I was executing when training my characters in 2020 was probably about the limit of what I could personally do over an extended period of time by repeatedly pressing the controller button manually.....


--Jay

[MasterWuuf]

Nice to hear the early experiences you went through with your Atari. I bought an ST and later the Mega??? ST. I eventually purchased a colored monitor ($350.00 or more). The better graphics were great. After all these years, Dungeon Master is still pretty decent, even when I compare it to the newer games. When I found DM on PC, I nearly tore the box up, getting it into my hands (ah yes, my old pal, Ovaltine...er, Dungeon Master).

Great memories and the additions to the DM family, including Conflux, were and are a great treasure to me.

[pjs]

Not sure where you are getting the 48 experience point amount for a full power MA (Stamina) Potion spell

Correct. I misunderstood the website docs completely. The mentioned values only form a small subset of the basis based on which spell experience is calculated. The formula is much more complicated than that and at least roughly corresponds to the figures posted above. So the experience for spellcasting is by far more efficient than war crying or possibly also disrupt/fluxcage actions unless the key point is automating key presses using hotkeys.

Now that I'm looking more closely, I see the Encylopaedia states that a "Fatigue" point is supposed to equate to roughly 1/6 of one second.

Looking at the reference code, experience is halved for failed melee actions and in Atari ST 1.1 fatigue is also halved for failed melee actions. These likely also apply to essentially all other versions. I think this applies equally to missing a monster and "punching the air" on purpose. This might explain at least some of the things you are seeing. With CSBwin and keyboard shortcuts it is possible to make use of quickly repeatable actions (but I'm not fully sure what is the actual cap there, ie. whether it can deal 6 or 12 punches in a second - maybe not). With different kind of controllers there are obvious problems for repeated actions. The consequence of this is that making repeated actions while not facing the monster (as long as the monster is attacking you) vs trying to hit the monster is not so bad, this depends on how quickly you are physically able to repeat the action (compared to the fatigue of the action) and how often you would hit the monster if you tried.

[jayrshaw]

Some of the notes from my experimentation in 2020 is in several posts within the following "Best Way to Level?" thread:

viewtopic.php?f=27&t=31321&start=30

The gist of it is that the quirk in the SNES version of the game which causes actions executed through the character action menu to be roughly twice as fast when executed in the character inventory menu than when they are executed in the normal dungeon view of the game has a huge impact on which training strategy will earn you the most experience over time. This phenomenon generally tends to make it a far better strategy to either use attacks which do not damage the enemy or to execute the attacks at a wall or the air while the enemy is flanking you than it is to try using attacks which actually damage the enemy (despite the fact that the latter type of attack gives more experience per repetition than the former) - you can simply execute so many repetitions from the character inventory menu over time compared to the number you could execute attacking the enemy directly from the normal dungeon view that you generally tend to gain more experience this way; however, the real advantage comes from the fact that your characters' mana recovery rate is also twice as fast in the character inventory menu in the SNES version of the game as it is when you are in the game's normal dungeon view - as a result, training from the character inventory menu also enables you to earn Healer and Wizard levels far faster than you would be able to otherwise (this is part of the reason I was able to train my characters' Healer and Wizard levels so high compared to their Fighter and Ninja levels [the other major factor was continuous use of Dane / Wisdom Potion buffs]).

In the final page of the thread I linked to above, Automaton delves into some of the details of how the formulas for calculating experience gained from successful attacks against enemies work. The formula is a bit complex, but the gist of it is that each enemy in Dungeon Master has an "experience class" which is used in conjunction with the amount of damage you deal with your attack to help determine the experience bonus you get for dealing damage to the enemy. This would combine with the effect you mentioned about the amount of base experience a character earns from an unsuccessful attack being halved compared to the amount of base experience the character would have earned had the attack been successful; the ultimate result is that successful attacks (particularly against more difficult enemies with a higher "experience class") can potentially grant significantly more experience to a character per attack than unsuccessful attacks do. The problem (at least in the SNES version of the game) is that you generally can't deal anywhere near as many damaging hits against difficult opponents over time as you can unsuccessful ones while being continuously attacked by a harmless enemy like a Screamer.


--Jay