list of bugs and wishes

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Derakon
    Prophet
    • Dec 2009
    • 9022

    #31
    Buying a Wand of Magic Missile with your starting money is a pretty fun way to go sometimes. It's pretty expensive, which limits your other gear options substantially (and almost certainly means your first dive is one-way until you find a ?WoR), but it trivializes a surprisingly large number of fights. It's especially worthwhile for mages and priests, whose offensive options early on are very limited.

    Comment

    • Pete Mack
      Prophet
      • Apr 2007
      • 6883

      #32
      10 !oil, some ?Phase, and a !Heroism have nearly the same effect and is a lot cheaper than -MM. Dive fast, looking for high value target. Then instant CL7.

      Comment

      • Derakon
        Prophet
        • Dec 2009
        • 9022

        #33
        Oil misses more than Magic Missile fails to activate in my experience. It also has considerably less range. I will agree that your proposal is more versatile, mostly on the basis of giving you other consumables that are generally useful, but I'd give the -MM approach a shot once or twice before you shoot it down entirely.

        Comment

        • wobbly
          Prophet
          • May 2012
          • 2627

          #34
          Oil can be a bit of a disaster for a human mage, you don't have infravision so you can't actually see the doggies without light. Even with !heroism I'm not convinced a lvl 1 human mage can hit the dogs reliably enough, I think the little guys have +throwing? The hobbit certainly does.

          Comment

          • Pete Mack
            Prophet
            • Apr 2007
            • 6883

            #35
            Yeah Ive tried it. It is a high variance method: either die fast, or level up fast. Using oil doesnt work quite so fast, but it has much less downside risk.

            Comment

            • PowerWyrm
              Prophet
              • Apr 2008
              • 2986

              #36
              I never bother with oil. The chance of missing is way too high compared to shooting/throwing/magic.
              PWMAngband variant maintainer - check https://github.com/draconisPW/PWMAngband (or http://www.mangband.org/forum/viewforum.php?f=9) to learn more about this new variant!

              Comment

              • Pete Mack
                Prophet
                • Apr 2007
                • 6883

                #37
                The chance of hitting with oil is the same as with any other thrown object, and at a low level, the extra +10 to hit from !hero makes a huge difference in odds. It and -MM are the only way to take out Bullroarer at CL 1 or 2. It is useful at most once in a game. But it is a use with massive potential upside.
                That said, if you are grinding at low levels, it won't ever pay off.

                Comment

                • Grotug
                  Veteran
                  • Nov 2013
                  • 1634

                  #38
                  I have tried the MM approach but do find it so expensive. I will give oil and heroism a try next time!
                  Beginner's Guide to Angband 4.2.3 Part 1: https://www.youtube.com/watch?v=m9c9e2wMngM

                  Detailed account of my Ironman win here.

                  "My guess is that Grip and Fang have many more kills than Gothmog and Lungorthin." --Fizzix

                  Comment

                  • Pete Mack
                    Prophet
                    • Apr 2007
                    • 6883

                    #39
                    Total cost is ~60GP: 12 !oil and 1 !Hero (which you should be getting in any case in case of fear attacks.) Strategy is particularly good if you are skipping levels a la Quickband.

                    Comment

                    • Nick
                      Vanilla maintainer
                      • Apr 2007
                      • 9633

                      #40
                      OK, here's a response to the first list:

                      Originally posted by Carg
                      01 - map view often doesn't refresh correctly: you can see your own character in there a lot of times, and sometimes remains are left of beam/cone attacks. CTRL-R corrects the errors. Apparently erasing changes from the previous turn is imperfect.

                      02 - outdated dungeon.txt:
                      * "Each level has at least one up staircase and at least two down staircases. There are no exceptions to this rule." Level 99, 100 and 127 are, I believe, exceptions. But several types of special map levels might not have 2 downstairs either. Notably a maze level always seems to have exactly 1 up and one down stairs.
                      * The whole file also is based in the character interface, ignoring tiles. It should e.g. mention that with tiles you can see the visuals under '~', 'display xxx knowledge', select one item and press 'v'.
                      Filed as bugs.


                      Originally posted by Carg
                      04 - with the Shockbolt tiles I still don't see any difference between passable and impassable rubble in 4.1.3, although I read that was changed since 4.1.2.
                      Should be fixed, but doesn't seem to be - bug reopened.

                      Originally posted by Carg
                      05 - if there is a bunch of monsters, I am often in a corridor get only one attacking me, resting until the next one steps in. Their drops are piling gradually, and when there is more than one item under me, the screen gets redrawn every time because it displays the list of items under me. I wish there was at least one wait command key (of the '5', ',' or '.') that I can use to rest without picking up.
                      Making a keymap for resting one turn should work.

                      Originally posted by Carg
                      06 - why are monsters like Pukelman and Earth Elemental resistant to all attacks except the one they are atacking with themselves (acid)? That seems illogical. I suggest a different basic element vulnerability.
                      I don't necessarily see this as a problem, but will consider it with the monster list revamp.

                      Originally posted by Carg
                      07 - wish: (semi-)automatically determine character width and height based on the selected font for all terminals, like you can set "Term 0 Font Tile Size" to "Font" for terminal 0.
                      I don't think I quite understand this one.

                      Originally posted by Carg
                      08 - resolve naming inconsistency wirh wands and rods by renaming staff of light to staff of illumination
                      Sounds sensible.

                      Originally posted by Carg
                      09 - why does @m1 inscription work, but not @p1?
                      Because 'm' is the real command, and 'p' is just mapped to it. With new classes coming in it may be best to remove the 'p' command altogether.

                      Originally posted by Carg
                      10 - Wish: display monster level in monster memory (esp. with uniques). If there is not enough room, I wouldn't mind the monster symbol being removed, but I think both should fit. Especially with uniques it helps to quickly see which you might expect on your current dungeon level.
                      Monster level is displayed, once learned (from killing one, IIRC).

                      Originally posted by Carg
                      11 - Wish: change options permanently (e.g. I like show_damage) - e.g. save set like exporting autoinscribe. Every time I start a new character (I like playing several in parallel) it is off again.
                      I think it has been mentioned that you can achieve this by using an old save file. The defaults are actually currently compiled into the game; it's possible that they will be in a player-editable file at some time in the future.

                      Originally posted by Carg
                      12 - show_damage should not only show weapon damage, but also spell damage. In case of beam, ball or breath attacks the damage in the square targeted could be listed, or perhaps the nearest monster affected?
                      This has been requested before, and is problematic for player damage to monsters because spells sometimes don't have defined targets. It probably could be implemented for monster damage to the player.

                      Originally posted by Carg
                      13 - treasure balance seems off: too many weapons compared to armor pieces (you need 6 armor pieces compared to 1 weapon, so it would seem better if they occur 6x as often too, but they seem to have similar occurences currently). For artifacts my high level mage has found 25 weapons, 4 shooters, 28 armor and 4 miscellaneous (light, ring, amulet). I would also prefer not to find at least one of the *thanc daggers at one of the first levels so often; then all the fun disappears of first finding one with bonuses, then the first ego, and finally the artifacts.
                      This is a perennial issue, affected by many things. It's a constant balancing act, and will remain so. I do listen to specific complaints

                      Originally posted by Carg
                      14 - In obj-desc.spo listing the items power in the column level is confusing. Either list the min level the item can be found, like in mon-desc.spo, or call the column power, or list both. I consider the min level more useful.
                      Filed as a change to be made.

                      Originally posted by Carg
                      15 - Amulets of resistance are way too high level for their worth: since most artifact armors, shields, and even many ego armors, shields and cloaks offer the same elemental resistances, by the time you reach level 60 no one is ever going to need them. 10 levels later than resist acid en lightning, so 20, seems more appropriate. Or perhaps it should have 1 or 2 random extra resistances?
                      The usual argument for this is that they are not intended as a first source of the four basic resistances, but rather as a way of using the amulet slot to get the basics and free up other slots from having to cover them. I suspect that is still a decent argument.

                      Originally posted by Carg
                      16 - in principle it is a good thing that slow digestion is powered up, but still there is so much food found in the dungeon, not to mention loads of satisfy hunger scrolls, that even a half-troll warrior is unlikely to sacrifice a ring or amulet slot for it. Perhaps the ring and amulet should have some minor extra power, like slow poison?
                      They really only of any use early - maybe that's OK though.

                      Originally posted by Carg
                      17 - spell failure chances seem higher than the spellbook mentions. I tested this with a 21st level mage with listed spell failure of 4% on detect monsters, with no negative effects present as far as I could see. I cast it in total 335 times in a row, and had 26 failures, which amounts to more than 7% - almost twice as much as listed.
                      This is a common complaint, usually refuted by an appeal to advanced mathematical statistics

                      Originally posted by Carg
                      18 - stores will not accept unidentified armor, not even for free. (I had a robe [2,+2] {??} that wasn't accepted until I happened to find out the rune in another way. It was resist fire.) I think they should, just like unidentified negative potions, scrolls, wands, etc. are bought. This is also contrary to the description in dungeon.txt. Weapons that are partially unknown and contain a curse are also unsellable.
                      This has been noted before, I've filed it as a bug.

                      Originally posted by Carg
                      19 - Why don't shops have consecutive letters for their inventory? It is confusing and I don't see a use.
                      Answered upthread; it's to avoid conflict with shop commands.

                      I'll get onto the rest of the list soon.
                      One for the Dark Lord on his dark throne
                      In the Land of Mordor where the Shadows lie.

                      Comment

                      • fph
                        Veteran
                        • Apr 2009
                        • 1030

                        #41
                        Originally posted by Nick
                        This is a common complaint, usually refuted by an appeal to advanced mathematical statistics
                        As an expert in advanced mathematics , the binomial test for 26 failures out of 335 at 4% failure rate gives a one-sided p-value of 0.00114. This means that you can expect to observe 26 or more failures about 0.11% of the times.

                        Usually, in a single, non-repeated experiment, a result is considered statistically significant if the p-value is 0.05 or below. So yes, that seems a statistically unusually high failure rate, and worth investigating. There is usually a reporting bias, in that only 'unusual' results get reported to the devs, so the best thing would be collecting completely new data to test it.
                        Last edited by fph; October 28, 2018, 22:21.
                        --
                        Dive fast, die young, leave a high-CHA corpse.

                        Comment

                        • Pete Mack
                          Prophet
                          • Apr 2007
                          • 6883

                          #42
                          BTW: for the rest one turn keymap, I suggest 's', and make it rest 5 or 10 turns (or until disturbed.) I used to use the old search command for this all the time: wait until a nearby monster moves, or appears around a corner .

                          Comment

                          • fph
                            Veteran
                            • Apr 2009
                            • 1030

                            #43
                            Originally posted by fph
                            There is usually a reporting bias, in that only 'unusual' results get reported to the devs, so the best thing would be collecting completely new data to test it.
                            Ran some tests: created a human mage, advanced him with debug commands to lvl21 and enough int to get 4% failure rate, and cast detect monsters 335 times. Got 14 failures. p-value of 0.47. Verdict: the RNG seems to work OK. Complaint "refuted by an appeal to advanced mathematical statistics".
                            --
                            Dive fast, die young, leave a high-CHA corpse.

                            Comment

                            • Pete Mack
                              Prophet
                              • Apr 2007
                              • 6883

                              #44
                              For low probability events, a poisson distribution is generally appropriate. It always has a fat tail (because you can't have fewer than zero events.) Yes, it will eventually converge to normal (as any distribution will.) But even at expectation value E=14, while the curve looks close to normal, the tail is significantly fatter at 1.5*E. The estimate of 0.11% is too low by a significant amount.

                              Comment

                              • fph
                                Veteran
                                • Apr 2009
                                • 1030

                                #45
                                Originally posted by Pete Mack
                                For low probability events, a poisson distribution is generally appropriate. It always has a fat tail (because you can't have fewer than zero events.) Yes, it will eventually converge to normal (as any distribution will.) But even at expectation value E=14, while the curve looks close to normal, the tail is significantly fatter at 1.5*E. The estimate of 0.11% is too low by a significant amount.
                                I don't follow you. These are independent trials, so their exact distribution is binomial/Bernoulli. I'm not approximating anything with a normal. 0.11% is what is returned by
                                Code:
                                scipy.stats.binom_test(26,335,0.04, alternative='greater')
                                --
                                Dive fast, die young, leave a high-CHA corpse.

                                Comment

                                Working...
                                😀
                                😂
                                🥰
                                😘
                                🤢
                                😎
                                😞
                                😡
                                👍
                                👎