Angband 4.0beta status

[Nick]

Update cb2b719 fixes:

• OS X graphics (thanks myshkin)
• Make lib/tiles/list.txt look prettier
• On Powerwyrm's advice, do an explicit check that a monster is never placed on top of the player

No bugs for a while - apart from the OS X sidebar update problem.

[Derakon]

Originally Posted by shreesh

Why are fail percentages for spell etc. never more than 50%? Even at other percentages, the spells fail probability is lot more than what is reported. Can't we have a better estimate (I don't have any idea)?

The spell failure rates are accurate. You just notice the "unlikely" failures far more than you notice the "unlikely" successes, because humans are bad at coping with randomness.

Fun fact: one way to recognize when someone is trying to generate "random" data by hand is that it doesn't have long streams of the same result. In sufficiently large datasets you should expect to see a few such streams.

I'd like to see Angband collect more stats on the player. A breakdown of how often each spell is cast and how its nominal failure rate compared to its actual failure rate would be pretty neat to see.

Also, birth option of preserving artifacts still does not work well. I don't see a few basic artifacts after long hours of play. In debug mode, for example, if you generate hundred of very good objects, you can see phial generated on the ground nearby and bam after some repeations it is no longer generated. My guess is when it is generated in the inventory of some monster (possibly unique) it is no longer generated again. No proof, just guessing.

Remember that once you learn the name of an item, it cannot be generated again. If your workflow for this test included picking up the phial, then you would expect to never see it again. What exactly were you doing?

[Ingwe Ingweron]

Angband 4.0 dev cb2b719

This is an error in OSX. On initial character creation, there was an error - "failure to save subwindow preferences" (or something like that). What was apparently happening was that the Documents/Angband directory was not being created before the attempt to save .prf and savefiles. After the failure, there was a Documents/Angband directory, but it was empty - no subdirectories, no files. Reattempting character creation worked fine, I believe because now the Documents/Angband directory existed.

[Carnivean]

Originally Posted by Derakon

I'd like to see Angband collect more stats on the player. A breakdown of how often each spell is cast and how its nominal failure rate compared to its actual failure rate would be pretty neat to see.

I'd like to second this request.

[myshkin]

Originally Posted by Ingwe Ingweron

Angband 4.0 dev b66dd1d

A possible insight into the failure of AC to update. In the attached savefile, @ has a Robe and has Full Plate Armor. With the Robe, @ AC is 22. With the Plate Armor, @ AC is 84, but the weight of the armor encumbers @'s movement. Switch between the armors and you will see the information at the left of the screen fails to update the armor class. Even after moving and engaging in combat with a townie, the AC display does not update without a redraw. Attachment 1276

Could be totally off base, but I'm wondering if the failure is related to the receipt of the encumbrance / move more freely messages, similar to the run-past bug when a message was received? Oh, and this is in OSX.

I also see this with your savefile. Oddly, the basic character subterm updates just fine. Also, for me, becoming encumbered doesn't trigger an update, but becoming unencumbered does.
In both cases, update_sidebar() and update_player_compact_subwindow() both get called at least once, and prt_ac() eventually prints something with the correct new value. I'm not sure yet why this would fail just on OSX, or why it fails at all. I should note that wearing the full plate triggers enough message text to have a -More- after the wearing message, while wearing the robe allows everything to fit on one line. I'm guessing it has something to do with that.

[shreesh]

Originally Posted by Derakon

The spell failure rates are accurate. You just notice the "unlikely" failures far more than you notice the "unlikely" successes, because humans are bad at coping with randomness.

I have read about this. I will try to get some statistics today.

Originally Posted by Derakon

Remember that once you learn the name of an item, it cannot be generated again. If your workflow for this test included picking up the phial, then you would expect to never see it again. What exactly were you doing?

Initially I was beta testing different classes. In some of the games I notices I did not see star and arkenstone at all. So I tried the following in debug mode with both randarts and standarts. I use roguelike controls on linux console.

Start a new game.
Know about objects. (99l)
Jump to some random level. (j10 - j999). Max you go down is level 127.
Create 100's of very good objects. (V99 - create 40 or so objects)
Wizard light the level. (w)
See the objects on the floor in whole level without picking anything. (])
Repeat by jumping to another level.

For testing purpose I do sometimes pick daggers, and sometimes phial but nothing else.

I will try to collect some statistics today for this to see after how many iterations I do not see any star or arkenstone. I think it is not many if I jump to a deep level.

Also it seems level feelings relate only to objects on the floor of the dungeon. If an artifact is generated in the inventory of a monster it does not contribute to the level feeling and yet the object may be lost if you leave the level without killing that monster.

[AnonymousHero]

Originally Posted by shreesh

Jump to some random level. (j10 - j999). Max you go down is level 127.
Create 100's of very good objects. (V99 - create 40 or so objects)
Wizard light the level. (w)

Note that some artifacts (including the Phial have a max depth of 100.) I'm not sure the game actually respects that setting or not, but it could be an explanation for the "missing" Phial on very deep levels.

[Ingwe Ingweron]

Nick, have you decided that Bug #1881 is not a bug or that it just can't be fixed?

#1881 Object sub-window delayed 1 turn. E.g., pick up a potion or scroll and it still appears in the sub-window until the next move by either player or monster. Unlike the monster sub-window, which works correctly, kill a monster and it immediately is removed from the sub-window.

[Ingwe Ingweron]

Attack wands and rods now have the magic device boost percentage explicitly stated in their description. Thank you, Nick.

But, what about activatable items such as weapons, armor, gloves, and dragon scale mail that have an attack activation? Are these not "magical devices" within the definition of receiving a boost, or is it just that the description is not reporting their boost as was the case for wands and rods before?

[Nick]

Originally Posted by Ingwe Ingweron

Nick, have you decided that Bug #1881 is not a bug or that it just can't be fixed?

#1881 Object sub-window delayed 1 turn. E.g., pick up a potion or scroll and it still appears in the sub-window until the next move by either player or monster. Unlike the monster sub-window, which works correctly, kill a monster and it immediately is removed from the sub-window.

I seem to have thought it was fixed. I've re-opened it now, but I'm not going to delay release for it.