MantisBT - VCMI
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0000335VCMIMechanics - Adventure Mappublic2010-01-01 19:442014-05-30 17:39
ReporterZamolxis 
Assigned ToTow 
PrioritylowSeverityfeatureReproducibilityalways
StatusclosedResolutionsuspended 
PlatformOSOS Version
Product Version0.75 
Target VersionFixed in Version 
Summary0000335: VCMI pathfinder seems to go for the most cost effective path at end of day, while H3 went for the most intuitive choice.
DescriptionVCMI pathfinder seems to go for the most cost effective path at end of day, while H3 went for the most intuitive choice (from the perspective of our eye perception). This report is mainly to point out a difference between VCMI and H3. If the devs are aware of this and coded it as such intentionally (as a feature), we can close this as "no change required".
Steps To Reproduce
Additional InformationI discovered this while investigating 0000327 on an edited map (see attachment). The current logic seems to follow the principle identified by TowDragon during 0.71 (see forum post: http://forum.vcmi.eu/viewtopic.php?p=2403#2403 [^]), however making an exception for the last move at the end of the day, for the least cost in the beginning of the next day. Perhaps a bit of feedback from other devs would be useful: does the path choice feel somehow "weird", as compared to H3 where path arrows would perfectly overlap the road, or not? Would players notice/appreciate the minimal gain in MP, or more perceive this as a bug?
TagsNo tags attached.
Relationships
related to 0000327assigned SXX Minor difference(s) in path calculation (last move rule + penalty from slow creatures) 
Attached Files?  VCMI_Tests2.h3m (11,995) 2010-01-01 19:44
https://bugs.vcmi.eu/file_download.php?file_id=210&type=bug
jpg 2010-01-01_H3Path.jpg (26,752) 2010-01-01 19:45
https://bugs.vcmi.eu/file_download.php?file_id=211&type=bug
jpg

jpg 2010-01-01_VCMiPath.jpg (26,793) 2010-01-01 19:45
https://bugs.vcmi.eu/file_download.php?file_id=212&type=bug
jpg
Issue History
Date ModifiedUsernameFieldChange
2010-01-01 19:44ZamolxisNew Issue
2010-01-01 19:44ZamolxisFile Added: VCMI_Tests2.h3m
2010-01-01 19:44ZamolxisIssue generated from: 0000327
2010-01-01 19:44ZamolxisRelationship addedrelated to 0000327
2010-01-01 19:45ZamolxisFile Added: 2010-01-01_H3Path.jpg
2010-01-01 19:45ZamolxisFile Added: 2010-01-01_VCMiPath.jpg
2010-01-01 20:05ZamolxisPrioritynormal => low
2010-01-01 20:05ZamolxisSeverityminor => feature
2010-01-01 20:05ZamolxisSummaryPathfinder does not always look for a path going down as much as possible (with respect to the condition of the least cost) => VCMI pathfinder seems to go for the most cost effective path at end of day, while H3 went for the most intuitive choice.
2010-01-01 20:05ZamolxisDescription Updatedbug_revision_view_page.php?rev_id=521#r521
2010-01-01 20:05ZamolxisAdditional Information Updatedbug_revision_view_page.php?rev_id=523#r523
2010-01-02 07:41WarmongerNote Added: 0000559
2010-01-02 11:28ZamolxisNote Added: 0000560
2010-01-02 11:29ZamolxisNote Edited: 0000560bug_revision_view_page.php?bugnote_id=560#r534
2010-02-10 03:44TowNote Added: 0000615
2010-02-10 03:44TowStatusnew => resolved
2010-02-10 03:44TowResolutionopen => no change required
2010-02-10 03:44TowAssigned To => Tow
2010-03-02 21:30ZamolxisNote Added: 0000745
2010-07-13 10:51ZamolxisNote Edited: 0000745bug_revision_view_page.php?bugnote_id=745#r955
2010-07-21 15:42ZamolxisNote Edited: 0000745bug_revision_view_page.php?bugnote_id=745#r979
2010-07-24 14:41ZamolxisResolutionno change required => suspended
2010-07-24 14:53ZamolxisNote Edited: 0000745bug_revision_view_page.php?bugnote_id=745#r988
2014-05-30 17:39beegeeStatusresolved => closed

Notes
(0000559)
Warmonger   
2010-01-02 07:41   
I guess in original game road bonus works only if we move according to its direction.
(0000560)
Zamolxis   
2010-01-02 11:28   
(edited on: 2010-01-02 11:29)
It's not really that. If Ivor would have enough MP to move one extra tile on the road, the path would overlap perfectly in VCMI as well. You can test that by editing the map attached, putting Ivor 1 tile lower. It really looks like VCMI is smart enough to handle the last remaining ~70 MP of the hero, depending on his next move the following turn (which would cost him only 50 MP, i/o 70 for the diagonal step in H3).

While this is a smart behavior, I am just a bit concerned the players will fall into the same trap I fell when first noticing this, thinking it's a bug.
(0000615)
Tow   
2010-02-10 03:44   
Pathfinder has been designed to find the most optimal path, using rules of movement from H3.
In this case I'd treat current VCMI behaviour be that as bugfix for problem present in H3 (offering suboptimal path).

So - it's not a bug, it's a feature! :-)
(however we should consider documenting it somewhere)
(0000745)
Zamolxis   
2010-03-02 21:30   
(edited on: 2010-07-24 14:53)
FYI - If you don't mind, I'm going to leave this one in Resolved, without moving it to Closed, until we document it properly somewhere.

EDIT: Another reason why I prefer to not close it yet, is because a few times the path chosen seemed quite unrealistic for the eye, so it may bother the players. I want to keep monitoring this for a while.

I'm also moving this from "no change required" to "suspended". If the path chosen seems unrealistic for the player, perhaps we should still consider one of the following:
- go back to the H3 system (sacrificing an almost insignificant bonus, for the sake of a more realistic looking choice)
- or else maybe find a way to "store" the sacrificed movement points lost, and give them as bonus in the beginning of the next day (it'll be like hero "going to bed earlier" in order to "wake up earlier" the next day)

//Apologies if I get annoying insisting on this, but if I have the feeling the hero makes strange path choices - even though I know it's for the sake of most efficient use of movement points, I can imagine other players will get the feeling VCMI implemented pathfinding rules "worse" than H3 (and we can't expect them all to read all that we documented to understand why it is like this). But of course, this is by no means a priority. It can wait for finer tuning till after the 1.0 release.//