Jump to content

_Xq_

Junior Defender
  • Content Count

    7
  • Joined

  • Last visited

    Never

Community Reputation

0 Neutral

About _Xq_

  1. _Xq_

    Item Check Thread

    May I have this checked please. https://imgur.com/a/FlTY1Dd Thanks
  2. In response to CrzyRndm's message on previous page. It seems I misunderstood the earlier question about the multiple types of characters and gave a reply according to how I understood it at the time. Regarding the approach you described I see a couple of possible issues too, at least when I consider how I use my characters. While my roster does handle several different roles, there is normally only one role for each character. As an example: I'll define rule sets for a builder and a fighter in comparer settings. I start a map, build it, and swap to my fighter for the combat. Now at some point I get a loot drop that is highlighted. I'll go and check the item and find it to be a nice piece of tower gear, obviously highlighted because of the good score from builder rule set. Well, my fighter has no use for it due to having a different role so I'll store the item for later comparison. The combat ends and I'll swap to my builder and see that while the item's general strength was stronger than my fighter's gear it is inferior to my builder's gear. So the item ends up not being useful for either of my characters in this example. The problem as I see is that by keeping it just with the active character it may be hard to get a reliable comparison from an item unless the item fits with the purpose of the said character. And that is one reason why I suggested comparison for multiple characters simultaneously. Regarding the color coding, the item tiers for mythical and above are already color coded. Would adding two (or more?) colors into the mix might make it harder to interpreter? Of course it may be just that I haven't correctly understood your approach. Regarding the two issues you brought up for my approach, local / remote data sync shouldn't be too much of a problem in my view. I tried to describe a possible solution to this with the chapter where I was describing the save / load routine additions and the method there was supposed to be one that would not require any changes to the actual remote save files. Though, a player would still experience a loss of comparer settings if he/she would change computer without copying/transferring the setting file. Regarding the issue "yes it's green, now explain why", maybe a name tag with the name of the receiving character could be places somewhere? Or then just rely on the player knowing approximately what kind of gear each of his/hers characters are looking for. One thing that has also come to my mind is that regardless of approach it would probably be wise to keep the (improved) non-biased comparer as a default one and the biased one as something the player needs to manually activate before coming into effect (saved setting). The reason being that a new player may not be sufficiently aware on how the game works to create proper rule sets from the start. What comes to [[9935,hashtags]] method, I fail to see how it would fall too badly as long as we aren't comparing weapons that use different mechanics (such as rifle vs sword). As long as the method answers to the following questions: #1 If we had a blank version of the new item, how many upgrades would it take to turn it into the actual new item for the select stats. Upgrading uses the upgrade rules of the new item, the result may also be negative. #2 If we had a blank version of the new item, how many upgrades would it take to turn it stat wise into the same item as the receiving character currently has equipped for the select stats. Upgrading uses the upgrade rules of the new item, the result may also be negative. #3 If we take #2 - #1 do we get a value that is less than the amount of free upgrades on the actual new item. If yes then it should take less than the available upgrades to turn the new item equally powerful as the equipped one. The above questions are enough if the equipped weapon is fully upgraded and similar enough. If not then we would need to make an educated guess where the remaining upgrades would go (damage in case of a weapon?) and "upgrade" the equipped weapon according to its upgrade rules to get the "actual new item" stats for #1. What comes to the example about the strong weapon vs weapon with less potential maximum damage. Yeah, comparing a pawn shot to say, blaster rifle could become challenging so maybe if the player has selected dps stats in comparer settings (damage, fire rate, extra projectiles) the comparison algorithm should combine those values into a dps value and see how many upgrades it takes to push the dps on the evaluated item to its current value (and same for the equipped one) and use that along with other upgrade count sources. When it comes to high damage overshadowing hero stats: if say, hero damage is selected as stats of interest and the new item spawned with -200 in that stat then it would take -200 upgrades to get from 0 to -200 so that stat alone would cause a large penalty to the item's score on its own I see bigger problem being a much greater value in selected hero stat overshadowing damage (like -100 equipped, +400 new). Then its up to the player to set the stats of interest so that specific hero stats are left out from comparison if high overall value of said stats on said character make changes caused by single item in that stat mostly irrelevant (due to logarithmic investment-return scaling). Other approach would be of course to add hero damage stat and its logarithmic scaling into the mentioned dps calculation. - Xq
  3. Thanks for the reply. Regarding the flaw with multiple heroes... One way to approach this could be to make the comparer settings separate for each character on the account. Then during comparison process the comparer would need to compare a new item's score against list of score values created from equipped items of same type across all characters on the account. If the new item's score exceeds any score on the list then the item is a potential upgrade to at least one of the account's characters. This however can create another inconvenience as sometimes the player may want for example to keep a separate set of characters with lower stats for lower level (or more challenging) play. Depending on the strength of the gear on those characters the comparer may end up showing half the drops as green on higher level maps even though the player wouldn't be looking for upgrades for the secondary characters. To solve this maybe something like a check box type list on which characters to include for the comparison would in my mind be pretty much required. The next issue would of course be managing the settings information as to avoid changing save files the information would need to be stored elsewhere but would still require additions to the save(load) process. A possible solution to this could be to create a function/method that is called after regular character save(load) process that would load(save) the comparer data from(to) for example local files (default values in case of missing / corrupt data) and then (in case of load) add it to the new comparer data fields under character data. Regarding complexity of non-armor items... I used a piece or armor in my initial message to keep the message more compact and simple as the purpose of the message was to transfer the idea I had on the matter instead of loosing too much focus into specifics. One way to simply non-armor (or accessory) comparison could be to use the rule set of the new item to break down all the stats of interest of the equipped item into how many "upgrades" it would take to turn an item with all zero stats into the equipped item (at least when selected stats are concerned) and add unused upgrades to the value to get a score. Then using again the rule set of the new item, break down the respective stats on the new item to get how many "upgrades" the item already has in its select stats. Add unused upgrades of the new item to this value to get a score to compare against the score value of the equipped item. Example Take an imaginary system where weapons normally get base 100 damage / upgrade, 1 attack rate / upgrade, and 1 reload speed / upgrade. If a weapon is of ultimate quality, the damage/upgrade value receives the 1.4 multiplier. The character used in comparison has a has damage, fire rate, and reload speed selected as stats of interest in comparer settings. New ultimate weapon: damage = 1000, fire rate = 3, reload speed = -33, 1/123 upgrades used breakdown of existing stats: 1000/(100 x 1.4) + 3/1 + -33/1 = -22.857 (about), now add (123-1) unused upgrades >>> score = 99.143 (about) Currently equipped non-ultimate weapon: damage = 10000, fire rate = 6, reload speed = 4 50/52 upgrades used note: breakdown uses the rule set of the new item breakdown of existing stats: 10000/(100 x 1.4) + 6/1 + 4/1 = 81.429 (about), now add (52-50) unused upgrades >>> score = 83.429 (about) Result: Green thumb. And indeed, even if we fix the bad reload to the same value as in equipped weapon and up the fire rate, the remaining upgrades will push the new weapon's damage to 12480 while the current weapon's maximum damage with the same side stats is 10200. One flaw in the method above is that depending on the player's choices the unused upgrades may or may not have greater value under the rule set of the new weapon due to the new weapon's 1.4 multiplier on damage / upgrade value versus 1 of the equipped one. To this I haven't found a definite solution but one approximation would be to assume that the player would dump the unused upgrades only to damage. To include this the comparator would need to simulate the upgrade process to get a fully upgraded version of the equipped weapon and break down the equipped score from that. Though, this assumption could potentially cause a large skew in the comparison if large number of unused upgrades exist on the equipped weapon. We can't know for sure, only assume, how the player wants to spend the points after all. The comparison gets even more difficult if the character (say, a jester) has a blaster rifle equipped and the item being looked at is a bloodshot staff or a gladius. To take weapon type mismatches into equation can fast turn the method overly complex. Yet making too many assumptions results in poor comparison results in these kind of mismatched cases. Anyway, sorry for the wall of text. I don't really know how to explain this in compact form. - Xq
  4. The opening also brought up matter with the item comparer. Reading this thread I haven't seen too many replies addressing that so here is my thought on the matter. Currently the comparer tries to make decisions fully automatically without any input or guideline from the individual player and as mentioned it often leads into results that don't suit the player's needs. In my opinion a better approach to this would be to allow the player to give s general ruleset based on which the comparer makes its decisions. For example, an "item comparer settings" button could be added to character gear overview screen on the right pane and clicking it would open a comparer settings screen on the left pane. the settings pane could look similar to the character stat screen with all the possible stat icons. There the player could highlight the stats that that he/she deems interesting. Then during item comparison the comparer would simply take the stat values of the highlighted stats on the item plus any unused upgrades then factor in tier bonus and compare it to a value received from the player's currently equipped gear piece of same type. Example Player has selected tower attack and tower hp as stats of interest. Currently equipped: mythical gloves with 122 tower hp, 170 tower attack, and 32/60 upgrades used >>> score = (122 + 170 + (60-32)) * 1.3 = 416 Item being under comparison: mythical gloves with 100 tower hp, 120 tower attack, and 1/117 upgrades used >>> score = (100 + 120 + (117-1)) * 1.3 = 436.8 Result: green thump as even if its slight, there is a potential gain in the stats selected in comparer settings. Of course other possibly useful settings in the comparer settings pane cold be check boxes like "must be of same material", "must be able to equip", "must be [weapon, pet, gloves, boots, ...]" and such. This kind of simple comparison wouldn't by no means be perfect and it is ultimately up to the player to decide if the new item is really an upgrade or not. It would however reduce the amount of running around checking every item as the player would know that there is at least possibility in stat gain for the stats he/she has selected when a green dot lights up in the mini map. Regarding on where to store the new setting data... It was mentioned that changes in save files is pretty big no. In my opinion new setting data the comparer would require could be saved even locally to a new? separate file. Any modifications to it would have no impact on item generation but only change which items are shown as green. Thus I can't see any risk of exploiting the system. Only drawback with local files would be possible loss of those settings if one would sometimes use another computer to play. This could be possibly mitigated by including the said settings as part of the data steam cloud sync handles. - Xq
  5. _Xq_

    Item Check Thread

    And now its destroyed. Thanks for the check.
  6. _Xq_

    Item Check Thread

    I just realized the first screenshot wasn't taken from proper screen. The link below contains full item info. http://imgur.com/r1yqsyk Sorry for the inconvenience.
  7. _Xq_

    Item Check Thread

    Hi Can I have item check for my drill. http://imgur.com/vZOTdmv The drill was gifted to me and since another player questioned it, I can't be sure anymore if its clean or not. Thanks in advance
×
×
  • Create New...