Random in Boss fights?

A good find about how good or bad it is for encounters in PvE to be heavily or lightly scripted, which is interesting given how much the degree of “random” affects fun.

Tobold wrote:

In my opinion there is too little randomization in MMORPGs. They are nearly totally deterministic. Before combat even starts you know what the monster will do, and what keys to press in which order to optimally defeat it. Thus combat involves no thinking, only execution.

As Craig Stern says, the solution is not making the result of button presses unpredictable, but to make the opponent unpredictable, or the starting situation. That is why card games work: The cards you draw are random, but what you can do with them is not. And in a MMORPG the monsters could be made more unpredictable as well. Why do people need to know in advance what the boss mob is going to do after X minutes to beat him?

posted by Tobold Stoutfoot at 8:54 AM on Nov 18, 2012

To be fair, its better to read the entire post and linked articles.

Would introducing more randomness be good?

I think it would be terrible – consider that if you really want to make the bosses challenging then you make them more intelligent…..say they don’t have a threat table anymore, they just crush anyone who tries casts a heal.

Same issue in pen and paper – for some reason (story) the big bad orc (high CR monster) wails on the fighter, not the cleric.

I can see the point about execution and predetermined strategies removing “creative” aspects, but really the 4th wall is so present in an MMO that it is just differing personal perspectives for how far the slider between real and scripted the entire game is. I also think that some random is ok, but too much will be very disruptive to the players, as their outcome for victory is too much outside their control.

The card game in Tobold’s example is not a good match to the game style events in an MMO, given how different the perspective is, but does help my point: who likes playing solitaire when the game cannot be won?

Trial of the Champion had the pvp based fight which was excellent, but it still came down to a set of abilities, and a priority based kill order. It was also hated by a few raiders.

Another random fight was Lord Ryolyth in Firelands – who was basically impossible if the randomness didn’t go your way. That is a shitty way to spend an evening.

Go have a read as the discussion on both Tobolds and Craig’s blog are good stuff. Happy Killing.

Just a few random screenshots from leveling…

Just a few shots taken while I wander aimlessly through Pandaria.

LFD Quirks

I’ve noticed that the first group I get in LFD is typically a fresh run where everyone starts from the start. If for some reason somebody does not accept, then the next group will almost always be a mid-way through run of some form. Sometimes its early, other times is at a hard boss.

I have no idea if this is by design, and I think its good. But is it just me – does anyone know if this happens for Dps, or just Tanks, or if its just a strange consistency in recent randoms?

If its planned you’re typically going to wait for a group, it is sometimes better to get into a part way group so that the wait time (for the guys that wait) is overall the same from queue to completion.

A counter though is that joining a group that has issues might have you joining a group of terrible players, rather than replacing a terrible player. I guess there would be no pattern to that.

the loot system is farked

In another post’s comment, I wrote that “the loot system is farked in wow“. It really got me thinking that such a negative statement could gather up how I feel about PvE raid rewards in World of Warcraft. It is worrying. Not only for PvE raiders, but for the fact that other games will look to WoW and copy the reward system as a fundamental. In this case Blizzard has it fundamentally wrong.

I promise that this will be one of the last posts about loot, gearing, and the cycles involved for a while. Just one last rant to get off my chest.

a skinner box

A skinner box session in progress. I think the bird is DPS...?

The system is based upon grinding for gear, in a most inefficient manner. When you kill a boss a random sub-set of items is presented as awards and the team get to decide who gets what. I like the fact that we get to decide how we divvy up the booty, but there is no consideration to who participated, or what is needed in the team. And that is just broken.

We have entire systems (ilevels, gear points, reforging, enhancements and enchantments, etc) to classify gear, and improve the gear we have, but no way to reward repeat raiding effort to gain specific items. These items are often needed to move forward efficiently.

You might think that I am forgetting that you get widget-points for killing a boss (who cares what they are called – widget points is fine), which can be use to purchase items? I’m not. That is a good system, which is logical, and places the control back with the player to participate in the gearing choices. Widget points are well loved as a system, the real contention is how fast you acquire them, not if the system is useful.

I’m raising the idea (yet again) that we need the game to facilitate improvement through effort in a better manner for boss loot.

Examples of the issues:

  • There is no point dropping Priest specific gear if we don’t have on in the raid. It is useless.
  • There is no point dropping an item that cannot be used by anyone in the raid at all, like Spell plate with no Paladins.
  • No prohibition on rolling for items which are irrelevant for a spec as a “Need” item.

Oh, but am I forgetting that it will get converted to a shard, and shards have value? No. An item being “sharded” in early content is a total waste, as the team really need upgrades to progress, and it late content you have so many that they are almost irrelevant. Shards only have high value and high use in the middle of the game, when you have gear worth using the shard on, but also don’t have 90% of your raid geared.

It is all really illogical.

But then is it really illogical to everyone? The drop randomness keeps a percentage of the players raiding each week. They can do nothing except raid to clear, and hope an item drops. This means they keep at it, and keep playing longer. It is not a stretch to suspect that if the loot system was improved then we could see more players switching to alts, or switching to other games.

Perhaps it is only illogical to keep doing it.

Well then you might ask what can be done. Well if you go digging through this blog you’ll find all manner of suggestions. In the wider wow community you’ll see more, and if you look as far as outside the wow community (gasp), there are plenty of good examples of other ways to handle rewards and loot. Far too many to bother with a summary here, but a few driving principals apply:

  • Minimise the amount of wasted loot
  • Balance loot amongst the classes and specs present
  • Guide loot distribution using the assigned role
  • Minimise the impact of greedy players

However here are a few I prepared earlier: Better gear rewards from runs, and Algorithm for need vs greed by role.

The opportunity to change this in the current 4.2 game is not a reality. It is probably too large a change to be made except by change in major version like the next expansion, but it could be made.

May your dice roll high, and you always see your loot. TyphoonAndrew

Better gear rewards from runs and raids

a lot of gears

Not that type of gear! Damn it.

Another enhancement (adding to a post about Need vs Greed by class & role) which would not hurt the Warcraft game at all would be to constrain the types of drops to at least the classes present in the run/raid. This an old idea, but a feature which seems blindingly obvious as a value add to players. It is the kind of assumption that a new player has, which is adjusted when the gear drop discussion happens.

Would you rather play for a slim chance of reward to your team, or a consistent reward to the team? Is there a significant disadvantage to adding this?

The suggestions is to mod the loot system mechanics so that:

  • Don’t drop gear for a class that is not present. Especially the class tokens, or class specific gear.
  • Don’t drop the same item, unless there is more than one of the class.

Continue reading

algorithm for need vs greed rolls by role

I’ve bashed the LFD Call to Arms as not going far enough, and not being an effective solution; and as a blogger with no leverage or responsibility in the game that is easy to do. So instead of a blog post which grumbles, here is a blog post with a suggestion.

Item stat based rules for Need vs Greed, for group roles.

The random dungeon finder tool now filters on armour type, but no further. I’m proposing that there is an algorithm which can filter the Need vs Greed option even further, so that gear is far less likely to be awarded to the wrong role. Using this the players would be getting gear with a far higher (but not imperfect) degree of focus on the role they joined to execute.

We have four roles and a range of special cases for gear within the class combinations created by those roles (Range DPS, Melee DPS, Tank, and Heal). It’s a bad thing to see a Need roll on an item which will never be used, such as a Death Knight rolling need on a Plate item which has Int on it, or a Hunter rolling on a staff with Spirit.

Here is a suggested solution to improve the algorithm which filters Need vs Greed selection:

Use role based filters for all Need role which restrict based upon:

  • Apply a primary stat filter, where the item must have that stat present.
  • Apply a secondary stat filter, where any one of the listed stats must be present.
  • Apply a secondary stat filter, where if any of the listed stats are present then the character is excluded from the Need roll.

This allows for the armor type restriction to hold as currently implemented, and the armor type restriction would still act as a restriction. It allows trinkets, rings, necks and other non-armor type items to be classified using the stats assigned to the item. This mean that it is not an arbitrary assignment of priority, but one based on the stats which the item has and how they are valued as recognised by item weightings.

The rules allow for an item within a armour type that a role might current role on to be corrected, like Int plate vs Dodge plate, or a trinket with only one stat, and allows for ignoring the Stamina values on gear. While this might disadvantage Tanks slightly, I am not sure of many items which would otherwise be selectable which would not also be valid choices for DPS characters – so the degree of impact is less.

All that we need to do is decide on the primary stat for each class & role combination, set a valid list of secondary stats, and also set some stats to exclude. It is worth noting that the inclusion of a stat in the included or excluded list is not mean to dictate a priority amongst them, it just means that this is a potentially valid stat for the class-role.

For example: It would be a shame for a range Tank to miss a role for an upgrade with Dodge on it just because a DPS in the same group rolls need.

The syntax is = Class Role: Prime-Stat / Stats Included / Stats Excluded

DK mdps: str / mast, crit, hit, expertise, haste / !int, !spirit, !dodge, !parry, !block
Pally mdps: as above
War mdps: as above

Druid mdps: agi / mast, crit, hit, expertise, haste /  !int, !spirit, !dodge, !parry, !block
Rogue mdps: as above
Shaman mdps: as above

DK tank: str / dod, parr, mast, exp, hit / !block, !int, !spirit
Druid tank: agi / dod, mast, exp, hit / !block, !parry, !int, !spirit
Pally tank: str / dod, parr, block, mast, exp, hit / !int, !spirit
War tank: as above

Druid rdps: int / spell power, hit, crit, haste, mastery /
Hunter rdps: agi / hit, crit, haste, mastery / !int, !spi
Mage rdps: as per Druid
Priest rdps: int / spell power, hit, crit, haste, mastery, spirit /
Shaman rdps: as per Druid
Warlock rdps: as per Druid

Druid heal: spi/ int, mastery, crit, hate, spellpower / !hit
Pally heal: as above
Priest heal: as above
Shaman heal: as above

As a basic this does it. It could be expanded to understand the ratio of Spirit to Int in comparison of caster items, where one is valued to dps higher than healers.

It could be expanded to ignore the rules if the person who could roll Need already has an item in that location which exceeds the item weight by ~20%. Meaning a player could not roll Need on a ilevel: 333 item if they already have a ilevel 356 in that location. For them it becomes a forced Greed.  That might mean that another player is the only one who can greed, or that nobody can roll need. You could also open the rolling so that if nobody is normally allowed to roll Need due to this rule, then the system lets them role Need if it is in their armour type. This way a offset item goes to somebody who might actually use it, not to disenchant or vendor.

And I’m sure there is a huge range of exceptions that might apply for funny items. No way in hell I was going to try and parse all of the WoWHead data to validate this.

On plan I think this is a solid idea. What do you think?

Happy Gearing