How I tried to save Divine Divinity
Posted on February 13, 2012
Back in 2001, Divine Divinity was in serious trouble. Several people at our publisher, CDV, wanted to kill the game because it was late, and our publisher’s producer needed help defending the game at an important internal publisher meeting. The goal of that meeting was a re-evaluation of their entire portfolio, and I was asked to write up a list of what I considered to be the strong and weak points of the first Divinity.
Being quite the idealist in those days, I made what I thought was a fair assessment of our own game, not realizing that it’d actually be used against us afterwards.
I figured it might be interesting to share the mail I sent to the producer with you.
This game really was a lot of work - really a lot
It’s quite long, but it reflects a lot of the hopes and aspirations we had at Larian in those days. Re-reading it, I recognize the idealism that drove us, as well as the hope that the publisher was going to forgive us for being late and give us the extra fuel we needed to finish the game the way we wanted to finish it.
For the record, it didn’t work out that well – several months later, just after the release, I had to downsize my team to 3 people from the original 30, because I refused to accept that I had no budget anymore, and spent everything on trying to finish the game as good as we could . My thoughts were that if the game was good, money would find us somehow. It didn’t of course. I caused quite some traumas with that attitude, and to this date I have regrets of how I handled the situation back then, but that’s another story.
Here’s the mail:
Hi,
As requested, I made an overview of what we think are the weak and strong points of Divine Divinity. It became quite a document, but I probably barely scraped the surface of our thoughts on the topic. I hope this helps you somewhat – more than willing to give you more information on certain points.
One thing which will make your presentation a lot easier – use the CTRL key for targeting npcs while hacking – you’ll find life is a lot easier.
Current weak points
Dialogs
While many bugs are present in these and they are being fixed, that’s not the main problem with them. A large degree of gameplay is to be had in a role playing game from the options you get in a dialog. In the quest to mix hack & slash with regular role playing, we probably oversimplified the available options in certain dialogs. This is not really visible in the demo area, where we think the mix is right, but later on in the game this is definitely an issue. The fixes for this are often rather simple, but complicated severely by the fact that the localizations are almost ready, and we need to maintain synchronisation between the three dialog base files (i.e. German, French and English). We have quite some tools in place to facilitate that, but a tool is just a tool, and given the enormous complexity/size of the dialogs in game, operating the tools/maintaining synchronization takes time. We are not to the level of Baldur’s Gate when it comes to the dialogs, though we could’ve been. One example that is present in the demo is the scene with the resurrected necromancer. The dialog there does not contain the richness that you would find in similar AAA games.
Savegame : Thelyron, click on the lever just in front of you, go to where the mummy appears and talk to him. There are no real options in the dialog, which is not such a good thing, but still acceptable in this particular situation. The fact that they stand still is just a bug.
Player expectation mismatches
Player expectation mismatches means that the designer of a certain quest/story scene designed something in a certain way, and that completing it in the way the designer intended it, feels “wrong” to the player. In general you want the player to “think” he’s smart while always making sure that whatever he tries, he’ll stumble into it. There are areas in the game where this is currently not the case, and this breaks the players gaming tempo, and causes unnecessary frustration. Another aspect of the same problem is that there are things in the game which don’t offer the gameplay value a player would expect from something. That’s a hard concept to describe, but an example should make that clear :
There is a small village on the first map which is quarantined because a mysterious plague broke out there. Upon coming closer to the area, the player sees a small scene where a certain Doctor Elrath chats with the guards. The guards ask him if he found the cause of the plague and he responds that he hasn’t yet. Control of the game jumps back to the player then. What is wrong with this scene is that the gates do not close, so the player can just walk in in the quarantined village. This is wrong, because if it is under quarantine, it should really be under quarantine. A player would subconsciously expect that he has to find a way to break in. At present this is not the case (though it is scheduled for fixing).
The next thing that is wrong in this particular quest is that when the player eventually finds a way in the quarantined village, there is no real reward for him. What he encounters is several people who are sick, and one meagre hint in the direction that it is strange that Doctor Elrath can just walk in without being sick. (The quest for which this is the setup has it that Doctor Elrath is actually poisoning the villagers). There is a well somewhere in that village through which the doctor distributes his poison, but nothing happens when the player gets near the well. Later on the player will discover that the well was the cause of them all getting sick, but what’s wrong here, is that figuring out that there’s something wrong about the well should be the “reward” for breaking into the village. Right now the player gains no real visible new knowledge by breaking into the quarantined village, though he would have expected something to be gained by the fact that he broke in. Not gaining that knowledge subconsciously decreases the immersion of the player.
I hope this makes some sense, as I realize it might sound a bit vague, but it definitely is a weak point. We became aware of this type of problem somewhat too late due to delayed QA feedback, and are fixing instances of this wherever we can. It is not the kind of thing a reviewer or player will spot as being a problem, but what he’ll say is “mmm…this was rather boring”, and that obviously is not a good thing. Actually, this type of problem is probably the thing we are currently spending the most time on, since the fixes are sometimes very simple, and can have a very big impact on the player’s gaming satisfaction.
No savegame available for the demo as this type of problem does not really appear in the demo.
Balancing
Currently not good, but getting better every day. This is by definition a task where plenty of QA is necessary, as the only way to get valid feedback is by checking over a large enough sample of gamers for comments like “Too hard, too easy”. The problem with Divinity is that it’s so large, and that you can do so many things, that there really are a lot of parameters affecting the balancing. Over the past few months we have made quite some radical changes to the engine to facilitate the balancing effort, and this has paid of tremendously, but the basic fact does remain that the more QA is available for balancing, the better the balancing is. We’re pretty confident that by march 15th we’ll have a good balance in the game, but I doubt it will be excellent (cfr. Diablo II)
Monster AI/Variation
This is a real pity. The game engine is capable of amazing feats of AI, but most of these have been deactivated for the time being, for the very simple reason that there were always more pressing issues. For the same reason, the thing which was always planned didn’t really happen – i.e. giving each monster a different behavior so that every encounter is more fun for the player. It’s not like there is no variation, but the engine is equipped with much more behavior features than are currently visible on screen. An example of this is the orcs. They are currently using the same AI routines as the skeletons. Given the capabilities of the engine, it is about 1 day of work to give them a drastically improved AI, and we do hope to be able to do so, but they are lower on the priority list because it’s not like the hack & slash isn’t fun. It just could be so much more fun and kick the @#|{ out of Diablo 2’s AI.
Savegame “Orcs”: Just hack at them, they come straight for you, don’t really show intelligent behavior. You are artificially boosted.
Roofs
While you don’t even think about it after 5 minutes of playing the game (at least we never had a negative test report from the external testers), they would be cool. But we did put them lower on the priority list (cfr. Other mail) because of the problems they could cause, and the fact that other issues are more pressing.
Item generation
This actually falls under balancing, but right now the variation of generated items is not large enough (compared to Diablo II).
Skills
They are both a strong and weak point. Because of prioritisation, some of the visuals aren’t as nice as we would’ve liked and planned. But comparing to the competition, we don’t really have to blush. The main advantage you have in Divinity because of the hack and slash content is that at some point you get so much action going on on screen, that it just looks impressive, even if some of the skill effects are “subtle”.
Polishing
Prioritization and cutting features were necessary because of the development delay incurred. Since we got into that road we will probably not be able to do what we really wanted to do – i.e. “USA-ize“ the game. In our observations it is usually the case that US games feature a much higher degree of polishing than their European counterparts. That means that in general European games ship with bugs which were categorized as minor, whereas AAA games from US houses don’t have these. Diablo 2 is the best example of this. The main reason for this is financial constraints, but that doesn’t mean it’s not a pity. If you look at most reviews of Diablo 2, the word “polished” is something that is going to keep on popping up. We are doing our utter best to get rid of all the imperfections, but it’s clear that we will not manage to get rid of all of them in time.
Not enough voice
Given budget and scope, it’s impossible to have voice on every single dialog in the game, but it would definitely add a lot if this had been possible.
Game strong points
User interface
So far we’ve had over 100 external testers of all kinds (meaning it’s a mix between experienced players and people who’ve never touched a RPG) come by to test the game. To this point, we’ve not had to explain a single time how to operate the game. Given the complexity of the things you can do in Divine Divinity, we are rather proud of that aspect. There are still bugs in the interface, and some improvements we are adding, but in general the accessibility of the game can be considered to be a very strong point of the game.
Action
If we get the balancing right, and fix the remaining bugs, then we think the hack & slash portion of the game is at least of the same level of Diablo 2, which is a very high standard to reach. While it might look simple on the surface, serving a decent portion of hack & slash is an incredible complex task, but we feel we’ve reached that design goal. As said, the only the thing that can stand in it’s way now is an issue of balancing.
Savegame “Thelyron 2”: Just right click like a madman to execute your special move, and drink lots of stamina potions (Note : You are artificially boosted to make things easier)
Interactivity and NPC reactions
I don’t think anybody will complain about the level of interactivity in Divinity. It’s usually the first thing people talk about after having played the game for several hours. The best example in the demo is the dwarf Otho, but there are plenty more.
Savegame “Otho”:
Take the herbs(Drudanai), Otho comes to complain and query you about what you are doing. Now follow him to his house. Click on the bones of his ancestor, Otho complains. Take one of the mugs of beer, drag it over the barrel. You just tapped a pint of beer. Open your statistics window, watch your intelligence and your strength, click on the pint of beer. Close the statistics plate, equip a weapon, walk outside, kill one of Otho’s pigs. Otho complains. Kill his second pig. While doing all of this, check the attitude meter for Otho.
Graphics
Opinions might differ, and while graphics are not “stunning”, we keep on getting reports from the testers that the game is so nice to look at. The same in the press – PC Zone UK for instance saying “It’s all very pretty”, or Power Unlimited (Dutch), “The gfx are gorgeous”. We strongly believe that when it comes to 2D RPG’s Divinity is one of the best looking games. An observation I’d like to make is that with probably the exception of the upcoming Dungeon Siege, I can’t think of a recent PC 3D RPG that was a hit. I’m rather convinced that if you were to release Diablo 2 or Baldur’s Gate 2 today, they would sell the same numbers. While the comment “It’s not 3D” is a very easy one to make, judging from the gamers feedback, they don’t care whatsoever. I believe the release of the demo will set a lot straight there.
Sound
If we are getting positive feedback on the gfx, we are getting amazing feedback on the sound. While there are still bugs in the sound engine that we are fixing, having two sound engineers and one musician on the project since day one of development is paying off. Not everything is working 100% yet on the sound system, but when it will, the sound image will be very complete.
Skills and character development
One of the things testers like the most is the fact that they get total freedom in the way they develop their character. If this will be balanced 100% correctly, this will in the end probably be the thing that players like the most about Divinity.
Mix between hack and slash and role playing
This was from day one a big gamble, but seeing how everything is coming together, we believe this is working out very well. The demo on its own actually tells that entire tale. In the village, you have traditional exploration and role playing elements. Going into the catacombs, you get more than your share of hack and slash.
Humor
This is something a lot of testers mark as something they like a lot.
Savegame “Confused”: Just walk south.
Exploration and size
If there is one thing you can do in Divinity, then it is exploring. There are plenty of things for you to discover, and you usually get well rewarded for your exploration efforts. And while it has been the biggest development problem, the size of the world is something we believe will be well appreciated by gamers. If the last remaining problems are solved, then people who buy the game will be playing it for a long time. That should help with word of mouth.
Weak point/Good point summary
The game’s weak points are mainly caused by time/budget constraints, too aggressive cutting only to find out that some things shouldn’t have been cut in the first play and starting with QA too late. Dialogs, player expectation mismatches and balancing we consider the most important weak points, and are what we are spending the most time on. Frankly put – these must be fixed. If they are not fixed, the game risks missing its potential. The other weak points would be nice to see totally fixed, but given the situation and the close master date, that’s not feasible.
On the bright side, we’re now more than ever convinced that Divinity is very close to the point of being an AAA RPG which will be liked by many players. There might be a positive bias among the testers we are inviting, but to this point we have managed to convince the most sceptic of them just by letting them play the game and not saying a word. Already word of mouth is flooding our mailboxes with requests for being allowed to come and test, and if we can get Divinity the attention it deserves in the press, we think this game could do really good. We think that we can get the most important parts of the mentioned weak points fixed in time, but admittedly it’s going to be a close call.
Current development status
First some numbers. There are currently 553 bugs flagged as open on our internal bugzilla server (coming from 1392 at start of internal QA). There are 132 bugs flagged open on CDV’s bugzilla – out of 875 reported bugs. That gives a total of 685 bugs out of 2267 remaining open, meaning that in a period of two months and a half 1582 bugs were solved, or about 633 bugs per month. That’s not bad, but it could’ve been better with more detailed QA, since a lot of time was spent by the developers trying to reproduce several of these bugs. Our internal QA however is overloaded with work (4 full time testers, 5 externals a day), so they can’t perform any better than they already are.
As related in the document I gave you and Martin, the typical ratio of bugs for a RPG of this calibre is between 7000-10000 bugs (numbers taken from Diablo 2/Baldur’s Gate). Since QA for Divinity started up rather problematic and late, it is feasible to say that the amount of total bugs in the end will be lower than the average over BG/Diablo (by the time QA started, a lot were already fixed by developers doing their own testing) – and when those 685 are solved, the game will probably be ready for release, but not as polished as we would want it to be. That’s not saying it will not be polished enough, it’s just saying that we would have preferred a higher degree of polishing.
Among the bugs that really count (except for blockers,crashes etc…), we are treating the gameplay ones as the most important. That means we give priority to something like “I don’t understand this quest” or “this area is boring”. This latter type of bug is usually related to “player expectation mismatches” and “Dialog” issues which are mentioned in the beginning of this document. Those two together with the balancing, are what currently stand between Divinity and its release.
All time estimates we currently have are based on the bug-reports – and at this stage given the amount of reported bugs, it looks like we will be ready by March 15th, but the problem of course is that we don’t know how much more bugs will be reported. If it wasn’t already, QA now really has become the dominant factor determining the release date.
For reference, I’ve included the bug reports from our internal bugzilla server as a text file.
It's a Divinity sequel, so not a chance.
Can't really imagine an isometric Action "RPG". Don't destroy my hopes and dreams.
Brain hiccup? Diablo 3's pseudo-isometric with a fixed camera and there are lots of others.
