Category Archives: Development Diary

Afterlife : Updates & Marketing

Posted on by
Teleporters Puzzle (screenshot by Logort)

Teleporters Puzzle (screenshot by Logort)

Here a post to do a recap of what happened to the game after its release, and some words about marketing & sales.

Updates

I just released a few days ago the version 1.0i of the game, which may (or may not) conclude a long series of updates done since the release of the game, mostly based on user feedback, requests & suggestions. And one of the users on Steam suggested I should promote all the things I added to the game since its initial release. Before to do it on Steam, I’m going to do it here with the main points of each update :

  • v1.0 – October 2015 : it was the initial release on Managames.com, without the Rogue class
  • v1.0a – November 6th 2015 : added the Rogue class with its 3 specializations, fixed a bunch of bugs and was the 1st release on Steam
  • v1.0b – January 2016 : added the Treasure Hunt, which is a chained series of riddles, more or less easy, designed mostly to test your memory and your observation skill, and nicely put some goal unity throughout the whole adventure
  • v1.0c – March 2016 : added a puzzle in the Level 3
  • v1.0d – July 2016 : added Soldier difficulty, which was best tuned for though dungeon crawler fans and regular tactical game players and filled the gap between the Adventurer & the Warrior difficulties, respectively for the average dungeon crawler fan and the tough tactical game player ; it also added a more complete custom difficulty that lets you define the damage and health for each type of enemy (normal, Elite, Boss)
  • v1.0e – September 2016 : added destructible barrels, food and drinks that restore health and mana when you’re outside of combat, as well as a new puzzle for the Level 2, and a few additional secrets in the first couple of levels ; this update, coupled with the puzzle added in v1.0c made a more consistent experience puzzle wise for the game ; on release, the 1st noticeable puzzle was on Level 4, which was way too far in the game for puzzle fans
  • v1.0f – December 2016 : added a tooltip when moving the cursor over the damage & healing numbers to show who did those and with what ability ; this is a close equivalent of the requested combat log, and it greatly helps you to figure out who did what and when
  • v1.0g – April 2017 : added the “Low Enemy Density” option which removes about half of the Common enemies, but not the Elites nor the Bosses ; you can enable at any time this option if you think there’s too much combat ; loot and XP will be automatically adjusted so the game plays the same. This one is a game changer for a lot of players who felt the game was too much combat based, by bringing the balance back to the exploration side, close of the average dungeon crawler game
  • v1.0h – June 2017 : added the Mini-map
  • v1.0i – July 2017 : added a Macro system to lower the repetitiveness of the combat, as well as a fully detailed explanation of each character when selecting the characters to create our party at beginning of the game, and a smoother rendering if you play with a refresh rate higher than 60hz

And of course, many other little & bigger features, changes, tuning and bug fixes. You can see a more detailed version of the update history here : http://steamcommunity.com/app/409450/allnews/ and the complete changelog here : http://steamcommunity.com/app/409450/discussions/0/490123727974362306/ .

Goals & Marketing & Sales

So let’s start with the good news : since last March, the game is making profit ! :-)

Ok, this is a totally arbitrary concept, but it means after all expenses, I earned a bit more than the average salary in France, which was the 5th goal in the following list I had written a couple of months before release :

1- make a game that isn’t a tennis game, because I was fed up with tennis since a few years => done !
2- make a game that I enjoy working on and that I’m pride of => done..! :P
3- make a game that at least 1 other person than myself loves to play => done as well !
4- sell 1’000 units : almost insignificant in term of money, but shows the game has some potential
5- sell 10’000 units : at a price point of US$20, 10% discount average, with 30% cut for Steam, it’d pay back my time & money investment, but not the risk taken
6- sell 100’000 units : with a lot of discount sales, maybe it’s reachable, but that would be a big surprise

So the first 3 points had been reached before the release of the game. For the 3rd point, if you wonder who was that person, it’s simple : it was Robert Engstrand, the musician & main Beta tester of the game..! :)

4th point had been reached on Steam release day.

5th point took waaaay longer than that… :P

Especially as I had to discount the game a lot to make decent sales, plus I discovered I had to pay more taxes than expected ( 8-O ), it actually needed about 20’000 sales instead of 10’000 to reach it, but at least we got to it.

Now, for the 6th point, it might be doable if I bundle the game, but money-wise it wouldn’t bring much, and I still hope to get enough earnings to make a bit more ambitious sequel.

And, here the 2nd good news : in last April, Steam updated their Discovery algorithm, which shows the game to potential customers through different means on their platform ; since then, Dungeon Guardians sales have almost doubled..! It means the new algorithm shows the game to the right people, instead of nearly random people like before. And as it does that, it leads to more sales, which encourages it to show it to even more people. And all in all, everybody wins..! :)

At least for now ; maybe it’ll stop soon, maybe not ; only time will tell.

This leads to another important point. In last September, to calculate the game global scores, Steam stopped to count reviews from people who didn’t buy the game on Steam, to avoid abuse and cheating by some shady developers. Because of that, DG dropped from 83% positive reviews to 81%. In the previous 10 months it had already dropped from 92% to 83%. This is a natural process for many games, as when selling more & more, they start to sell outside their target to people who are actually not that interested in that kind of games, and thus are more likely to put a negative review. For DG, it has been likely a bit more bad than for other games, as explained in the previous post, but nothing extraordinary either.

Anyway, since last September, the game slightly crawled back to 83%, while still having decent sales, especially since last April (thanks to new Steam Discovery).

Now, let’s speculate about the reasons behind that.

I guess one of the 1st point is that all the new features, new options and polishing done since the release have paid off. People do enjoy the game a bit more in a general way.

The 2nd point is that at beginning of the year, I slightly redefined the marketing of the game. This was actually very simple : I added the tag “Strategy” on the Steam page, so now the visible tags on the page are “RPG, Dungeon Crawler, Strategy, Indie”, which perfectly describe my game and lead to less people surprised to find tactical combat in a dungeon crawler. The other tuning has been to change the introduction of the game, which was quite generic for a dungeon crawler, and instead opening with “In this tactical combat & dungeon crawler mash-up” and ending up with “with your tank, healer & damage dealers party“, so again, now people know better what to expect.

The 3rd and last point is that the game got 2 additional translations : Russian & Japanese. As it’s harder to enjoy the game without a decent understanding of the game rules, these new localizations might help, especially Japanese, as I saw a recrudescence of posted screenshots from people with Japanese nicknames… :)

 

Anyway, now I think I could try to write a regular postmortem, before I forget everything… I’ll try to get to it soon enough. Meanwhile, have a nice one ! :)

FacebookTwitterGoogle+RedditEmailShare

Technical Post Mortem : Building a game with Unity3D Engine

Posted on by
A tough Boss !

A tough Boss !

Ok, I’m still not ready to write a full post-mortem, but after seeing this thread “Unity is broken” – What are people referring to? , it made me want to write a list of all the technical issues I got creating Dungeon Guardians, most of them being tightly tied to using Unity3D (real name = Unity Engine ; Domain name = Unity3D.com ;) ).

Before to go further, I think I need to make this very clear : Unity is a great engine and without it Dungeon Guardians would either not exist at all, or look way less good than it does ; and having used it made me save a lot more time than it made me lose.

Thus said, I have built my game engine using the minimal stuff possible from Unity, and every part I used had a least one really annoying issue, if not several.

Let’s start with a painful task that almost no one can do without : debugging

Conditional Break in the MonoDevelop Debugger never worked. It makes debugging much tedious as I had to add some lines of code and recompile everything every time I need to check a special case. Later on I switched to Visual Studio and its debugger which was technically working, but often terribly slowly, so not much more usable actually. Though this last point is likely the wrong of C#, not Unity.

To stay with the debugger, pressing the “Play” button while the Debugger is attached makes Unity stall for 20-30 seconds instead of starting in less than 2 seconds. After that, it executes normally.

Moreover, Live recompiling (ie: changing the C# scripts with the game running) isn’t practical and thus I never got it working. This is the thing I miss the most from developing in C++.

Let’s continue with some isolated issues

Physics engine handles really badly small fast moving objects ; I had to use & tune this to fix the issue : http://wiki.unity3d.com/index.php?title=DontGoThroughThings .

The FPS counter in the stats doesn’t actually count the FPS ; it estimates them in the most idiotic possible way and thus is completely useless and misleading.

There’s no built-in decal shader, nor shadow-receiving transparent shader ; so for putting spider webs on my wall, I had to grab an user-made shader .

GL rendering stops to work in OnGUI after an alt-tab (ie: switching to another application), but it works again after a couple of alt-enter (ie: alternating between fullscreen & windowed mode). There’s no fix nor bypass for that. This is one of the rare bugs left in the game.

There’s no documentation on how to handle a lot of sound sources. The sound parameters would let you think it’s automatically handled, but it isn’t. Fortunately, I have been able to just cull the sound sources with my custom-made occlusion system for the rendering.

And now for the big part : animation & importation

Legacy Animation system was duplicating the animation clips assigned at realtime, instead of referencing them, which leads to a gargantuan amount of wasted memory (more than 1GB!). I had to swap to the new Mecanim animation system which was severely feature limited. Fortunately, the possibility to change the speed of the animation per layer came with Unity 5.1. Unfortunately, this version made the legacy immediate mode GUI way too bright in editor mode (but luckily not for the built game). 5.2 fixed that issue but destroyed the animation system, making everything jerking around when using multi-layers. 5.2.1 fixed that but not completely, and 5.2.2 didn’t change anything, thus final version of Dungeon Guardians is built with 5.1.3p3 and I kept on burning my eyes every time I launched the game in editor mode (which was all the time till recently).

Disabling & re-enabling the Mecanim component of an object removes the current playing animation clip instead of only pausing it ; ie: it completely empties and then resets the internal state of the animation. It brought me a lot of issues and it has been very hard to figure out a solution. But the hardest part has been to understand that behavior, because it is not documented. I have requested it more than 1 year ago to be added to the documentation, and as of today, it’s still not there .

Still in the animation department, it is not possible to select multiple animation clips and change their import options all at once. You have to do it for each clip, which can have near of 10 parameters, 1 by 1 ! There are hundred of animations in Dungeon Guardians, I lost hours if not days doing a stupid grunt work that should have taken less than a hour..!

To go on with the resource management, the texture re-sizing on importation is done with a very simple algorithm which has the double-perk of being both too aliased and losing contrast. The 1st one happens when resizing to a much smaller size and the 2nd one when dividing the size by only 2 or 4. Using “Bucibic Sharper” with Photoshop gives a much better result. It means I had to do an awful lot of grunt work to get the best of all the HD textures I had, instead of relying on the built-in scaler.

And now, the true Unity specialty : Half-cooked Features

It’s really the plague of Unity. They’re in a race to always add more & more features, while regularly never really finishing the existing ones and letting them with blatant missing possibilities.

The most patent example is the “new” animation system Mecamin (it was released with Unity 4 in July 2013), as explained above, which needed nearly 2 years to get a layer & clip based speed setting. It’s something that a lot of games need. Think for example at an adaptable run speed animation for the legs and a defined speed for a fire animation for the torso.

In the same vein, the Unity Input doesn’t allow to be set in-game. It’s even not possible to know what key is assigned to an input ! Which means any developer wanting to offer in-game configurable input (instead of the default Unity launcher setup) has to redo the Input from the ground up, as I did for DG !

In bonus, the text input field doesn’t work on Linux. It may or may have not be fixed in latest version of Unity, but as I couldn’t use it (see above), it meant I had to re-write my own (which doesn’t support Ctrl+V & stuff like that) : you’d be surprised at how much people actually want to enter the name of their choice for their character in an RPG, even if they use Linux..! ;)

The “new” particle system, Shuriken, released in February 2012 can’t use a skinned mesh (ie: an enemy) as source for its particles ! It has been fixed in Unity 5.3, released at end of December, nearly 4 years later. For DG, I had to use an external add-on (and yup, I had to paid for it).

Conclusion

I skipped a good bunch of issues which were also annoying but have been fixed before releasing the game, and that didn’t require extra work on my part before that. I also skipped a dozen of little issues which were more or less trivial to fix once found.

I also had several undocumented changes of behaviors, some really strange.

All in all, I feel lucky I had a version that let me built the full game without any major bugs nor important missing features, but I’d have liked to make it without the GL render & Text Input on Linux issues. Maybe someday I’ll try to update to Unity 5.3 and check if it doesn’t break anymore my animations, without breaking some other new stuff.

For now, I think I’ll be still using Unity for my future games, although for the very next one, I’ll still use my old custom framework so I’ll be able to use my old game engine without porting it to C# right away.

And to end on a positive note on Unity, it has been really fast & pleasant to be able to build all versions (Windows / Mac / Linux all for Steam and for Mana Games, and the Mac AppStore one), while it was really cumbersome with my old homemade framework.

Final Weekly Update : Treasure Hunt

Posted on by
Hey, that's me ! :-)

Hey, that’s me ! :-)

And this is the final weekly update.

At least for the time being ; maybe in the future I’ll come back to add a Rogue Mode to the game. Or Dungeon Guardians 2. We’ll see… :)

But for now, let’s see what has been accomplished the past 10 days.

I finally found the energy and the time to put the Treasure Hunt in the game. This is a chained series of riddles, more or less easy, designed mostly to test the player’s memory and observation skill.
This hunt features a new extra non-combat item, as well as alternate weapons (eg: 2H tank weapon).

This little extra in the game makes it now really complete, especially as there was a series of rooms & places that didn’t make much sense till now, because they were destined to be part of the Treasure Hunt.

I also wrote a complete documentation with a mini-tutorial for the Map Editor : http://steamcommunity.com/app/409450/discussions/2/458604254469667698/ .

On the Steam front, I added the Trading Cards. And in bonus, you should be able to get the OST as free DLC on Steam within a few hours !

At last,  the after-release service is also more or less done, as I got only very few bugs and requested features to add in the game since a few weeks.

So the main work is done now, almost 4 years after the development started, back in February 2012.

I will still take care of any reported bug, and will very likely add some extra requested features if they are easy to implement, but my mind will now be able to switch to my next game (Tennis Elbow Manager 2, a tennis manager game).

I’ll still likely write a postmortem for the game at some point : maybe in a few weeks, or at a milestone for the sales (ie: when the game will have paid me a somewhat decent salary :mrgreen: ).

I will also maybe write a mix between a game design analysis and a game theocrafty article on the game. A full game design analysis would feel too much of self-cannibalism, as I don’t have enough distance to be objective enough on the perception of the different gameplay mechanisms.

I will also try to contact GOG again, now that the wishlist got more than 300 votes. Maybe it’ll be enough to make them change their mind. If not, maybe knowing that Dungeon Guardians is the best selling FPV Dungeon Crawler released recently will help as well… :P  (ok, this is a relatively easy feat, as there’s not much FPV Dungeon Crawlers, but still, it’s nice 8-) )

That’s all for now. Thanks to have followed this blog, I hope you have enjoyed the ride. Me, I surely enjoyed it, and I’m a bit sad it’s kinda over now. DG took the best part of my mind for most of the past 4 years, and it has been an incredible experience !

See you in not too long time (hopefully!), and Happy New Year ! :)
(I’m french, we have till January 31st to wish it ! :-P )

54th Weekly Update : Steam Workshop

Posted on by
I finally will be able to rest a bit ! (Screenshot by Logort)

I finally will be able to rest a bit ! (Screenshot by Logort)

Last week got its up & downs.

I’m nearly dry, and working on system stuff (ie: the Steam Workshop integration) is really not as fun as creating games.

But I did it, I’m finally finished with it : the game now loads Workshop Dungeons, export and then load Mods to & from the Workshop.

Hopefully, it’s bug free. It may require a bit of polishing, though ; I’ll see what need to be done depending of user feedback.

Right now, in the Workshop, there’s only the No Spiders Mod that I have uploaded by myself (for our arachnophobic friends :) ) : http://steamcommunity.com/sharedfiles/filedetails/?id=573850100 .

Note: to see the workshop items, you have to own the game, as it’s still not officially released at this point. I’ll do that once I’m sure everything runs smoothly, hopefully for Xmas… :)

So now, I guess I’m good for some serious holidays. I’ll stick around to fix any reported bug or glitch, though. I might also take a few hours to add Steam Trading Cards & Badges to the game. Hum, that sounds like half-holidays, but I’ll take that, that’s better than what I got lately..! ;-)

I’m still thinking about the Scavenger Hunt, but still no plan nor release date for it…

So I’ll do at least a couple more Weekly Updates, plus a Post Mortem by February 6th at maximum, to celebrate the 3-month anniversay of the Steam release… :P

Cya soon !

53rd Weekly Update : Store Stories

Posted on by
Maneuvering around big Dragon boss (screenshot by MardoG)

Maneuvering around big Dragon boss (screenshot by MardoG)

Ok, on last week, my productivity lowered to the average one of a corporate employee… :mrgreen:

I took care of the Mac Appstore version on last Monday, managed to do a couple of small updates, but I didn’t have too much mood to get into Steam Workshop, so I only started the heavy work on Friday, and on Saturday I released an update that allows you to upload your custom Dungeons to Steam Workshop, but now it’s still not possible to play them for other users. It should come very soon, though.

But don’t run away right now, I have 3 important news ! :)

So this week, hopefully, I’ll finish everything for the Steam Workshop, so maybe we’ll see a couple of small user dungeons by the end of the year… :)

Cya next week !

52nd Weekly Update : Modding SDK

Posted on by
Dark Crypt (screenshot by fR0z3n.s0u1)

Dark Crypt (screenshot by fR0z3n.s0u1)

Last week went smoothly.

I still did a little bunch of bug fixes, but they were mostly for very special cases.

So I had time to finally work on a new important feature : the Modding SDK, including multi-mod & incremental mod support for the game !

You can find the SDK here : http://steamcommunity.com/app/409450/discussions/2/492378806378228784/ .

It even comes with 2 small Mods : 1 to replace the spiders by skeletons, for players with arachnophobia  ; and 1 with all the built-in portraits of the game, to change them or add new ones.

Except that, I got the answer from GOG : they said it looked like a good & interesting game, but were afraid it was too much of a niche game for their store (translation : it didn’t sell enough on Steam). :(

Actually, sales on Steam are not too bad and I’m relatively confident we’ll get to the point where the game will be worth the money and time invested in it.

Now, only the Steam Workshop support is missing. I hope to able to do it all this week, although I don’t have much idea of the amount of work it represents : it’s going to be a surprise..! :-P

I also should be able to publish the game on the Mac AppStore.

After that, I’m not sure where it’ll go. I still want to add the Scavenger Hunt to the game, so maybe I’ll finally do it… Or maybe take intensive holidays… Or maybe start to fiddle with my next game (ie: Tennis Elbow Manager 2, a tennis management game)… Or write documentation for the Level Editor and Modding the game…

Cya next week to see how everything will turn out ! :)

51st Weekly Update : Blog’s 1st Anniversary, time to look back a bit ?

Posted on by
Xmas in a Haunted Dungeon ? (screenshot by Hellbishop)

Xmas in a Haunted Dungeon ? (screenshot by Hellbishop)

Exactly 1 year ago, I wrote the 1st post of this blog.

Development had started more than 2 years before that, but that post can be seen as the limit between “I’m doing that to enjoy myself” and “I’m going to have to sell this game if I want to do more like this ; and to do so, I have to finish it”.

After that post, I almost never stopped to hurry, or at least feel impatient to finish the part I was working on to switch ASAP on the next one, hoping to finish the game soon.

Now, the game is released, but it’s still a bit early to do a real Post Mortem, especially as I’m still not sure where we are headed in term of sale volume.

So let’s talk about past week.

Work rhythm finally noticeably slowed down : most of the days, I had free time between end of my work and going to sleep ! :-P

I still did a good bunch of little fixes, changes & enhancements, as you can see in the ChangeLog here : http://steamcommunity.com/app/409450/discussions/0/490123727974362306/ .

I finally fixed the Linux input issue (by doing my own version of the text input), but I still didn’t start to work on the modding stuff. It should be very likely for this week, as the bug reports & the feature requests start to dry down… :)

I still didn’t hear back from GOG. I’m not sure if it means they’re not interested, or if they are crawling under the game submissions, or if they are just very busy with the end of year business. Hopefully, I’ll know soon enough.

Cya next week for more news !

50th Weekly Update : Patch Week

Posted on by
Pally vs Skeleton (thanks to Agent Cooper for this nice screenshot!)

Pally vs Skeleton (thanks to Agent Cooper for this nice screenshot!)

Alright, it’s getting better ! :)

I still worked a lot last week, but less frenetically (at least since last weekly update) and got some well deserved sleep, but probably still not enough rest, because I spent a lot of time watching Let’s Play videos on Youtube.

These videos shown me that a lot of players don’t want to spend 2 minutes to read in detail the description of their abilities, and thus just smash the buttons without much idea of how to organize their fight. Game is forgiving at start, but the difficulty quickly raises up so I wonder if they’ll make it to the level 5 or not.

Except that, Dungeon Guardians should arrive soon on the Humble Store (but still through Steam Keys), and I contacted GOG but still haven’t got any answer from them.

So next step is to put the game on the Mac AppStore, but before that, I’d like to add the Modding support to allow adding custom portraits.

And even before that, I have to create my own text input field for the Linux version, because my attempts at bypassing the Unity3D bug seem to have failed… :-?

Cya next week for more news !

49th Weekly Update : Steam Release !

Posted on by
It's done !

It’s done !

Pfffffff… (for the last time ?)

If you follow this blog, you know that normally I publish the weekly update on Sunday or Monday.

Today is Wednesday. That says it all, no ?

No ? Ok, so let me explain..! :mrgreen:

Last Friday I released The Fall of the Dungeon Guardians on Steam, like this :

Naively, like I always am, I thought it’d be a bit rough, with a few little bugs & fixes to do here and there, but it’d settle within 24 hours.

How wrong I was ! It took me almost 4 days to fight bugs and integrate in emergency popular requested features into the game, so everybody could enjoy the adventure nicely.

I think with the last update of this morning, it should be (nearly) ok. And I will be able to finally get some rest, before doing a bit more bug hunting (text input doesn’t work on Linux coz of an Unity3D engine bug :roll: ), and then switch to Modding support (SDK + Steam Workshops) and getting the game published in other Stores (GOG, Humble Store, Mac AppStore).

In term of sales, the game did an ok start, but nothing fancy, I’m still not on my way to buy a Porsche (or more important, fund a more ambitious sequel). But within 6 months or 1 year, I may be able to get my investment in money and time paid back.

Hopefully, it’ll catch on with Christmas season, with the help of the additional stores and maybe a bit of marketing (if I get some ideas or get lucky :) ).

That’s all for now, cya soon for the next weekly update !

48th Weekly Update : Rogue Done !

Posted on by
Hungry Bone Dragon!

Hungry Bone Dragon!

Pffffffff (again), I finished the Rogue and I’m glad I did because I’m near implosion ! 8-O

The week has been ok without too much bumps (except in my neighbourhood, but that’s another story, hum, stories actually, which include birds, dogs, bugs & dead mouse amongst other annoying things), so I managed to do the 2 final specializations of the game : the Duelist, a Rogue tank spec, and the Ranger, the only spec able to use bows & crossbows in the game.

For the Ranger, I cheated a bit… :oops:  I didn’t do anything too special for him (except the bow/crossbow) so he’s kind of a patchwork between several other specs. But except if you do plenty of runs with many different parties, you will likely never notice that… :-P

Anyway, the game is now fully done ! It just needs to be proofread by an English native speaker, as well as to get the French translation (I’ll get to it :-? ) and the German translation, and then I’ll need to build all the different versions correctly (Windows / OSX / Linux) and it’ll be all set to be released for next Friday on Steam, yippee  ! :mrgreen:

Cya soon for more news !