Spellsworn - Wizard duels with sumo-wrestling rules

Hello there!

We’re Frogsong Studios, and for the past year we’ve been hard at work creating a free to play game called Spellsworn in Unreal Engine 4. In approximately one month we’re going into early access on Steam, and we haven’t really had the opportunity to talk a lot about our game in full on the forum yet, so we thought we’d give it a go. We’re still looking for people willing to test out our game, and we still have quite a lot of keys to hand out - so if you want to take it for a spin give us a heads up.

The inspiration for Spellsworn is actually a mod for Warcraft 3 : Frozen Throne called Warlock. We’ve all played the hell out of that mod, and still love it to death. It’s always been a favorite on any LAN-party we’ve been to, and will probably remain so for a long time to come. So we decided we’d want to make our own standalone game based on that mod, and so Spellsworn (and actually Frogsong Studios) was born. Spellsworn is all about hectic arena based combat, with the tactics and strategy of a MOBA. You have people on one side trying to knock you out into the lava, and people on the other side dying to be scorched by your fireballs, and somehow you need to make a split second decision in the middle of it all. It’s disorienting but fun at worst,exhilarating and making you feel unreasonably good about yourself at best.


The gameplay can sort of be compared to Super Smash Bros in its’ hectic decision making, but also to LoL, DotA - even Counter Strike in its’ tactic and strategy. You start off with only one spell, but have the option to buyi and upgrading a couple of more spells using gold before the first round of combat. Between each round of combat there’s a grace period and a shopping time for you to increase your spell arsenal - or upgrading already bought spells - in order to wreak more havoc in the following round. It’s important to keep track of what your opponents are starting to upgrade and buy in order to counter them and their spell lineup. Sometimes it’s best to skip buying an offensive spell if you notice the enemy has a strong offensive combo; in which case it may be best spending your cash on a transportation spell. There’s many different spell lineups to explore and master ( I usually go for a glass-cannon build which means I die half the time I try to execute a massive combo )](, and equally many that will destroy and stomp on your specific lineup. It’s what makes the game fun and balanced.

There’s a total of five rounds of combat before a victor is declared, and then XP and rating points are distributed based on your performance. Casual play is of course totally possible (ie, locally hosted custom games) - but we personally like the rating system found in many pvp games these days ( and the glory and titles that come with it ). Before a match starts you gather in a lobby, so you can taunt each other and show off your awesome skins and characters. There’s currently a myriad of team setups which we may want to limit in the future; but at the moment it’s possible to play 1v1, 2v2, 3v3, 4v4, 2v2v2v2, 2v2v2, and FFA with up to 8 players. We’re trying to focus the balancing on 3v3 and 2v2 at the moment, but that may change depending on where the players want to take things.

Before entering a lobby you have a main menu with a few buttons giving you the option to change skins and characters, which brings us to monetization. It was very difficult for us to establish a method of payment for Spellsworn. None of us like ‘pay to win’ titles, and some of us have a neutral feeling towards ‘free to play’ in general - but we still felt it would be best to go the free to play route since the player count is what makes the game fun to play. Spellsworn without any foes makes for a boring game. With this decision came a huge amount of work, which we didn’t really expect starting off. First; you need to be able to find a sustainable way to keep the match servers up and running around the clock, which isn’t cheap, without an initial income - second; you need a large pool of characters and skins for the players to use for customization and purchasing, otherwise nobody will buy anything, and lastly; the infrastructure behind such an item-based game is rather difficult to setup tech-wise. In hindsight we would probably never have gone down the FTP-route had we known of these problems, but now we’ve dealt with them and feel confident we can pull it off.

It’s been a pretty long and at times shaky journey thus far, but we’ve stayed true to our goal and see no reason to quit. The development will most likely stretch out into early next year (before we’re ready for full release), and we have lots of features planned. The spell-idea-department is running a little dry, so if you have any suggestions please do share.


EDIT: I forgot; here’s our kickstarter which ends in roughly a day :eek:

Hi! Really nice game! I like the art style and the definition of the effects.

Since this is a UE forum, I also have some technical question. We are also building an arena fighting game of some sorts (but not top down view), and we could use some of you experience:
a) can you elaborate more on the difficulties you have encountered when going free to play; was the difficulty in creating the persistent database server where you keep per-players data and communicate with clients, or the amount of work required to build all the additional buyable items. I think the latter could also be done iteratively (you don’t need 10 skins at the start).
b) how do you host games; do you create a dedicated server somewhere and link it to all players when the “matchmaking” server finds a match? Also, do you have any more exact figures how much do servers cost, and how many cores/ram do you need to host the game? By the way, does Steam offer any way to host servers there?
c) how long are you in active development, and how big of a team are you? How many programers/designers/artists?
d) how big of a portion of the project is done in blueprints (in our case, we do most of the logic (90%) in blueprints, just heavy stuff in C++?

Best of luck!

this looks awesome. looking forward to checking it out!

Hey! Thanks!

Sure, I’ll answer best I can.
a) The four major points of frustration for the FTP model are the techwise infrastructure which needs a solid database handling (however you may chose to handle accounts), the transaction infrastructure, the content production and the “initial cost without initial income” hurdle. Even though you may do an iterative process for the purchasable content it will still need a lot of the production time, much more than level designing or props. Making a new character with a new skeleton and a new mesh painting takes us about three to five weeks from start to finish (polish), and we’re aiming for up to ten characters on release - which equals roughly thirty weeks, and then 10 characters are still a bit few from a players viewpoint. The tech side of things is always something that can be understood and learned, but it’s still time being dedicated towards something other than in-game features and bugfixing - which for a five-man team can be crushing. Using steams API’s has been a bit of a blessing, but there’s quite a few features missing in the Unreal department when it comes to Steam API wrapping. So we’ve often needed to work around Unreal’s wrappers.
b) We have two ways of creating games at the moment. There’s custom games which is your standard hosted game ( menu -> lobby -> game ) which shows up in a server list, and then there’s the matchmaking games which are hosted on dedicated machines around the world. In this case we’re using Steams matchmaking lobbies to find opponents and then from those lobbies find a suitable dedicated server, which we then let the lobby join (and reserve). When joined the dedicated server sends an API call to our dedicated server manager which fires up two new dedicated servers, just to keep up with demand. When they’ve idled for an extended amount of time they automatically shut down. Steam has no dedicated server handling, that would be way too expensive on their end, but they do have a master server which you can use for server listing. A quick google search for different hosters will most likely tell you the cost of a single VM over a monthly period, and the one we’re going with costs us roughly $40-$60 per month per VM.
c) We’ve been working full time for about a year now, and we’re two programmers, two artists and one designer.
d) Almost everything is done in C++, only the UI elements are done in UMG/Blueprints

Thanks, and same to you!

I’m glad you like it!

Looking good!

Thanks a lot, Chris! Couldn’t have done it in such a short amount of time without the Unreal Engine. A true blessing.