In my last post I talked about how I decided to make a match-three-endless-runner hybrid. The next task was to work out how such a thing would work. I don’t know of any magic short-cuts here. My ‘strategy’ was mainly to think about it for a while!

Straight to the Source

The first thing I did was to identify what the important elements of each of my target genres were.

Common Match Three Mechanics

• Grid or hexagon tiled play space
• Multiple tile types
• Swap tiles to edit a play space
• Match lines of three to make them disappear
• Tiles can only be swapped which lead to matches (non-matches are cancelled before play continues)
• Board rearranges in some way after match to fill the space (e.g. everything drops down to fill the space and new tiles appear from the top)
• Points are scored for matches, more points for longer lines
• Combo bonus points are usually given for chains of matches within a single move (i.e. If the tiles fall down and a new match is formed) or if two lines are created in one swap
• There are usually special tile types e.g. Wild cards (match with anything) or bombs (clear an area on the board when matched)
• The objective varies a lot depending on the context. Often tile types are used to represent an action (Puzzle Quest), all of a certain type of non-spawning tile type must be removed (Candy Crush) or just a high-score must be achieved (Zoo Keeper)

Common Endless Runner Mechanics

Endless runners are less formulaic than match three but there are some common themes:

• The game involves moving in a direction
• The play area moves on continually and is not entirely within the control of the player
• The player must overcome obstacles to keep moving with along with the play area
• There is often a speed based risk-reward mechanic (e.g. Combos and speed in Alto’s Adventure)
• Play generally continues indefinitely until some kind of failure condition occurs.

What Mechanics to use?

Looking at these lists, there are no doubt many ways that the two genres could be combined by picking different subsets from each list. Previous games like 10,000,000 and We Must Build a Boat have already combined the genres in a way but have opted, like Puzzle Quest does with adventure games, to have a Match Three sub-game which is used as a metaphor for the player performing an action such as fighting.

However, to truly combine the genres into one I wanted to make the endless runner be actually within the match three board so the two existed in the same play space. I decided that meant I needed a constantly moving board of tiles which could be somehow matched in lines.

Skiing down a mountain of tiles seemed like a natural fit. As the tiles would represent the terrain it didn’t make sense for new tiles to drop down the board when matches happened so I decided replacing them with snow tiles was a better option.

I considered how the player would be represented in the game for a while but couldn’t think of a nice way for the player to control the skiing character. In the end I opted for a skiing character with an automated pathfinding algorithm. I tried to make the path finding algorithm do the ‘expected’ thing as often as possible (note: this is often NOT the most optimal thing) but after play-testing decided that even that wasn’t enough and visualising the player’s path was also required.

The final part of the ‘make it a game’ requirements are that there is some kind of win (or actually lose in the case of endless runners) condition. I experimented with a constant camera movement speed and losing if the player dropped off the back but it felt unnatural so in the end decided to just have the game end if the player crashes. Once I’d tried that, it seemed so appropriate I did wonder why I hadn’t just done that in the first place! But that’s the point of the iterative process in game development. Try things and see what works the best.

That is the core loop of Ski Three and it’s what I implemented in the initial prototype. I thought it was a fun base for the game.

The level in which Ski Three plays out is procedurally generated with trees, houses and rocks. The procedural generation algorithm ensures no matches are ever generated and also has a gradually increasing difficulty as well as difficulty waves (where the density of tiles increases and decreases quite a bit locally as the overall density increases gradually).

Layered Mechanics

With the core loop in place I was able to layer some mechanics on top of that.

One of the things I really love about Alto’s Adventure is the risk and reward nature of combos and speed. I tried to recreate that in Ski Three using a match three combo system. If you make chains of matches close together without moving any other tiles then your character gets faster. Your scores for each match are also multiplied by the combo multiplier. It’s a fairly simple mechanic but I find it makes the game a lot more interesting once you’re getting good at it. And some people are REALLY good at it. The overall high score currently stands at 2,405,100 – which is approximately 2.2 million better than my best ever score!

The second layered mechanic I added was global leaderboards. Though I get really frustrated with global leaderboards in general as I tend to just be the N thousandth or million’th (depending on the popularity of the game) best player in the leaderboard giving me a strong sense of my mediocracy at playing games – presumably not the intended emotion by the developers?

I decided to do things a little differently. The game doesn’t actually tell you where you are in the leaderboard, just what the best score is. That would be pretty rubbish if it only ever showed the best EVER score which is way out of reach for most people so instead it has several leaderboards – one for the last hour, one for the last day, one for the last week and one for best ever. So there can always be a challenge for the last period of time which naturally is the one just better than you are. Of course, occasionally the best score in the last hour is the best score ever, but that only lasts for an hour and then it’s back to sensible scores (usually just a few thousand). I think that the leaderboards are the best part of Ski Three and I’ll definitely consider how to use them less directly as Ski Three does in future games.

The final ‘mechanic’ which is really a concession to Free-To-Play is that there is an in-game token system and you can only play if you have tokens. Tokens can be earned by:

1. Coming top in a leaderboard (ie. best score in last hour)
2. Waiting a day
3. Watching an in-game advert
4. Buying a Ski Pass (which gives you infinite tokens)

And that’s it.

What would I like to have added?

I intended there to be one further mechanic in the game which had to be axed to make the game fit into the development time I had available. I think it would be an interesting bit of game design though, so I’ll mention it briefly here.

It’s the Yeti! Here’s Ara’s awesome concept art for him:

The Yeti design was to be a character who occasionally appears ahead of the character (they appear on a tile as the character skis down the mountain) and upon getting a clear path to the character will chase them. The player must use the tiles as obstacles to stop The Yeti from being able to get to the player, which means avoiding matches – which would make the tiles clear away, while still matching three to progress. I think it’s an interesting idea and would nicely turn the match three mechanic on its head.

Summary

Ski Three is certainly a simple game, but it’s well executed and could be considered minimalist design rather than simple :-). I’m proud of it as my first attempt at designing and building a game from scratch. It’s been fairly successful by my standards in terms of number of downloads (over 25,000 so far), however I was unwilling to follow the Free-To-Play norms to the extent that would create a profitable game. Most of those norms feel like exploiting the player to me (eg. selling in-game currency the player cannot understand the value of) or at least a hindrance to the player enjoying themselves (eg. blocking gameplay at key moments with pay walls) which seems contrary to the job of a game designer. So if you’re looking for a free-to-play game that doesn’t exploit you – play Ski Three. I can tell it certainly doesn’t exploit anyone because I’m not making any money from it!

I’d set up Cape Guy, finished the 80 Days PC and Mac port and arranged a 6 month contract at Microsoft Lift to bring in the funds to cover a 6 month project.

I was ready to make a game of my own…

But Which Game Should I Make?

Deciding which game to make is essentially a blank sheet of paper problem. It’s a huge decision because it affects the entire project and decides what you’ll be doing for the duration of it. Getting it ‘wrong’ could also hinder a project before it has even really started. Though the paper isn’t really blank, it’s actually swimming with little proto-ideas (often inspired by playing other games) jostling to be the one you think is worth making!Continue Reading