Defining the style

The artist for Scratch is the very talented Jason Kingsley who I came in contact with through Board Game Geek. I imagine I wrote him substantially more and substantially longer emails than he is used to for a pre-engagement set of discussions, but the outcome was certainly worth it in my opinion.

Some back story reminders. Scratch gets its name because the game board and tokens were originally scratch paper and spare change (i.e. a little scratch). The casualness of the ancillary pieces of the game were supported by the casualness of the card design and all together gave Scratch a really accessible feel. It felt like the kind of game you could pick up with ease and toss down when you were bored. I liked that and didn't want to lose it, but I also realized that when I brought a professional on board the cards probably wouldn't look shabby and unassuming any more. So we needed to define a style that still made Scratch feel accessible and approachable while also looking professional and engaging.

The style itself. The benchmark we kept coming back to was "worn leather." High quality leather just keeps looking better and better, even as it gets scratched up and worn out. A solid leather chair looks great in your grandpa's den that stopped aging in the '60's or in a modern minimalist house. It's just a good look. All of the colors in Scratch are just a touch on the muted side and the presence of scratches, gashes, and nicks can be seen throughout. I love it.

We also wanted to incorporate this mentality into the actual content of the cards. So when there is a rocky outcropping, it is rugged. When there is a castle wall, it has been smashed and is slowly being rebuilt. The implied world of Scratch is itself scratched. There are battles here, fortresses and bastions are won and lost before being retaken, and that, too, we wanted that reflected.

You can also see how Jason left the occasional pencil line unfinished or just implied a part of the art. The idea here was to make it as though even the artist was harried or decided things were "good enough" and moved on, just as they would in rebuilding their walls. (To be sure, that wasn't Jason's actual mentality since I watched him go through revision after revision and it takes a master to make something look casual.) Collectively, it sets a style where the artist, the content, and the colors all push a vision of a world that is scuffed up but still worth fighting for, like an old leather chair that everyone loves.

Idle Hands

I have not been idle on Scratch. In fact, I've actually been working on it more diligently than ever this last month but it wasn't in a context that I was able or willing to really share. I was going through the slow and meticulous process of working with an artist to develop the art for the game. The process involved hundreds of emails and hours of back and forth as we hammered out the aesthetic. I am incredibly happy with the results and the end result is everything I hoped it would be. I plan to slowly start dolling out samples and sharing the story of its production in the days to come.

Scratching the Surface: faction powers

Each of the four factions (Kingdom, Monster, Wild, and Undead) have access to a power that no other faction has. In each instance, two cards in the faction have the power: a unique (the defacto 'leader' of the faction) and a rare. No other power appears twice in the same faction.

We'll run through each faction and their power in turn. What you'll see is that the faction powers are among the most powerful powers in the game but also oppose and offset each other in really interesting ways.

Kingdom. The Kingdom power is Loyal to the End. Loyal to the End reads 'Decrease all stats by 1 while [creature] is not controlled by original player.' So if Player A plays King and Player B captures King, all stats are reduced by 1 until Player A gets it back. Obviously, this is really powerful but only defensively. It only comes into play if Player B wants it to come into play by capturing King.

Monster. The Monster power is First Strike. First Strike reads 'Increase all stats by 1 the first turn [creature] is played.' So if Player B plays Dragon, all of Dragon's stats are increased for that first turn, making it really effective at defeating cards in play. The symmetry to Loyal to the End is hopefully apparent; First Strike is powerful on turn one and puts all the control in the hands of the player who plays it while Loyal to the End is powerful in every turn except turn one and puts all the control in the hands of the player reacting to it.

Wild. The Wild power is Winsumloosum. Winsumloosum reads 'If defeated, controller may remove [creature] from game prior to losing control.' The idea here is that if Player A plays Phoenix and Player B subsequently defeats it, Player A can decide to physically remove Phoenix from the board instead of allowing Player B to flip the token and gain control. This leaves an empty space on the board, but Player B's turn likely still ends. Player A can then play somewhere else or play in the empty space to refill it. At first blush this seems extremely powerful, but it is actually a lot more balanced than you might first believe because it is sort of like Player A skips a turn in order to not lose a card. Skipping a turn in Scratch is a big cost.

Undead. The Undead power is Festering Defeat. Festering Defeat reads 'If defeated, controller may swap [creature] with attacker before control is lost.' The narrative idea here is that the undead infected or began to turn the attacker even though they were defeated. Similar to Kingdom and Monster, Festering Defeat is the symmetrical opposition to Winsumloosum. They aren't quite as opposite, but they take very different approaches to the same idea of making attacking the cards inherently risky.

The Wild and Undead faction powers also introduce two interesting complexities to the game. First, Winsumloosum removes a card from the board, which means it reverse the relationship that whoever plays first will be playing last. Having to play that ninth and final card is challenging because if you don't defeat a card in play, your opponent wins. Fear of that ninth card is one of the best reasons to decline to go first. With Winsumloosum, you can play first and not play last, but only if you opponent decides to attack and defeat a Winsumloosum card. It isn't a unilateral decision, so it adds layers of interesting tactical complexity.

Festering Defeat is interesting because it makes Stand Your Ground (the power that denies efforts to swap places with it) more powerful. If you use, say, Knight to attack a card with Festering Defeat, the Knight can decline the attempt to swap.

Take that, Little Guy!

This follows up on my previous post, We're all international backers when it comes to shipping, and addresses some potential weaknesses of that first article.

I went onto Kickstarter and pulled the last six months of successful board game projects. That is, any project in the last six months that received funding and was selling a physical board or card game. I excluded a handful of projects based on some miscellaneous criteria, but I'm fairly confident in the data set being a good snapshot of successful Kickstarter board game projects.

There are two important findings that I think validate the takeaways in my prior article.

(1) As the cost of the game increases, the average backer donation increases. This is going to seem obvious, but let's walk through it real quick. At the outset of collecting data I decided to collect the dollar amount of the reward category where you actually got what they were selling. If they were selling a deck of cards at $20 but you could get three extra promotional cards at $22, I entered $20. If they were selling the base game at $50 and it included every promo they could imagine, I entered $50. It is the lowest reward tier where you actually got the physical product they created.

Next, I took the total amount raised by the project and divided it by the number of backers to get an average backing amount. I then sorted the data by the lowest reward tier where you got the product and plotted the average backing amount. So the left side of the horizontal axis are the projects with the cheapest reward tier to get the game and the right side are the most expensive tier to get the game. The vertical axis are those costs of entry to get the actual game. The trend line is pretty definitive--as the cost of entry increases, the average backing amount increases.

This should feel really obvious, but it is still important to confirm. People are mostly backing to get the game and if the cost of that game is high, then they pay a lot to get it. This validates the premise that more expensive games are required to ensure the margin required to advertise.

(2) The cost of the game has little bearing on the number of backers. This is much less obvious and should be, perhaps, downright shocking. My first article pointed out that the cost of advertising to customers was so expensive that it often exceeded the margin on cheap games, thereby pushing everyone to produce expensive games. This was just validated in (1), above. A rebuttal could be that the conversion rate will be much higher on cheaper games because convincing someone to back a $10 project is much easier than convincing someone to back a $75 project. The data doesn't seem to agree.

I sorted the data by the number of backers and once again plotted the cost of the lowest tier where you actually get the game. What we see is that there is almost no correlation between the number of people who choose to back the game and how cheap it is to back. Now, the data isn't definitive, we don't have unsuccessful projects nor do we know the marketing budgets of these projects, but the line is almost shockingly horizontal (indicating no correlation). This suggests that regardless of the cost of entry, projects tend to have the same conversion rate, meaning the cost of advertising is also equal. The $10-$50/sale figure in the last article holds true whether the game costs $10 or $100.

The next takeaway
A lot of games are simple, fun, and cheap to produce. This means they can be simple, fun, and cheaply delivered to your door... but for the realities of the system. People like watching a game climb into the $20-30k range and that is when Kickstarter really starts to promote it as an editor's pick or 'popular this week;' those are also the points where paid-for advertising takes a back seat to word of mouth. This shows that games with low costs of entry struggle to ever get there.

One conclusion is that we're overly rewarding expensive games. I think a more important conclusion is that we're encouraging makers of simple, fun, cheap games to complicate them and start adding on unnecessary features to justify a higher purchase price because without a higher price they will fail. That's a huge bummer.

A fresh perspective (design theory 2)

When I last talked about my design theory I argued that the fewer rules you have in place the quicker the learning curve for the game which facilitates sharing and getting to that competitive sweet spot with your friends. A really great example of that is a standard card game I invented a long time ago called Equals (1 page pdf). The entire PDF is under 500 words all in so I won’t waste too much space here describing play since it would be about as long.

When we first started playing Equals it was instantly addictive. It is a lightning fast game where all the strategies are pretty obvious early on, which ended up meaning it came down to psychology. You had to get inside your opponent’s head and bluff them with what you were playing while trying to read into what they were playing so that when you get to that last card you don’t find yourself drawing a whole mitt of new ones. It’s a surprisingly fun game.

Then I decided to make it even better. I decided that all those face cards (which are not at all special anymore) were a great opportunity to introduce some special cards. Maybe you could reverse play, skip people, or have wilds. Basically all the standard stuff you’d expect in this type of game. All of them made the game worse. They were flashy and fun and you got really excited when you drew a face card, but the game became more about the rules and getting lucky and less about having to really dig into your opponent’s strategy to predict their final actions.

Let’s take one example and really explore it. One of the changes we made was to decide that Kings would be wild—they match any card. It felt like a really simple introduction that made a lot of sense. Immediately, the dominant strategy was to hold a King for the last card to ensure you wouldn’t have to draw extra. All of the tension was gone. Obviously, we just fixed this by saying you couldn’t play a King as your last card (or, rather, second to last card—this will make sense if you read the rules).

All this did was shift the break in tension one card earlier. Remember, the goal of Equals is to discard all of your cards while trying to match your opponent’s card as closely as possible. This introduces really elegant tension because as you get closer to winning (fewer cards) the probability that you can’t match your opponent increases. If you introduce a wild card to the game, the dominant strategy will be to horde those wilds until as late in the play cycle as possible because that is when they have the most value. Over the course of a long game, the likelihood of getting a wild is extremely high, so all that is really happening is that we know that when you get to the arbitrary point where we decided no wilds may be played beyond this point, you’ll probably play a wild. There is no tension what so ever.

We also punish good play. Let’s say that I have a 4 and a King in my hand and my opponent plays a 4. That means I was really good at predicting what they would play by holding onto my 4. But I cannot benefit from that good play because I can’t save my King for the last card.  So I must play my King and get no advantage for accurately predicting my opponent’s hand.

Finally, ruining the elegant tension that grows throughout play and punishing good play comes at the low cost of adding extra rules that make the game harder to teach and harder to master. Instead of the first few games immediately being a psychological duel, they are a teaching phase where you have to tell your friend, “No, you can’t play that because of this arbitrary rule which had to be introduced to fix this other rule.”

We played maybe 20 games with the extra card rules before reverting back to the simpler format and it has been a better game ever since. If you have 10 minutes, you should give it a try. After two or three games you’ll have mastered the basic strategies and you’ll start to see the psychological side of the game emerge. For this reason, I actually prefer it with just two players to emphasize just how much of a duel it really can be.

We’re all international backers when it comes to advertising

One of the most frequent complaints I see on Kickstarter projects is the cost of shipping for international backers, and rightfully so. It is really expensive and no one feels good about a huge portion of the total cost going to something that doesn't add any value but merely facilitates acquisition. As I've been researching how to conduct a successful Kickstarter project, I've come to realize that we are all international backers when it comes to advertising.

Throughout this article we’re going to use banner ads to represent ‘advertising’ as a whole. There are different ways to advertise, but banner ads are the most common and since we’re talking about industry averages, should be reflective of the true cost of advertising. That said…

Banner ads are expensive. The general rule of thumb is $1 for 1000 impressions. Prices can vary, but most of the big sites seem to fall in this zone. The way they really differ is in the minimum quanitity of impressions they are willing to sell. Some sell by the buck, most sell in bundles, and some only sell in big bundles ($500+).

Regardless of quantity, what we really care about is click through rate (CTR). Every time a webpage loads, all of the banner slots are filled with a banner. That is an impression. If someone hits refresh or goes to a new page, that is a new set of banner impressions. CTR is the percentage of those impressions where someone actually clicks on the banner to be taken to some other site where they might learn more about the product and hopefully purchase it. Note that in all of this we never measure whether or not the person clicking is ‘new.’ If all 1000 impressions go to the same guy and he clicks every time it would look like we had the most amazing banner ad campaign in existence despite the fact that we only ever reached a single person.

CTRs vary by website and also vary during the course of the project. Because many of those impressions are delivered to the same people, and people are more likely to click something when it is new, CTRs tend to fall throughout the course of a project. Many websites will tell you their average CTR prior to purchasing and they seem to fall in the 0.4% to 0.2% range. Sometimes you can also get samples of successful ads that have CTRs in the 1-2% range.

This gets us to our first math: calculating the cost per click. The cost per click is how much we actually paid to get someone to click on our banner and see what it was we wanted them to see. Here’s what it looks like:

So if we assume our banner is going to be on the high side of average (0.4% CTR) we can calculate a cost per click of $0.25 ($1 / [1000 * 0.004]). That is $0.25 for every person that looked, even if they’ve looked before and even if it was an accident and they immediately close the browser window.

Because the point of advertising is to generate incremental revenue beyond the cost of advertising, we need to next calculate the conversion rate. Conversion rates are hard to calculate. The overall global average seems to be about 2.5%, but this is for mature companies who know what they are doing. Amateurs seem to average closer to 1.5% and a lot of reports I’ve been reading are indicating closer to 1% for Kickstarter advertising.

Sidebar—I have a hunch for why this is. For most advertising, someone clicks and makes a determination. For Kickstarter advertising, someone clicks and makes a determination, but then also becomes interested in watching the progress of the campaign. Many, many times I’ve seen a banner and thought, ‘Oh hey, I wonder how they’re doing.’ I think Kickstarter conversion rates are artificially low because the information the advertiser wants me to see (why their game is great) at the other end of the link is different from the information I might be going there to see (how much money they've raised).

Let’s take the global average, 2.5%, and presume that we’ll be more successful than most Kickstarter projects. The cost per sale looks like this:

So if we assume our 2.5% conversion rate along with our 0.4% CTR, we come out with a cost per sale of $10. Just getting the sale cost $10. Real quick, if we change our assumptions to a 0.2% CTR and a 1% conversion rate, that cost per sale jumps to $50. Note that both of these changes are actually closer to what most people actually experience. A cost per sale nearing $50 is not unrealistic!

This probably seems pretty shocking and you might even wonder why so many people bother advertising if what I am saying is true. The reason is that you need this advertising to generate a baseline of support. A lot of people will only back successful projects because they assume it means other people vetted it and it is worth backing. Kickstarter only promotes projects that are already doing well. When a project has a lot of backers, word of mouth kicks in and people like to check in from time to time and see how high the numbers will go. I checked the Double Fine Adventure Kickstarter daily for the first few weeks of the project just to marvel along with everyone else at its rise. Somewhere along there I decided to back it. Small, new game designers can’t get that level of exposure on reputation, so they have to buy it. They buy it through ads.

There are two important takeaways from all of this. 

First is that these costs, the $10 to $50 per sale, are real costs. They are real costs that are passed onto the backers because if a project isn't profitable it is going to fail. These costs also don’t add any value for us once we get the game. Just like shipping, they are this huge portion of the total we have to pay that we could avoid if we just had a better system.

Second is to think of the ramifications this has on what types of products are produced. The costs are per sale and whether I’m trying to sell a $200 game or a $10 game, those costs are the same. What differs is that the $200 game has a lot more margin to absorb the cost of advertising. The $10 game has almost no margin and so traditional banner advertising is a fool’s errand. Since banner ads are the main way to generate that initial enthusiasm that builds into momentum that culminates in a successful project, we’ve essentially rigged the system to ensure that fun, cheap games struggle to get produced. 

Scratching the Surface: card frequencies

Cards fall into one of three frequencies in Scratch: Common, Rare, and Unique. Each of the frequencies plays an important role in the game and conveys significant information. We’ll run through each frequency in turn and while doing so talk about what information it conveys and how it impacts strategy.

Commons

Common cards are the weakest of the three frequencies. Every card is built with formulas and the formula for building commons is the weakest. Commons also don’t have powers. These two facts are important when you remember that at least half the cards in your deck (5 out of 10) have to be commons. This speeds up play because it ensures that you aren’t juggling a bunch of cards with varying powers around and also introduces more action to play because commons are easier to beat. The game would be awful boring if cards couldn’t be gained and lost, so commons help increase the overall tension.

There are no restrictions on commons in play. You and your opponent can play duplicates of the same common. You could make a Goblin deck where you only play Goblin. Good luck winning with that, Champ, but you could give it a try.

Rares

Rares are the middling power of cards. While they are unabashedly more powerful than commons, you might notice that few rares are able to have a 7 while many (if not most) commons can do so. This is important because it ensures some baseline of competition between the two frequencies. Commons can often overcome rares, albeit only on one side.

A rare in play blocks the same instance of that rare. What this means is that if I play ‘Griffin’ and you also have ‘Griffin’ in your hand, you cannot play it. Similarly, if I have a second ‘Griffin’ in my hand, it is still blocked. We talked a bit about why you might want duplicates of the same card when you can only ever play one instance the last time we Scratched the Surface. This rule won’t come up too often if you have only a single set of cards that you are divvying up, but if two people each have their own set, it may. It adds an important strategic aspect to deck building and match play where you start to learn what your opponent might play and seek to block it in advance.

Uniques

Uniques are the most powerful cards. They have the biggest numbers and powers available and can really change the shape of a game. Because they are so powerful, they are also very vulnerable to being blocked. Only one unique from any faction may be in play at a time. So if I play a Kingdom unique, all other Kingdom uniques are blocked. That puts a lot of pressure on deciding which uniques to include in your deck and also when to play them. There are a lot of really attractive unique cards that you might want to play, but you’ll need  a good balance of factions. If you build a deck with three Monster uniques, you will only ever get to play one (or none if you get blocked early!) and will have to discard the rest. That is a big chunk of your deck to discard.

Recall that the first card on the board cannot be a unique. This simple rule really opens up the possibility of wanting to go second to ensure you get to play your unique. There is a constant sense of pressure to use any unique in your hand because you know that at any moment someone could play any unique from the same faction and make it obsolete. It doesn’t feel great to discard a really powerful unique that you were just about to use. On the other hand, it feels really great to watch your opponent discard a really powerful unique that would have won them the game. It’s fun.

Surface Scratched!

Card frequencies are a simple method of dividing the cards into various levels of power. They have been specially designed to ensure that they all can still compete with each other, so even if you just toss together a deck it will be competitive. The strong cards have weaknesses (uniques can be blocked) and the weak cards have strengths (unlimited commons which can still have 7s). Collectively, they ensure a fun mix of cards that develop a lot of action and make sure the game doesn’t get too cumbersome with too many complex cards.