The Hard Data Behind Play-and-Win

3 June 2014 | 31 Comments

UPDATE: If you represent a convention that has a play-and-win section, please list it on this Google Doc so publishers can send you copies of their games.


A few months ago, I wrote a post on this blog called An Open Letter to Small Game Conventions from a Tiny Publishing Company. With over 2,500 Facebook Likes, it is by far the most shared post I’ve written.

The post describes an innovative way for publishers to offer people exposure to their games at conventions. It’s called play-and-win, and it was created by Jay Little when he started a convention called Geekway to the West in St. Louis about 10 years ago. I had the pleasure of meeting Jay at Geekway a few weeks ago.

Play-to-win works as follows: Publishers donate games to a convention’s play-to-win section. Any convention attendee can check out a play-and-win game (or play any copy of that game in the game library or that someone brought to the convention). When they return the game, they write the names of everyone who played it (or taught it) on a piece of paper, which they deposit in a jar next to that game. At the end of the convention, one person who played the game will win it.

You can go read the original post about why this is a great system for publishing companies and gamers. Today I want to talk about data.

Marshall Jansen, one of the people who currently runs Geekway, was very generous to share with me the data from all 67 games in this year’s play-and-win section. I asked for his permission to write about this data on the blog because I wanted to see if there are certain patterns that publishers should take into account when submitting games to play-to-win sections.

So I took Marshall’s data and added a few columns (Board Game Geek rank, year published, max # players, and max game length). You can see the full data here, and I’ll break it down below.

Note that the metric I’m using to determine which game was the “best” at Geekway is based on total number of players over the entire tournament. The key for a play-and-win game is exposure–you want as many people to play it as possible. These aren’t unique views, as the same person might have played a game several times. Also, the data is a little skewed because most of these games had more than 1 copy in the play-to-win section, and as I noted, you could play copies of these games from the game library or those that attendees had brought to the convention.

sushi goTop 10 Games Overall

  1. Sushi Go (440 players)
  2. One Night Ultimate Werewolf (422)
  3. Rampage (317)
  4. La Boca (271)
  5. Skull and Roses (266)
  6. Euphoria (253)
  7. PIX (218)
  8. Glass Road (217)
  9. Lords of Waterdeep w/expansion (216)
  10. Hanabi (206)

The average BGG rank for those games is 547. Most were published in or after 2013. All accommodate at least 4 players, with an average max player count of 6.1 players (vs. actual average players per game of 4.1). And the average game length (as estimated on BGG) is 40 minutes.

The key is that these are new, short, larger-group games. That very much resonates with my experience of Geekway. As a gamer, I love long (60-120 minutes), complex games, but when I’m at a convention, unless someone is going to teach me how to play, I don’t want to read an 18-page rulebook. Does that mean that Sushi Go is a better game than Terra Mystica? Nope. Is it a better game for a play-to-win section? Absolutely.

Perhaps the most interesting thing to me about that data is the BGG rank average of 547. 547 is nothing to scoff at, but it’s still pretty low. I think that’s a good thing–us publishers don’t have control over our BGG rank, but we do control our player counts, publishing dates, and game length.

But let’s not discount the impact of games that fit into a different mold. Sure, getting 440 plays would be awesome, but it’s still great if your game saw 100 plays. Let’s dig deeper into the data and look at the averages from the top 25 most played games.

Averages from the Top 25 Most-Played Games

  • Average Players Per Game (actual): 3.7 players
  • Max Number of Players: 5.5 players
  • BGG Rank: 512
  • Year Published: mid-2012
  • Max Game Length: 47 minutes
  • Average cost: $30

This is fairly representative of what we saw in the Top 10. When you choose games to donate to play-and-win sections, choose games that can play 5-6 players but that play well at 3-4 players. BGG rank doesn’t matter much, but make sure you have enough plays logged on BGG to get close to the top 500. Your game doesn’t need to be brand-new, but it should be new enough that it’s still in the public eye. And your game should be playable in under an hour.

There were a few helpful comments after I originally posted this entry about looking at the play-to-win data from a different angle, so here’s one other perspective. I wanted to look at the return on investment (ROI) when a publisher donates a game to a play-to-win library. This formula may be WAY off, but it’s something to consider.

Instead of just looking at overall plays, let’s look at how much money publishers will make by donating a game to a play-to-win section. The formula is: (total # players) x (% who will purchase the game) x (cost of game)

The % who purchase the game after playing it at the convention is completely hypothetical, but I’m going to say 7% for a brand-new game, less 1% for every year since publication (i.e., 7% for a 2014 or future release, 6% for 2013, and so on, with a minimum of 1%). Of course, this is very inexact data, because if no one likes your game, they’re not going to buy it. Perhaps future play-to-wins can let players rate the game?

I pulled the cost of each game from the banner under each game on BGG. This isn’t the MSRP, nor is it what the publisher receives. It’s just an anchor price to give us a frame of reference to compare faster, smaller, cheaper games to bigger, more expensive games.

Top 10 Games Based on ROI

  1. Euphoria (ROI of $797)
  2. Rampage ($742)
  3. Lords of Waterdeep w/expansion ($724)
  4. Glass Road ($586)
  5. La Boca ($504)
  6. One Night Ultimate Werewolf ($502)
  7. Suburbia w/expansion ($488)
  8. Russian Railroads ($403)
  9. Space Cadets ($403)
  10. Tomorrow ($392)

Some very different games occupy that list, and it’s clear that it’s absolutely worth the small investment from a publisher to donate a game to a play-and-win section. Here’s a profile of those games:

  • Average Players Per Game (actual): 4 players
  • Max Number of Players: 5.6 players
  • BGG Rank: 547
  • Year Published: 2013
  • Max Game Length: 65 minutes
  • Average cost: $47

Compared to the previous schemata, the player count, length, and cost per game are higher here.

Also, just for kicks:

LordsTop 10 Games by Plays from the BGG Top 100

  1. Lords of Waterdeep w/expansion (216 players; BGG rank: 27)
  2. Russian Railroads (164 players; BGG rank: 54)
  3. Lewis and Clark (155 players; BGG rank: 98)
  4. Love Letter (153 players; BGG rank: 86)
  5. Terra Mystica (144 players; BGG rank: 6)
  6. Tzolk’in w/expansion (128 players; BGG rank: 15)
  7. Pathfinder Adventure Card Game (126 players; BGG rank: 49)
  8. Suburbia w/expansion (122 players; BGG rank: 53)
  9. Kemet (112 players; BGG rank: 89)
  10. Eldritch Horror (111 players; BGG rank: 41)

What did you learn from this data? Do you see patterns that I’m missing? Thanks to Marshall and Geekway for sharing this data!

Leave a Comment

31 Comments on “The Hard Data Behind Play-and-Win

  1. After attending Origins 2015 this year, I love what Asmodee did with Splendor. Instead of doing a play-to-win, they created new game components for Splendor. The new components do not change the game, they are just some upgraded pieces. If you entered the tournament, you would automatically get a small cardboard punch-out of new nobles. If you won your way into the second tier, you received one set of upgraded gem tokens (just like the normal ones, but with a translucent color chip instead of solid color). If you won your way to the next level, then you got a complete set of new gem tokens (but not the golds!). And finally, if you won it all, you got everything plus the new golds and a game mat to play on.

    I like that it gives your entire fan base a reason to play. It was exciting to play against people who know and understand the game. There were tons of demo sessions, so I knew beforehand that I wanted to (and did) buy the game. But unlike most games that I buy, I will probably seek this out and play again in Origins 2016 to keep updating my components. But for games like “Bring out yer dead” which did a play-to-win (that I couldn’t make it into – so I bought that game too), I will check for expansions and then likely just move on quickly.

    I think that’s what makes this a great concept for game companies, because now you have a larger draw by bringing in your fans and it will help drive interest as other gamers walk by and see lots of people playing the game. I also think it solves the problem of people not holding out to buy thinking they might win (and in the meantime, perhaps purchase a different game instead of yours!)

  2. Ah, you know I love data! I have a few thoughts. I wrote a long note, so I’ll summarize my thoughts here and give reasons below:

    (1) Make your BGG comparisons against the same timeframe; if the games on Play-N-Win are from 2011 to 2013, restrict BGG search to games from 2011 to 2013.
    (2) Make players more likely to buy the game if you want to boost return from donating:
    a) Reduce MSRP if players buy right after playing.
    b) Offer credit instead of game for Play-N-Win if they win the game but bought it that weekend
    (3) Get larger games to boost BGG posts
    a) Offer an additional game in the giveaway if a certain number of players is reached.

    Long thoughts:______________________________________________________

    For Charles and other Geekway organizers – what if you added a “How Likely Are You to Purchase This game?” question to the back of the play-n-win tickets? That may not be a reliable metric but it’s at least something. It could even be “Do you plan on buying this game here today?”

    A related thought – I hold off on Play-N-Win purchases because I might win it. If there were a rule of “You could win this – OR if you’ve purchased it already, we’ll give you a $25 Miniature Market credit” then I’m more likely to buy because I don’t wind up with two copies of a game when all I wanted was one.

    For Jamey,
    I like your BGG comparisons. I’m not surprised that the average rank is mid-500s, though, and for the reason that these are newer games that haven’t yet risen (in BGG’s ranking algorithm) to the top 100. Sure, some games really rise to the top quickly, but the Top 100 are full of older games. 36 of the top 50 (as of today) were published prior to 2012. In other words, the Play-n-Win selection is biased to begin with towards new games because it’s promotional in nature. A more apt comparison group on BGG would be games released in the past 18 months.

    I think the Number of Players also reflects a Geekway skew. You’ve gamed at my table, you know our number of players tends to match whoever comes that evening so it could be 3 or 5 or 7; but Geekway tends to be lots of twos – couples, me and my buddy, a person who played by himself playing with another person who came by himself. If you looked at a histogram (bar chart with number of players 1 to 8 in the bars, and size of the bar determines the frequency of that number of players at a table), I bet your evens would be taller than your odds with 4 as the mode (most frequent bar). That’s just how we socialize. There’s also a competitive advantage to 4 players – 3 players is the smallest group in which one player can be defeated purely because of an alliance, so I’d argue we prefer 4-player games to 3-player. We even design that way.

    Why don’t we see more 6-player games? I think because it comes back to 2’s – it’s easier for 2 people or 2 pairs of people to negotiate and agree upon a game, than it is for 3 people or 3 pairs of people. Again, humans prefer a simple negotiation (balance my needs versus yours) to a complex one (balance my needs, your needs, and his needs).

    I really like the ROI metric and I see where you’re trying to incorporate “cheap” as in “hey this game fits in what’s left of my budget for the weekend, I’ll play it and see if I like it.” That’s the opposite of my approach – “Hey, I can’t afford anything else this weekend so I’ll play the most expensive game here because I won’t be able to play it otherwise unless I can get Jamey to buy it.”

    Something else you said:
    “The key is to get lots of people to play your game. If you have a 120-minute game that attracts 50 players (6000 total minutes played), is that really the same as a 20-minute game that attracts 300 players (also 6000 total minutes played). ”

    No, it’s not the same, but which is more desireable?
    From the perspective of boosting revenue, it’s a matter of (units sold) x (price per unit) – (cost of donation) = Net Revenue, and you elaborate (units sold) further to (players) x (likelihood to purchase). Maximize likelihood to purchase, even if that means lowering MSRP (for example “Thanks for playing in Play-to-Win! If you bring this card to the booth within 20 minutes of finishing, you can purchase for $5 off!” In other words, if likelihood to purchase is 1% and players is 500, you sell 5 games. It’s easier to double sales by doubling likelihood to purchase than by doubling number of players, so make the game easy to purchase.

    From the perspective of boosting BGG visibility, it’s a matter of maximizing number of plays, which means have more units available and encourage more players per game. Think of it this way:
    Number of new BGG posts = (number of players) x (likelihood to write a post) and (number of players) can be thought of as (number of games) x (average number of players per game). It’s easier to boost number of players than number of games, and you can do so by having a conditional reserve, say “We have 3 copies of . If 500 people play this weekend, we’ll give away a total of 4 copies!”

    1. JT: This is a great analysis–thank you for sharing. The point that sticks the most with me is finding ways as a publisher to encourage people to buy the game after playing it in play-and-win. Including a special coupon with a play-and-win game could work really well. The problem for us is that we ship all of those games to conventions from Amazon fulfillment, so it’s not like we can tape a little card to the box before sending it. But perhaps I could find a way to start to personally mail coupons/vouchers to conventions that receive play-and-win games.

  3. Do other conventions do play-to-win or just geekway to the west? How does one add a game to the play-to-win event?

    Thanks for the post

    1. Seth: It’s my impression that after my original post about play-to-wins went viral, a number of conventions started adding play-to-win sections. You can look on any convention’s website for contact information–I’m sure they’d appreciate your donation.

  4. Thanks for pulling this info together, Jamey. It’s a very interesting read. There are so many factors to consider when trying to determine if a Play and Win title was “successful”. It’s a three way evaluation — the player, the convention, and the publisher each have a stake in it. Each target has its own metrics for success, whether it’s unique plays, unique views, player involvement, the lottery excitement, etc.

    One thing that I find especially appealing about the Play and Win model is how it encourages active participation during the convention. Rather than spending a lot of down-time wondering what to do or trying to find gamers, the Play and Win games create a type of excitement that attracts a wide range of players.

    You rarely have to worry about finding enough people to participate in a Play and Win title — the possibility of winning the game at the end of the convention is often attractive enough for someone to sit down and try something that may otherwise not be their preferred style of game.

    That excitement, and the overall enthusiasm of the players, helps combat some of the introverted-ness seen in many gamers when dropped into such a large, context-less open-game environment. If nothing else, Play and Win games add a layer of context. It creates an immediate common ground with the other players. We’re sitting down not just to play this game, but to get involved in the meta-game surrounding all of the Play and Win games. We’re participating in something that extends beyond this one session, and it’s something we can continue to participate in and a shared experience we can continue to enjoy over the course of the entire convention.

    1. Jay: Thanks for sharing the convention’s perspective on play-to-win. I really like what you said about creating a very friendly, meta-environment that is particularly welcoming for introverted gamers like myself. As you said, there’s also an immediacy to it–it eliminates downtime and gives attendees a default thing to do. It’s an interesting juxtaposition with the gaming library (Geekway has a great library), which had an abundance of games I wanted to play, but psychologically I only wanted to play the play-to-win games (that’s probably better for the publishers than the convention, though).

      Thanks so much for coming up with such a great innovation, and I look forward to hearing from all the various conventions that try it for the first time this year!

  5. Those are definitely great exposure numbers and I can see why you’re excited about play to win. But for conventions like ours, which aren’t getting 1000 attendees (yet), and aren’t open 24 hours, those numbers aren’t going to happen. Not even close. We don’t have a lot of people taking advantage of open gaming. Maybe that would change if there was a chance at winning games.

    That being said, since your blog post we have discussed play to win a lot and have tried to figure out how to make it work on a smaller scale. We plan to have a few play to win games during the convention. Few enough that we can be confident each will get enough exposure. We are also hosting monthly game days around the city and are featuring games at each to be played and then given away. Then as our attendance numbers grow, our play to win plan can grow as well.

    1. Kelly: Thanks for your comment. I think this line of yours is key: “Maybe that would change if there was a chance at winning games.”

      I’d recommend giving it a try. I think you’ll be very pleased with the excitement it generates. I’d love it if you followed up on this post after your event!

  6. I would think the real value as a publisher would be the marketing impact. What would it have cost to put the game in front of that many (unique as Charles mentions) gamers? Is that something you can even measure?

    But yes, an addition of “likely to purchase” would be a great addition. I know that the play to win table has lead me to buy Sushi Go as a direct result. Never would have given it a 2nd look without this event. But someone in my group got it, and I love it.

    Luckily, I don’t have to buy Euphoria, since I happened to win it!

    1. Congrats on the win! That’s great.

      That’s a good question about what it would cost to put the game in front of that many gamers. I honestly have no idea. The alternative is probably still convention related (demoing at a booth), but I think the ROI of a play to win is much higher than a booth in terms of dollars spent vs. earned. Of course, there is something intangible about getting facetime with a publisher or designer at a convention.

  7. Like all data mining, it depends on what you want to get. The number of unique players that played the game? The number of plays? Utility (how much was the game played over time – did the game(s) sit on the table half the time)?

    As a publisher, I probably want to know how many unique players tried the game and the utility rate. As the Geekway (who is buying a lot of those games), number of plays still doesn’t mean much to me as much as utility. The person buying games for the table wants the availability of a games to be higher so that people have a better chance to find that game they really want to play, so they need to know if more copies of X is better than more copies of Y.

  8. As someone who was addicted to Play to Win at Geekway, I’m curious if you took the game length from BGG and then standardized it for the number of plays, for example, ten plays of Love Letter is worth one play of Glass Road, if you would come up with a different top ten.
    (Apologies if this is actually how you did it, but I’m not 100% sure what formula you used for your top list!)

    1. I didn’t use a formula–I just used the total number of players who played each game.

      Standardizing it is one way to look at it, and it might give a better perspective for longer games…but I’m wondering if that data is relevant. The key is to get lots of people to play your game. If you have a 120-minute game that attracts 50 players (6000 total minutes played), is that really the same as a 20-minute game that attracts 300 players (also 6000 total minutes played). Perhaps there is way to compare the two, but I’m not sure it can be standardized.

      I’ll test out one formula I have in mind to see if it checks out.

      1. Oh, and for point of reference to your ROI breakdown, in a regular gaming group of five people, DURING Geekway, just from the titles available on Play to Win, we bought 7 games. This doesn’t include titles we already owned or wanted but had to prioritize to a later time, or that were sold out/not available.

        I’ll be interested to see if you come up with anything interesting weighted by play time!

        1. Morgan: That’s a very helpful point of reference! I like gamers like you. :)

          I kind of think my ROI breakdown is the most helpful formula for publishers–playing time is built into the formula in a subtle way based on the number of players, but the key is the earning potential, not how much time people spent playing it. That said, you’re free to play with the data and post something about it on your blog–after all, you were at Geekway too! :)

  9. I think the number of copies needed for saturation really depends on the game. If the game is a medium heavy euro and plays 2-4 taking 90-120 minutes, you need more copies than you would Love Letter (a 15 min game).

    Which brings up the point that number of plays isn’t a great metric. Sushi Go had 440 players (plays?) ok – it is a 15-20 minute game. In two hours, a group of three would get 18 plays in to one play of say Russian Railroads. Does that make Sushi Go “hotter”? Just a guess that Rampage, Euphoria and Lords of Waterdeep might be “hotter” than Sushi Go based on their play times (Euphoria also benefited from its event).

    And as Jamey points out – any play of the game counts towards the PtW counts, so copies in the library or other personal copies help the numbers.

    If I were trying to push my game, the number of copies you want on the PtW table really depends on length and demand. A 90 minute game always takes no less than 2 hours at the Geekway, because SOMEONE is new and needs rules explained. So if your game is a two hour game AND is currently HOT (Russian Railroads, Lords of Waterdeep, Lewis and Clark), more copies (4 vs 2) matters. If your game is longer AND hard to learn (Archipelego), 2-3 copies is enough, because people will avoid it. If your game has a specific target audience (Eldritch Horror), 2-3 copies is probably enough. If your game is shorter, more copies is better. There are always people looking for a filler game to play.

    Having said that, it seems to me that having more games is always better – Giving away $100 worth of your games (your cost) means a ton of exposure. That seems like pretty effective marketing to me.

  10. For Play and Win this year, we had a larger number of ‘purchased’ rather than donated games. With 1000 attendees, we try to have enough Play and Win titles to make sure the table isn’t completely bare, with most titles having 2 copies, and a few having three.

    Geekway will continue to push Play and Win as our signature event every year, and we hope to keep improving it!


  11. without knowing the number of each game donated, it’s kinda hard to be certain. How many extra copies can be donated before saturation? A second one would probably double your players, a third, a fourth?

    1. It’s really hard to tell. Remember, as noted above, you don’t have to play a copy of a game from the play-to-win section to be eligible to win the play-to-win game. You can also play a copy of the same game from the game library or that an attendee brought to the convention.

      The key isn’t the number of games available, a number that would be impossible to calculate. The key is the number of players who chose to play the play-to-win games over the 4-day convention.

© 2020 Stonemaier Games