This blog was first posted by FunFair founder Jez San on his Proof of Steak website on 9 July 2017. Among the aspects that make FunFair the leading blockchain gaming tech solution are speed and efficiency, enabled by our Fate Channels. In this detailed cost analysis – with numbers accurate on the dates of testing – Jez compares FunFair to several competitors in the gaming space.
Can Ethereum casino games disrupt the online gaming industry?
The cost of gaming on the blockchain
Here at FunFair we believe that games have to be both Fun & Fair…. One of the things we didn’t manage to get into those alliterative F’s was that they also had to be low cost. Frugal, Fun & Fair. Oh yea, and Fast… but by then, we’ve had too many F’s and it just didn’t have a ring to it.
We’ve been focussing our effort on the creation of a unique gaming platform together with the protocol and a suite of launch games that will bring ethereum casino gaming to the mass market with the goal to disrupt the incumbent online gaming market — currently worth some $40B/year.
We’re also aware that the land based casino gaming market is worth about ten times as much as the online equivalent and we feel that its artificially low and barely scratches the surface because of a variety of reasons that we believe are in part related to trust, regulation and monopolies which make online gaming harder to compete with its land based equivalent.
When you think about how other land based incumbent operations have been disrupted by online equivalents, whether its online shopping, uber taxicabs, ebay and so on, you’ve got to wonder why online casino gaming has had such a mild impact on the land based equivalents, especially when there are so few actual casinos, and they’re often sited quite far from where people live and its often inconvenient to make the trek to get there.
When we talk anecdotally to people we meet, especially those at land based casinos and ask them if they play online, we hear a common tale that often they just don’t trust online casinos. They think either the games are rigged and they might not get fair odds, or worse, they worry that leaving their money in the custody of the casino is a risk too great and that they might lose it or not have access to it when they want it back. Although stories of casinos cheating players are rare, these are valid concerns as some casinos have on occasion been caught stealing players funds or spending players’ funds from un-ring-fenced accounts.
The trust issue is more one of perception than reality but in players’ minds its still very real, and if we can do something to improve player trust — and confidence — and at the same time improve regulation by automating much of it, we think we can help improve the market share of online versus land based gaming. FunFair’s games CANT cheat. Everyone in the ecosystem benefits. The players’ get trustworthy reliable games that they can audit in real-time, the affiliates get paid on time and can see all their players’ spending, the operators get transparent low cost games that give them extra margin to recruit new customers, and the regulators get a level of transparency and audit-ability they’ve never seen before. Everyone wins!
We believe the advantages of bringing ethereum based smart contract gaming to the mass market are great from a player fairness and protection point of view that we think it has to happen for those reasons alone.
However there are significant headwinds in delivering blockchain gaming to the mass market. One of those headwinds is cost. In Ethereum gaming all transactions that are processed by the network are charged a ‘gas cost’ to send the transaction to the miners for processing and eventual inclusion in the blockchain. Every bet, and every win is usually a transaction, and each one incurs a gas cost.
At FunFair we have our own unique solution to this problem (Fate Channels) but first, lets quantify the problem by comparing and contrasting the existing solutions.
In a traditional incumbent online casino game, there are two or three places where the player might incur a cost. At the credit card deposit stage, during game sessions at each bet and at withdrawal time. With incumbent online gaming, usually the casino covers the cost of the customer making the credit card deposit, and even though that might be as high as a 2–5% charge on the deposit, the casino eats it and thus the player thinks there was zero cost for making that deposit. Its unlikely casinos would get much business if they charged customers to make a deposit. Lets say theyre depositing $50, that’s the only charge that will appear on their credit card without additional fees.
Similarly, if the player wants to bet $1 on a blackjack hand or roulette game, they bet their $1, and that’s all it will cost them to play that game. Their wins and losses are the usual casino odds and the house edge is just a few %.
When we look at the various blockchain alternatives… there are two types. Bitcoin gaming has been around for a few years and brings with it some privacy features (no itemised credit card statements, which lets people be more discreet about their gaming activity) but the down side is an additional player risk and also the lack of fairness and player protection — in fact, the risk to the player is far greater when they play a bitcoin casino game versus a credit card one, as the player has to trust the casino with a deposit of funds and leave it with the casino for a long period of time that one day the casino could easily run off with. To solve these fairness and player protection issues, there are also newer Ethereum gaming solutions (discussed in detail, later).
However, this starts to get quite bad when you consider the bitcoin casinos, as the cost of sending a transaction using the bitcoin network is becoming quite expensive and is borne by the player. If you factor in the two transactions, once to deposit, and once to withdraw, it can cost several dollars of bitcoin transaction fees to send and receive this cash.
However, to play the games, once your bitcoin funds are deposited, playing the games has zero additional cost as the games are played entirely off chain. While this means there’s zero ongoing fees to play each hand or each game, it also means there’s zero transparency, just like the incumbent online gaming equivalents. If the casino wanted to cheat you, they could and you’d never know whether you received the right odds. They might be able to prove they chose a random number (provably fair RNG), but they cant prove what the game did with that random number, or how the game’s outcome was decided. If you were supposed to win but didn’t, you will never know. You also can’t know the stats of all the other players, since none of the games are published unless the casino wants to publish them and they can be selective and only publish the ones they want you (or the regulator) to see.
In Ethereum gaming, we took a look at the existing solutions… and although on the most part, they are providing a fair and player protected experience (the RNG in many implementations is provably fair — often committed in advance of the games, or chosen independently by the blockchain using techniques like block-hash etc, and the gameplay is processed inside a smart contract where the game outcome decision can also be provably fair)…. However, the new problem is transaction fees — notably Gas Costs. If each transaction — each bet. each win — is done on-chain, then each bet and each win requires a payment of gas to the network. A slow and expensive way of implementing the games that is uneconomic and uncompetitive.
If blockchain casino games are supposed to disrupt the incumbent online gaming industry they can’t just be more fair — they also have to be lower cost! And so far at least, every ethereum casino game is significantly more expensive to play than the incumbent games people are already playing, which would mean there’s zero chance they will succeed in disrupting the incumbent online gaming business. Players simply wont pay extra for fairness! (and we’re talking a LOT extra)
We think that this new generation of ethereum casino games are to be applauded for their innovation and in most cases their fairness. They have definitely improved the experience of player trust and we think that all of our peers have made great strides to help the player receive a fair game. At FunFair our motto is Fun, Fast and Fair and we mentioned above that we think Frugal is important too. We think that most ethereum games often can most likely tick the Fair box, but they usually are much worse on the other F’s (Fun, Fast, Frugal etc) — and for them to compete with incumbent online games they need to have all of the above! Just being Fair isnt enough. Some of these games are just science experiments that prove the game was possible, but theyre so unplayable and un-fun we think players will never give them serious consideration (apart from in blockchain, where players are more tolerant of high gas costs and pretend they don’t really exist).
We tested all the existing Ethereum casino games and provide you some example gas cost here. We tested the dice games: vDice & Etheroll. We also tested the blackjack games: Winsome & dao.casino. And we also tested the Winsome’s Rouleth, as well as an early version of our own home-grown effort. The testing was done as follows and we urge those that would question the methodology or findings to try them out for themselves as we encourage independent testing to validate these findings. Simply go onto each of the gaming web sites and play some games.
Make note of your ethereum address so that you can look at the transaction afterwards and see how much gas was paid (both by you, and sometimes, also by the gaming site). Note the price of Ether on that day. Also, if these games are using test ether (on ropsten, kovan etc) you can still see how much gas was consumed and that translates into an actual effective cost had the games been live on the main network instead of a test network, as the gas usage is the same. Even on test ethereum networks, etherscan reports how much ether was used in gas costs, which you can multiply with the actual eth/usd price to get the transaction cost, were that game on the main network.
In most cases, the per game cost, every time you play is between $1–4, at present ethereum prices (circa $240). And these gas costs scale linearly with the price of ether… so as eth goes up to $400 (as it did, last month) so too does the price to play these games increases… making most of these games exorbitantly expensive to play, for each game or hand or dice roll.
Its no wonder that Etheroll & vDice decided to opt for a very high minimum bet size to ensure their gas cost doesn’t overwhelm the bet size. The min in Etheroll is 0.1 Eth (today its $24) and vDice was 0.22 Eth. And even then, the gas cost is over a $1 for most of those games.
Here’s the statistics, complete with actual transactions to assist in validating these findings. But better yet, try it yourself and see what your own gas costs are when playing these games. They will no doubt scale relative to the price of ether on the day of your test.
Note, in all of these tests, im only talking about the gas cost, which is in addition to any bet, and doesn’t include the cost of the House Edge!
Tests performed at ether at approx = $240 on July 9th.
The dice roll took about 2 minutes. Was a little excruciating even though I won! And it cost me $1.09 (and it cost Ethe roll $0.72 in gas that OracLize paid for during their callback). And presumably cost Etheroll some fees to pay OracLize for the service. Etheroll says their house edge is 1% — which is charged on top of the gas cost.
My transaction had a gas limit of 250,000, and an actual cost of 210,000 ether which was about $1.09 and is here:
The callback from OracLize to Etheroll to give the RNG result cost them an additional $0.72, on top of OracLize’s own fee:
The video of this actual test is on YouTube so you can see its a normal customer use-case, nothing unusual:
Now lets turn our attention to some the other main dice game to see how it fared. Next up is vDice (by vSlice)The transaction cost about $0.75 for a dice roll so it was a little cheaper than Etheroll, but less user friendly. No interactivity at all, very reminiscent of the original SatoshiDice on Bitcoin, ie, you send money to an address and a minute or so later you receive back either your winnings or a 1 wei consolation prize.
The transaction id was:
And the win/loss transaction was:
There was an earlier failed transaction (that took 30 minutes to be paid) at:
Which took 30 minutes for the result to come back:
The video recording of the test that worked is at:
Weirdly i’ve not used vDice that much but more than once i’ve had it fail on me and presumably, the 30 minute payment delay is one where their systems failed to notice the transaction and they detected a fallback condition and then re-submitted the transaction. Since this has happened to me a few times, it can’t be that unusual. I’ve had reason to email them before about lost bets (that were later processed automatically by their fallback recovery).
Next up is Winsome, who have two games — Roulette and Blackjack, entitled Rouleth and Blockjack. These are on-chain games — both utilise the block-hash method of generating random numbers which isn’t that fast (requires waiting multiple blocks) and has gas cost — especially in the case of the blackjack as its very reliant on blockchain speed and slow / tedious to play (and expensive). The roulette isn’t fast either (a minute or so per spin) but not quite as bad as the blackjack which can seriously take minutes to play each hand. Rouleth cost about $1 per spin, and took just over a minute.
The rouleth transaction is here:
Blockjack cost between $2.50 and $4 per hand to play, and took several minutes to play each hand!The blockjack 1st hand was here:
And Stand was here:
The other hand that had a low card count and needed a few extra cards cost about $4 and wasStarting hand here ($1.65):
Hit next card here ($1.40):
Hit next card here ($0.59):
And finally, Stand here ($0.66):
Total gas cost to play that hand (!) : $4.30
Both their prototype games, a Dice game and a Blackjack game are available to test on their prototype showcase (running on Ropsten). I wasn’t able to get the blackjack game to run because of a lack of Bankrollers (their server app), but I was able to look at the contract and see how much gas someone else’s hand of blackjack had consumed.The dice game was surprisingly fast, as it seems to pick the RNG and tell you if you’ve won quite quickly (a few seconds)… but the transaction takes a minute or so to appear on the blockchain. It still consumes a lot of gas. Personally, I think their Dice game is the best dice game so far.
A dice roll transaction (cost $1.63):
A blackjack start hand transaction (cost $1.85) :
A blackjack confirm hand transaction (cost $2.69):
But there were two more unexplained transactions involved in this hand:
The total cost of playing this blackjack hand was well in excess of $4.00 ! The video of this test can be found here :-
And finally, here’s a test of FunFair’s Euro Roulette game (as yet unannounced, unfinished, and still in early alpha test, but it will be available in a day or two for everyone to play in the www.funfair.wpengine.com showcase section).This game uses our (now quite infamous) Fate Channels, which are an advanced state channel technology that supports realtime random number generation, as well as micro-payments for bets & wins, and also real-time player interactions. It does other cool things as well related to audits and game provability.Comparing just one spin of the roulette wheel wouldn’t be any sport now would it, so I took the liberty of playing about 8 times in a similar time that the other games did it in one game .. hope no one minds ! But the point is that I could’ve played it 100 times and the gas cost would’ve been the same. You see, using our Fate Channel tech the player only pays gas costs for opening and closing the state channel (at the start and end of the gaming session).. and if you played 100 games in an hour, it still costs less than just one game of the other company’s ethereum games.
The ‘start game session’ transaction is here (cost $1.04):
And the ‘end game session’ transaction is here (cost $0.42):
Roulette test video here :
This tech and these games are a big leap forward in playability and economic viability and allows Ethereum gaming to be truly disruptive to the incumbent online gaming industry for the first time! And its fast and fun too, as well as ‘frugal’ (gas efficient). We see the quality of the games, and interactivity of the experience as well as the low cost deployment to all be key factors in competing to bring blockchain transparency into the online gaming industry.