[Discussion] Betting on difficulty

[rex]

Some people may already know that we are working on building a platform for gambling on bitcoin’s difficulty value, I hope this post may help more people know what we are doing and why we are doing this. Generally speaking, we invented a game about gambling on the future difficulty value. (Detailed information about difficulty, click here.)

Why we want to gamble on difficulty?
Gambling by using bitcoin or on bitcoin is not a novel idea, and we guess that Satoshi Dice is known here. However, Satoshi Dice depends on pure luck. This kind of games doesn’t sound interesting for us who believe that the fun of winning a game originates from how to outmaneuver your opponents, not merely depends on luck. Therefore, we want to create a more interesting gambling game like horse racing. If you know much about the horses and the riders in the race, your chance of winning is higher than others. Luck is also important, since unpredictable things will happen in the real race.

Mining difficulty in bitcoin network will fluctuate along time like stocks’ prices. Yet unlike random number, difficulty value is determined by us, who are watching the bitcoin network. So we believe that betting and predicting on difficulty value is more interesting than guessing random number on Satoshi Dice. Therefore, why not making a game that let us bet on the difficulty? Currently we are building a platform for people to gamble on future bitcoin network difficulty value.

Details – How to play?
Difficulty value is a number we can get from every bitcoin block. Everybody can check the difficulty value. It’s open and transparent to everyone, e.g. you may check difficulty value here. Difficulty value changes every 2016 blocks, approximately every 2 weeks. So beginning with the first block which has different difficulty value from previous blocks, to the last block that still holds the same difficulty value as before, we call that those 2016 blocks form a round. Blocks in the same round share the same difficulty value. People can bet on every future round’s difficulty value. When you are betting, you should specify your round, your estimated difficulty value, and your betting amount. We will collect all the bets on the same round to form a pool, that is to say, every round will have a pool, which composes of all the money betting on that round. Noting that next round’s difficulty value is calculated by information of current round, everyone can easily deduce the exact difficulty value right before the beginning of next round. So we only allow betting on the round that will begin in at least two weeks, that is to say, you can’t bet on next round, and the available rounds for betting start from the round after next round.

When to pay out?
The payout time is in the beginning of every round. Once the difficulty is fixed for a round, we will start to pay out the money in the corresponding pool.

What’s our paying out strategy?
Generally, we follow below three principles.

• The earlier you make your bet, the bigger share you win from the pool.
• The closer your guess is to actual difficulty, the bigger share you win from the pool.
• The greater your money in your bet, the bigger share you get back from the pool.

The first rule is to award players who are willing to take risk to make a long-term guess. The more risk you take, the more gain you should get.

The second rule is to award players who have keen sense of bitcoin computing power. The smarter and luckier you are, the more gain you should get.

The third rule is to award people who are willing to take risk by putting a big fortune. The higher risk you take the more gain you should get. Also, this rule is fair for all the participants. If you are betting on a big fortune while others only put one nickel, it is unfair to lose all your bets to others who are not brave enough to take risks.

If you are interested in details of how to pay out, this link leads to payout formula we designed according to the principles mentioned above. Details on Payout Formulas

What’s our business model?
We maintain the gambling platform not for free. We will get a small percent from every pool as house fee, e.g. 3%.

Current progress
Currently we are polishing the front-end of the website, and the back-end part is already completed. Our server is built on Google App Engine, therefore, security and scalability won’t be big problems for us. The estimated date is on Feb.

Should you have any comments or questions or suggestions, please kindly let us know. Usually we will be online from 7:00am to 11:00 am EST.

[1715019849]

1715019849

[1715019849]

1715019849

[1715019849]

1715019849

[Sukrim]

What are the minimum betting amouts?
How do users communicate their exact bets to you (e.g. can they only bet on difficulties that are pre-set by you or can they bet on values they choose themselves)?
Will you be auditable via the blockchain (so it's visible which addresses you own), similar to satoshidice?
Will you publish the values of Alpha, Beta and Epsilon beforehand?
What about taking a fixed amount based on actual costs instead of percentages? This would make it safer for you to calculate and encourage people to bet higher and more often, as your cut actually gets smaller the more the platform is used.
Is "current block depth" measured by you, derived from the block in which the actual transaction get included or something else?

An idea:
Minimum bet amount: 0.001 BTC
For each 2016 blocks round you publish an address.
To bet, you take the logarithm of the difficulty you want to bet on (e.g. LOG(3 500 000) = 6.54406804) and add the 5 most significant digits to your payment. In this case you pay 0.00165440 (= betting on 3 499 451.67) or 0.00165441 (= betting on 3 500 257.54) to the address in which block you guess this difficulty. The few hundred µBTC more or less can either be taken by you as a fee or simply considered as blonging to the bet amount.
Of course you can offer calculators on your site etc. to make this easier.
You consider the block in which the transaction is included as the time the bet has taken place.

Downsides: Bets can only be done in 0.001 BTC increments and it might sound complicated ("OMG maths!") - also small margin of error and I can't bet on very exact predictions (in this example +/- ~500 difficulty on a bet on 3.5 million difficulty)

Pros: Auditable, can be verified by anyone who cares to write a script for that, you can keep your private keys for these 2016 block addresses even offline, as all you need to do is sign a single large transaction with it once. That transaction can be generated on an internet PC, signed offline on your secure station and introduced to the network. I guess that amount of work shoud be doable every 2 weeks, even with maybe low profits. Also you don't have to do any kind of fancy webforms or anything - a logarithm calculator, a few addresses and you could be running today!

[nimda]

What formula are you using? How are time, accuracy, and amount weighted?

[rex]

Q: What are the minimum betting amounts?
A: According to recent BTC to USD ratio, we set 0.01 as the minimum betting amounts. Besides, usually transaction fee is around 0.0005BTC. Thus sending amount below 0.01BTC cost high percentage of transaction fee.

Q: How do user communicate their exact bets to you?
A: We don't have pre-defined difficulty values. We encourage users to freely bet on the values they like. Our betting strategy is like this.

We will encode the betting difficulty value into the betting amount. The minimum bitcoin amount in a transaction is 0.00000001. Yet our minimum betting amount is 0.01. Therefore, we will use the digits that after hundredth to encode for users' betting difficulty. For example, if you want to spend 0.01BTC to bet on difficulty value 123,456, you can make your bet 0.01123456 to us. The extra money we received will be also added into your betting amount. So, actually, you spend 0.01123456 to bet on difficulty value 123,456 in this case.

However, since we only have 6 digits while currently the difficulty value is at least above 2 million. To tackle this situation, we will release more than one address, and every address corresponds to different levels of scaling. For instance, if you want to spend 0.01BTC to bet on 3,512,345. You may send your money 0.01351234 to our another address, which will automatically multiply your encoded difficulty value with 10. So in this case, actually you spend 0.01351234 in betting on difficulty value 3,512,340. Though you may lost a little bit accuracy, it is relatively very small comparing to the actual difficulty value.  

Q: Will you be auditable via the blockchain?
A: Yes, we will publish our own addresses on the website. So users can check every transaction we receive and send. It's open to everyone.

Q: Will you publish the values of Alpha, Beta, and Epsilon beforehand?
A: Yes, we will publish those values when the platform is online to use. Currently we are conducting some simulation. Through the simulation result, we hope that we may set those values to meet two ideas. The first is that we hope in every round, at least 10% percent people will win. The other one is that the biggest winner has possibility to win more than 150X. Besides, we will publish as many statistics as we can to help users to analyse this game. We are ready to help you to extract the data that you are interested in.

Q: Is "current block depth" measured by you, derived from the block in which the actual transaction get included or something else?
A: When the main chain is stable in the bitcoin network, we will get the transaction from main chain blocks to judge the depth.

[rex]

An idea:
Minimum bet amount: 0.001 BTC
For each 2016 blocks round you publish an address.
To bet, you take the logarithm of the difficulty you want to bet on (e.g. LOG(3 500 000) = 6.54406804) and add the 5 most significant digits to your payment. In this case you pay 0.00165440 (= betting on 3 499 451.67) or 0.00165441 (= betting on 3 500 257.54) to the address in which block you guess this difficulty. The few hundred µBTC more or less can either be taken by you as a fee or simply considered as blonging to the bet amount.
Of course you can offer calculators on your site etc. to make this easier.
You consider the block in which the transaction is included as the time the bet has taken place.

Downsides: Bets can only be done in 0.001 BTC increments and it might sound complicated ("OMG maths!") - also small margin of error and I can't bet on very exact predictions (in this example +/- ~500 difficulty on a bet on 3.5 million difficulty)

Pros: Auditable, can be verified by anyone who cares to write a script for that, you can keep your private keys for these 2016 block addresses even offline, as all you need to do is sign a single large transaction with it once. That transaction can be generated on an internet PC, signed offline on your secure station and introduced to the network. I guess that amount of work shoud be doable every 2 weeks, even with maybe low profits. Also you don't have to do any kind of fancy webforms or anything - a logarithm calculator, a few addresses and you could be running today!

we really think that we share the same concepts - using betting amount to encode the desired difficulty value. However, logarithm may have difficulty in providing enough accuracy for difficulty value.

As to the extra money, uBTC, we will add that into users' bets.

As to fixed amount of house fee, according to our simulation, we find that splitting one big bet into small different bets is a better strategy. So in this case, taking a small percentage from the pool is better for users' different betting strategies.

As to your other questions, we make a QA section in this post above. We really appreciate your insightful questions and great suggestion!

[rex]

What formula are you using? How are time, accuracy, and amount weighted?

Thank you for your question. We have a detailed formula, and this is the link to it. Detailed Formula.