Resolution:

The Block number 869688 will resolve this coin flip, this block should be mined on November 10.

Nonces over 2,147,483,647 are HEADS (Resolves as YES), nonces under this number including 0 count as TAILS (Resolves as NO). You can find the resolution of this coin flip on your favorite blockchain explorer like Blockchain.com or Blockstream.info

For convenience I'm also tracking multiple coin flips, recent results, a leaderboard of the biggest farmers, and more here:
https://docs.google.com/spreadsheets/d/11BAYqjXPjZ9szGV2L7JV-sCWC2QWl0muy1AH1vRRC7Q/edit?usp=sharing

What is a Profitable Bitcoin Coinflip?:

I came up with a way of using Bitcoin blocks as a source of randomness. This should be a superior source of randomness than the @FairlyRandom since the cost and difficulty of attacking and coercing Bitcoin block producers is fairly high, specially when doing so only for play money rewards: Mana (Ṁ). (More on the source of randomness at the end).

What makes this coin flip special unlike most coin flips and most markets on Manifold, is that all participants end up with more rewards than they started with over time, meaning that this is a positive-sum game, unlike the usual zero-sum nature of gambling and prediction markets.

How you make Mana Ṁ:

All traders regardless of their position or number of trades will receive Ṁ5 upon resolution of the coin flip. This means that you could bet Ṁ1, and make Ṁ5 times your bet even if you didn't predict the coin flip, and if you call the coin flip, you could make Ṁ6 in profit upon resolution or +600% on a Ṁ1 bet.

How to make more than Ṁ5 a day from this coin flip?

Participate in as many Bitcoin coin flip mirrors as you can. If you participate in 10 coin flips a day, you would be able to farm Ṁ50 a day, or Ṁ350. This is a way of passively building up your mana balance without a lot of risk. You could also try your luck and bet higher amounts or even trying to break the source of randomness.

More Details on the source of randomness:
Bitcoin's blockchain as a publicly verifiable and objective source of randomness

Inside every Bitcoin block that is mined there is a portion of data that makes up the block called a "nonce", this variable was specifically designed as a dedicated buffer for trying random numbers during the Bitcoin mining process. Miners repeatedly hash the block header, a combination of block data (bitcoin transactions happening globally) and a nonce (a random non-significant number), this hashing is done in an attempt to produce a hash value that is less than or equal to a predefined target also known as the network's difficulty. The network’s "difficulty level" adjusts approximately every two weeks to ensure that blocks are mined, on average, every 10 minutes.

Originally the entropy required to mine Bitcoins was fairly low, therefore this buffer is only made up of 32 bits or 4,294,967,297 different potential combinations. Currently Bitcoin uses more sources of randomness that far exceed 32 bits, however, the nonce is still widely used in the Bitcoin mining process. There are some scientific papers that have theorized that the nonce shouldn't be used as a randomness source since it can be targeted and determined by a powerful miner willing to expend the resources of increasing the difficulty of mining bitcoin by additionally requiring a certain nonce to be included in a block. While this is theoretically possible, it it very expensive to do and generally not worth it when big miners with enough equipment to target a specific block in time would most likely be far more interested in just producing as many blocks as possible in order to get the Bitcoin reward which is 3.125 BTC in 2024, or around $240,000 every 10 minutes, and this reward can be taken by any miner, so to reliably pick nonces, the attacker would need to control most of the Bitcoin network's mining power, and even then it would fail sometimes.

Later I might improve the randomness source to make it even more expensive to alter by using more provably random data from the mining process.

Some results:

From a small test sample of 1,005 recent coin flips that happened live on the Bitcoin blockchain during October there were:

498 heads (49.55% of all flips)

507 tails (50.45% of all flips)

Source: https://docs.google.com/spreadsheets/d/11BAYqjXPjZ9szGV2L7JV-sCWC2QWl0muy1AH1vRRC7Q/edit?usp=sharing

Coin flip mirrors:

All mirrors target different blocks, which means that they will have independent results, on average if you bet on all of them, you will win 5, and lose 5 flips.

You can find more Profitable Bitcoin Coin flips here:
https://docs.google.com/spreadsheets/d/11BAYqjXPjZ9szGV2L7JV-sCWC2QWl0muy1AH1vRRC7Q/edit?usp=sharing

Have fun!