If I was mining in a pool with a different share system, this luck would have simply not been rewarded, at least not to that extent.
That's false, or more accurately, it's just a wild assumption on your end, obviously, you did mention that you don't understand how TIDES work, which means you can't prove that your earnings were not going to be higher had you mined to another non PPS pool, if you provide all the numbers, I would be able to calculate things for you, but regardless, you just got lucky, your pool got lucky -- it means nothing more than that.
But really at least the people at ocean could provide a more comprehensive breakdown of how TIDES benefits miners.
TIDES is just another fancy word for a modified PPLNS where the window is shares based instead of time based, any benefits it adds to miner A, it would contribute to a disadvantage to miner B, if you want to understand how TIDES work you need to understand how PPLNS work.
In a very simple way to explain it is that that pool would set a time window N (which is what N in PPLNS stands for), and let it be 1 hour, if the pool finds a block at 11AM it would count all the work shares between 10AM and 11AM, and then calculate the % of each miner, so assuming there are only 2 miners, miner A sent 500 shares, and miner B sent 500 shares, the block subsidy + rewards was 10 BTC, the pool will give miner A 5 btc - fees and the same to miner B. simple right?
Now with TIDES, they don't have that N, but rather a pool of shares that is capped by the current difficulty x 8 , this means if the current difficulty is 100T, then all new shares will go into that share pool without older shares having to leave until the maximum number of shares is reached, so once the number hits 800T, every new share will wipe out the oldest share, so it's "last share in - first share out", when a block is found, the same PPLNS principles are going to be applied in distributing shares.
Does this make miners earn more? Nop, it all depends on what happens during the time a pool finds a block.
Let's take a pool with 1 hour N, say there are 10 miners each sending 100 shares per hour, and the pool is small they can hit a block every day, so for the first day, only the 10 of them are there, block found, each get 10%, the next day, 1 hour before they hit the block, a large miner joins and adds 2300 shares in 1 hour, this means, the new miner would earn 50% of the rewards and the remining 10 old miners would share the 50% which you might want to label as "unfair" because the new miner was not there for the 23 hours, obviously, if you want to cherry pick a scenario like that and say PPLNS is not fair -- that's doable, but it makes no sense.
If the same pool used TIDES, the math is not going to be a lot different, assuming the pool hit a block every day, same miners, same shares, so at hour 23 of day X, the share pool has 2300 shares (we would assume those 2300 = than current diff x 8 ), what's going to happen here? he would be adding 2300 shares every hour, which means in 1 hour he would wipe out the 2300 worth of shares for the 10 miners, but given that they will also submit new shares, so once that hour is finished the new miner would have 2300 - 1000 shares which is 1300 shares and that would get him 56.52% of the block rewards.
In the above "valid" example TIDES favored the new miners at the expense of the old miners, it paid them 6.5% less, but obviously, depending on the pool hashrate, the timing of the new miner, the overall luck of the pool, every block is going to be different, if you point the same hashrate to two pools, one with TIDES and another with PPLNS, you are going to see days when TIDES pay you more, and other days when PPLNS pay you more.
In fact, the same logic applies to different PPLNS pools with different N value, it's all random, there is no such thing a new payout scheme that would pay out more to miners, such theory doesn't exist, unless the pool finds a way to play with consensus and manages to increase block subsidy, or reduce their fees, how would a different payout pay you more? where is the
BTC going to come from?
@Kano, your math checks out, here is a simple python code for those interested in running their own numbers, you can plot the % straight away so you don't get confused by division.
from scipy.stats import erlang
# Define the mean (expected number events)
mean = 12
# Define the percentage change you want to evaluate
percent = 80
# Calculate the threshold (don't change)
threshold = (100/percent )
# Calculate the CDF for Erlang distribution (don't change)
cdf = erlang.cdf(threshold * mean, mean, scale=1)
# Format the output to 2 dec points (change '2' to whatever decimal places you want)
formatted_result = "{:.2f}%".format((1 - cdf) * 100)
print("Probability of having {}% or more events in {} blocks is: {}".format(percent, mean, formatted_result))
50% for 12 blocks is 0.25%
80% for 12 blocks is 18.48%
50% for 36 blocks is 0.00% (given that I use :.2f.format in the code, so it's only 2 decimals)
80% for 36 blocks is 7.42%
Obviously, pool size is very important here, I also think or at least would like to think that most miners understand that variance goes both ways and it's only the fact that they don't want any variance is the reason why they use PPS pools, not with the assumption that PPLNS pools have variance only to the downside.
