Hearthpwn

Yes it's soo possible! Did a quick count, winning 62 times in a row with a 1% win rate is a bit less (around 10000x less) possible than getting hit by 7 meteorites and a lightning strike simultaneously. Losing (surrendering) 6138 times would take you at least 5 hours a day for a month. Winning 64 times in a row with an amazing win rate of 80% (I don't think anyone can reliably achieve such a win rate, luck is a deciding factor too many times) is around 1/50000000. And I don't think surrendering can be aplied to count an actual win rate. It all depends on how you understand these words, if you insist on understanding it as games won/games played, doesn't matter how you lost (by surrendering in this case, if you 'tried' to actually play these 6000 games, getting to legend would by no means be possible due to time limits), if someone was a real pro and devoted himself to getting legend with a 1% overall month win rate, it would be possible. If 'win rate' is the measure of how many of the games you actually try to win you win, getting legend with 1% win rate is not possible.

Ran the script again, but! I noticed I'm a moron and the first time I ran the script the numbers seemed a bit low compared to the numbers given by other people so I adjusted it for the '0 star' (1 more star for each rank, that is simply not correct). This means the math part was ok but the 'yes that is how hearthstone ladder works' part was not. Changed it now, here are the numbers, given in jpgs so that my post does not get deleted by spam bot again :P Also an example graph of a great streak at 80wr which brings us to legend and an average streak at 52wr which does not. Did the sims for 12 mins/game and 2/4 hours played a day, which gives us 300 and 600 games to achieve legend in a month.

The applications are: Do not try to get legend if you play only 2 hours a day unless you're super pro. Even with 62% win chance getting legend is not that reliable (below 90% chance).

Unfortunately priests will never be arrested, the only hearthstone police is the 0-3 police.

Ok, obviously this data won't be exactly correct because the real chance of winning changes and would be much lower for a worse player at higher ranks, but I wrote a quick Matlab script which simulated a player going for legend, his winrate doesn't change over time. I used 2 time frames - one is 7.5 minute matches (+0.5 min queue) and 24h/day playing (impossible by any human standards but theoretically possible with for example 3 ppl changing and playing a hyper aggro deck) - 5400 game deadline, and another is 9.5+0.5 min per match (real average is 12 min from what I've read) and 16h/day - 2880 game deadline. with 3-10k simulations for every winrate and time frame.

First column is winrate, second is an average of matches if legend was achieved within a month, for first time frame, third one is avg for more realistic time frame, fourth one is the chance of getting legend for first time frame and the last column is the chance of legend for the second time frame. I wouldn't say I'm a master coder and never do any mistakes, but I don't think there are any errors in the script.

80%  125    125    100%    100%
70%  185    185    100%    100%
65%  240    240    100%    100%
60%  334    334    100%    100%
55%  552    552    100%    100%
53%  745    745    100%    100%
52%  901    894    100%    99.9%
51%  1177  1135  100%    98.8%
50%  1568  1389  99.9%   92.3%
49%  2172  1646  95.2%   74.0%
48%  2729  1831  75.9%   42.3%
47%  3101  1930  40.8%   17.2%
46%  3291  1981  12.3%   4.8%
45%  3509  2100   2.1%    0.7%

What does it mean? Even if your winrate is 52% and you play 16h a day for a month, the chance of getting legend is 'only' 99.9% ^^ And even with a winrate as low as 45% there is as much as 2.1% chance of getting legend if you play 24/7. The reason for the second column being lower than the first one is because with lower amount of games before deadline, the simulations that actually get you to legend are much faster than those in the first time frame.

'Kill a minion that has a type' (mech, pirate, beast, dragon...) would be pretty neat but a bit broken I guess.

Aggresive sorcerer super powerful, lurker very meh, armored warrior should give the debuff to your side only, it's a mistake in wording I guess? Reinforced Murlocs would be very nice if it reappeared in your hand at the end of the turn ^^ Like Headcrack but without combo.

Is it called 'Kolento synergy' because it has synergy of his level?

That is not right. The sequence does matter and both D-D-H and D-D-D (which can not be achieved, so after D-D it's always D-D-H, but the chance of it happening is not 1/7 but 2/8) kill the dragon. The chance of killing it is exactly 50%.