Big hashpower short-term vs small hashpower long-term

I've been renting mining rigs from miningringrentals and nicehash for solo pools like ck, viabtc, brains, and pyblock, treating it like a lottery. Also, I'm working on debugging cgminer and some solomining Python scripts. From what I've noticed (I get the tech stuff like sha256 and how the poisson distribution works), renting a lot of hashpower for a short time seems to be more effective than renting a little for a longer time. For example, renting 10 terahash for a day feels better than renting 1 terahash for 10 days with the same budget. What do you guys think? Any experiences or ideas? I feel like renting a ton of hashpower is like having a big memory. You’re scanning a larger area that’s definitely bigger than the target hash. By seeing, remembering, and cutting out the bigger parts, you get closer to the target hash. So, even just keeping track of failures and not working on them brings you one step closer to that target hash, which seems key when renting big hashpower. On the flip side, when you rent lower hashpower for longer, each new block brings in more randomness. You can only scan a smaller section of the target hash range in a block. I've seen demo runs with different hashpowers over the same period, and the big hashpower hits tougher difficulties way more often, even in super short time frames. So, my final question for tweaking my gambling strategy is should I rent 300 terahash for a single day instead of 10 terahash for a whole month? Or do you think that even changing one bit of data makes the sha256 totally random, and I'm off base? I've checked the sha256 bit operation.

What's better to play at a casino: lose $300 all at once or lose $10 30 times? I don't see any difference, since the outcome is the same. It depends on how you like to play: fast or slow. I think it's very important that this doesn't become your addiction.

Bitcoin mining difficulty usually is increased, so i would recommend choosing "rent 300 th for a day" for slightly better chance.

Considering memoryless of SHA256 mining, if not the viewpoint, I see the variants as the same. The days and hash were only varied. So far the total hashes are equal, the expected value will be the same (10th for 1 day/1th for 10 day or "instead of renting 10 terahash for 30 days, should i just rent 300 th for a day").

It is certain that the larger rentals are better. At least 1 ph. Here is why: 1ph needs 4 s21xp's or 5 s21's or 7 s19xp's 100th will not gain you access to any of the gear above they are too big. 100th will like be 1 s19 or 2 s17 units.  So you very likely will be renting from a guy at 1 ip his connection could suck. 1ph you are more likely to be renting from at least 2 ip's The more ip you pull from it is likely to be at least one good connection