As expected, the malicious stakeholder gets about 51% (0.52019) in the scenario of a 3 of 5 requirement, but in a 1 of 5 requirement, the stakeholder now has 97% control of the blockchain. It get worse quickly with increasing malicious stakeholder wealth with the unanimous system; at 30% we see 83% control of the chain, and at 40% we see 92% control of the chain.
I don't disagree with your calculations. In fact, they accomplish just what I want. We don't want the same things (you prefer work; I prefer stake; you think energy use is okay; I think it is a big waste).
Edit: However, I thought about it a bit more and decided it is not worth making a fuss over. Here are approximate percentages on attack resistance (this is for permanent 51% attacks, not lucky streaks from minority attackers; majority voting performs better against lucky streaks).
4 out of 7 majority voting: (99.95% work 5% stake; 99.5% work 10% stake ; 94.4% work 20% stake)
3 out of 5 majority voting: (99.90% work 5% stake; 99.1% work 10% stake ; 94.1% work 20% stake)
3 out of 3 unanimous voting: (99.985% work 5 % stake; 99.86% work 10% stake; 98.46% work 20% stake)
5 out of 5 unanimous voting: (99.99996% work 5% stake; 99.998% work 10% stake; 99.9% work 20% stake)
3 out of 5 majority voting is adequate when paired with a nonnegligible PoW block reward. I was fanatic for PoW protection because I wanted minimal energy use. I forgot about your big PoW rewards.
There is another issue I forgot, the waiting issue. To me it is no big deal, but I expect a huge fuss from others. You should remove this weakness to avoid future problems.
You can always hoard coin-age by waiting (see numerous rants by killerstorm). To attack, you need 5% of stake-days, not 5% of stake. (e.g. so if you have x% of stake and bide your time for y voting cycles, then you can attack just like someone with xy% stake)
How about you buy most tickets with straight coins rather than coin-time? Coins are escrowed up until the tickets win or get invalidated. This approach sidesteps the waiting issue.
Sadly, it also excludes small holders from direct participation, though they could still use lottery pools, banks, etc. You have what ~ 60000 tickets. That is 160 coins a ticket if there are 10 million coins. A bit steep. To let small holders participate directly, perhaps allocate 10% of tickets based on coin-age and 90% based on straight coins. That would allow for some token grassroots participation. Honestly there will be a bulky blockchain to store. Smallholders would end up PoS mining via a banking system anyway. No point in designing a system that allows for infeasible use cases.
Question: When you purchase a ticket, could you then send the ticket key to a pool to manage for you (in exchange for a cut)? Or does transferring your ticket impose some risk? In my opinion, it would be good to attach some risk to letting other people mine for you (e.g. allow the ticket holder to steal the winnings). It would be bad to attack a huge risk to letting other people mine for you, (the principal used to purchase the ticket should be safe from theft). This strikes a balance between discouraging centralization and encouraging participation in PoS mining.