Tweak inclusion delay rewards and set BASE_REWARD_FACTOR

[JustinDrake]

Substantive changes:

1. Split the inclusion delay reward between attester and proposer to add up to at most one base reward. This is analogous to the reward logic in slash_validator, and makes the BASE_REWARDS_PER_EPOCH constant include proposer rewards.
2. Double BASE_REWARD_FACTOR to 2^6 (addressing item 4 in Phase 0 state transition checklist—take 5 #1054). When the total effective balance is 2^17 ETH then maximum annual issuance is a bit below 2^21 ETH. Maximum annual issuance happens when a) all validators make perfect attestations (matching source, target, head, as well as consistent crosslink data), b) all attestations are included as fast as possible (in particular, no skip blocks), and c) there are no slashings.

BASE_REWARD_FACTOR = 2**6
SLOTS_PER_EPOCH = 2**6
SECONDS_PER_SLOT = 6
BASE_REWARDS_PER_EPOCH = 5
GWEI_PER_ETH = 10**9
MAX_TOTAL_EFFECTIVE_BALANCE = 2**27 * GWEI_PER_ETH
TARGET_MAX_ISSUANCE = 2**21 * GWEI_PER_ETH

def integer_squareroot(n: int) -> int:
    """
    The largest integer ``x`` such that ``x**2`` is less than or equal to ``n``.
    """
    assert n >= 0
    x = n
    y = (x + 1) // 2
    while y < x:
        x = y
        y = (x + n // x) // 2
    return x

MAX_REWARDS_PER_EPOCH = MAX_TOTAL_EFFECTIVE_BALANCE * BASE_REWARD_FACTOR // integer_squareroot(MAX_TOTAL_EFFECTIVE_BALANCE) // BASE_REWARDS_PER_EPOCH * BASE_REWARDS_PER_EPOCH
EPOCHS_PER_YEAR = 365.25*24*60*60 / (SECONDS_PER_SLOT * SLOTS_PER_EPOCH)
MAX_REWARDS_PER_YEAR = EPOCHS_PER_YEAR * MAX_REWARDS_PER_EPOCH

print(MAX_REWARDS_PER_YEAR / TARGET_MAX_ISSUANCE)

[djrtwo]

This seems reasonable... The counter argument is it's just as simple to bound issuance with as BASE_REWARDS_PER_EPOCH + 1 / PROPOSER_REWARD_QUOTIENT, and we want to insure that we incentivize validators to not delay in creating their attestation.

Also need to remember there will be participation rewards on shard chains.

This also might not be the final adjustment for [2] until we lock down slot time

[JustinDrake]

This also might not be the final adjustment for [2] until we lock down slot time

I suggest keeping the slot time at 6 seconds for the spec freeze to "stress test" the cross-client testnet. So it could be the last adjustment before freeze (though possibly not before finalization).