Uploaded February 2020 | Updated September 2026, 1 hour ago
This chapter is more complicated than it looks, because it is the first time when enemy stat variation can start affecting reliability. Furthermore, the core strategy has numerous fallbacks but many of them are conditional based on other values, making it difficult to produce a truly comprehensive figure without simulating the scenarios.
The stat check on a Brigand on T2 is to see if Wil needs to help chip it to enable a clean kill on T3. 8%~ of the time the Brigand will have rolled up on HP and Defence meaning Sain + Lyn isn't enough to kill. If that Brigand doesn't need to be chipped, Wil attacks the level 1 Brigand in front of Lyn instead. Lyn doesn't need that chip to kill but the marginal chance Wil crits and kills it reduces her chance of death a bit (in fact, she can no longer die if that Brigand is dead)
If one of the level 2 Brigands that Sain fights on EP2 rolls high on Str (75% chance total), then he is hit twice, his chance of death versus the boss will be too high to be worth taking, as it will on average add more turns than just taking an extra turn to be safe will. Furthermore, there are scenarios that can occur where Lyn, Kent or Wil are unable to cover for too many misses on PP3, leaving at least one Bandit alive, who can potentially increase Sain's chance of death if his HP is not high enough, to the point taking an extra turn is a safer option. And finally, if Lyn dodges too many attacks and Sain is hit twice, the Archer will be in the incorrect position because he will choose to attack Sain instead of Lyn, forcing a 4 turn.
An easy way to work out if it's worth taking a risk of resetting without performing all the possible calculations is to use the expected turncount formula for the turn you are on and the chance of a reset, then see how it compares to the amount of extra turns you would have to spend if you did not take the reset chance. In this case, when Sain's chance of death increased to 30% on Turn 3 conditionally, you end up adding 1.3 turns on average, wheras you can reliably get a 4 turn by just waiting for a cost of only 1 turn. When it is only 20%~ conditionally then you add about .75 of a turn on average, so less than 1 turn, making it worth it.
Ultimately, the combined chances of a 3 turn are still rather good though. The only chances of death worth taking are Lyn's on EP2 and conditionally Sain's on EP3.
64% of the time, Lyn has enough strength to be dropped on the Forest on T1, which clears the whole way. The expected turncount for that strategy is 3.4~ but the difference is still negligible to this one, and the exp is better on Sain. For this map overall, you can't go any faster than 3 turns, and 4 turns would make the clear essentially 100% by removing all chances of death.
The rough calculations for the expected turncount are
63%~ chance for a 3 turn
23%~ chance for a 4 turn
14%~ chance to reset (5%~ turn 2 Lyn, 9%~ turn 3 Sain)
The conditional mean clear turn is 3.3~
The conditional mean reset turn is 2.6~
The reset penalty is 0.4~
The expected turncount is 3.7~
Pessimistic Mode is a hack by Doesnt on feuniverse that only allows player character stats to increase if there is a 90% chance of them having increased by the level you have just attained. It is close to 0% growths in reliability but brings stat growths into the equation to reward training units and to help compare units better than by using averages.
feuniverse.us/t/pessimistic-average-growths-patches/5134
This chapter is more complicated than it looks, because it is the first time when enemy stat variation can start affecting reliability. Furthermore, the core strategy has numerous fallbacks but many of them are conditional based on other values, making it difficult to produce a truly comprehensive figure without simulating the scenarios.
The stat check on a Brigand on T2 is to see if Wil needs to help chip it to enable a clean kill on T3. 8%~ of the time the Brigand will have rolled up on HP and Defence meaning Sain + Lyn isn't enough to kill. If that Brigand doesn't need to be chipped, Wil attacks the level 1 Brigand in front of Lyn instead. Lyn doesn't need that chip to kill but the marginal chance Wil crits and kills it reduces her chance of death a bit (in fact, she can no longer die if that Brigand is dead)
If one of the level 2 Brigands that Sain fights on EP2 rolls high on Str (75% chance total), then he is hit twice, his chance of death versus the boss will be too high to be worth taking, as it will on average add more turns than just taking an extra turn to be safe will. Furthermore, there are scenarios that can occur where Lyn, Kent or Wil are unable to cover for too many misses on PP3, leaving at least one Bandit alive, who can potentially increase Sain's chance of death if his HP is not high enough, to the point taking an extra turn is a safer option. And finally, if Lyn dodges too many attacks and Sain is hit twice, the Archer will be in the incorrect position because he will choose to attack Sain instead of Lyn, forcing a 4 turn.
An easy way to work out if it's worth taking a risk of resetting without performing all the possible calculations is to use the expected turncount formula for the turn you are on and the chance of a reset, then see how it compares to the amount of extra turns you would have to spend if you did not take the reset chance. In this case, when Sain's chance of death increased to 30% on Turn 3 conditionally, you end up adding 1.3 turns on average, wheras you can reliably get a 4 turn by just waiting for a cost of only 1 turn. When it is only 20%~ conditionally then you add about .75 of a turn on average, so less than 1 turn, making it worth it.
Ultimately, the combined chances of a 3 turn are still rather good though. The only chances of death worth taking are Lyn's on EP2 and conditionally Sain's on EP3.
64% of the time, Lyn has enough strength to be dropped on the Forest on T1, which clears the whole way. The expected turncount for that strategy is 3.4~ but the difference is still negligible to this one, and the exp is better on Sain. For this map overall, you can't go any faster than 3 turns, and 4 turns would make the clear essentially 100% by removing all chances of death.
The rough calculations for the expected turncount are
63%~ chance for a 3 turn
23%~ chance for a 4 turn
14%~ chance to reset (5%~ turn 2 Lyn, 9%~ turn 3 Sain)
The conditional mean clear turn is 3.3~
The conditional mean reset turn is 2.6~
The reset penalty is 0.4~
The expected turncount is 3.7~
Pessimistic Mode is a hack by Doesnt on feuniverse that only allows player character stats to increase if there is a 90% chance of them having increased by the level you have just attained. It is close to 0% growths in reliability but brings stat growths into the equation to reward training units and to help compare units better than by using averages.
feuniverse.us/t/pessimistic-average-growths-patches/5134
