My main advice (to everyone, and to
@Deamon in particular) is to pick fewer games, and probably also more upsets.
e.g., Take a QB-WR stack in the divisional round instead of two different QBs. Instead of taking Flacco+Goff, go with either Flacco+Cooper or Goff+ARSB, so that you're only counting on either CLE or DET to win rather than both. That way you're picking one less game.
There are 4 main things to aim for with your entry:
1. A relatively high chance of hitting on your bracket so that you get no (or almost no) zeros.
2. Relatively few other entries also hit on their bracket, in the cases where you do hit on yours.
3. Players that are expected to score lots of points, if you hit on your bracket & they actually play that week.
4. Some distinctiveness in players on your roster, compared to the other teams whose bracket hits when yours hits, especially if there are lots of those other teams
Skipping games helps a lot with point 1, and some with 2 & 4. The main advantage is that it's one less game where you can bust (e.g. your bracket stays intact no matter who wins DET-LAR if you skip that game). This can also help you face fewer intact brackets when you do hit on yours because an upset that spoils a lot of brackets won't mess up yours, and you'll get some player distinctiveness because you have to take players on other teams while the rosters that made the correct call on the game you skipped will have players on the winning team (probably mostly their QB).
Skipping games hurts you some on point 3, because you have a smaller player pool to pick from. But with reduced QB scoring this year, the first couple games that you skip hurt your expected scoring very little, because the top non-QBs score about as much as a typical QB. e.g., By ppg, ARSB actually outscored Flacco by half a point per game this year, and Amari Cooper was only 2 ppg behind Goff. Also, players tend to score more fantasy points when their team wins, and skipping some games can let you pick more players in games that their team wins rather than in games that you're counting on them to lose.
If you pick the favorite in all 12 games you'll have about a 1/200 chance of an intact bracket (that is, 0.5% of the time all 12 favorites will win). But if that happens then you'll be facing off against a bunch of other teams who also have intact brackets, because the favorites are very popular picks. If you skip multiple games you can get your chances of an intact bracket to better than 1/100. If you skip some games and pick some underdogs, then you can still have better than a 1/200 chance of an intact bracket, while substantially reducing the number of teams that will also have an intact bracket when you do.
How I'm reacting to your comments....let's say you and I correctly pick the SB contestants and are mostly a wash in the last two rounds....there are two schools of thought: 1. I should rely on the power of correlation and pick stacks in the early rounds, or 2. I should rely on the relative stability of scoring from the QB position and maximize QB count in the early rounds. I'm not sure which is the more sound approach but it occurs to me if you are relying on picking the correct early round stacks you have just decreased your overall probability of picking correctly for the duration of the rounds. I suppose there are enough entries to where someone will pursue your approach and NAIL the early round stacks, but you just made it harder to do, dammit!
Oh and I suppose there is a category three, which is to generally maximize early qb count, but then to have specific exceptions. Tyreek Hill comes to mind. Unless its 40 degrees below zero.
Let's say that we have similar rosters, almost entirely picking the same teams, with some differences in players. Except, in the divisional round you have Goff & Flacco, while I have Flacco & Amari Cooper.
If DET & CLE both win in the wild card round, then you probably have a very slight edge on me in the divisional round. Maybe 2 points higher projections, let's say a 55% chance that you outscore me. Though I probably have a higher upside, because Flacco & Cooper both having a huge game in the same week. So this scenario is maybe a wash, maybe a slight edge for one of us, I don't know. It's not the main reason in favor of stacking.
If CLE wins and DET loses, then I have a significant edge on you in the divisional round. You are taking a zero and I am not; I have Amari Cooper to your Goff zero. This is the great thing about concentrating on some teams and skipping over others - it increases my chances of making it through the whole contest unscathed by any zeros. This is why my main piece of advice is to pick fewer games.
If DET wins and CLE loses, then you have a significant edge on me but it doesn't matter much. You are taking a zero (from Flacco) and I am taking 2 zeros (from both Flacco & Cooper). But we're both are in fairly rough shape for winning the whole contest, even if we hit on the rest of our bracket, because probably there are a bunch of teams out there who didn't take any zeros.
If DET & CLE both lose then we are both in pretty bad shape, taking two zeros each, and very unlikely to win the contest.
As I recall, last year the winning team had some zeros in round 2 and I know I had a winning route in my bracket of 6 with zeros in week 2. I think where your thought might come into play is if Purdy/Cmac/Deebo are within a few points of Purdy/CMac/Aiyuk in the final round.
By the way, very good read.
I do have one question, how does a Flacco/Cooper in round one help you in round two if you can't use them? Seems like you would want someone from Houston like Stroud and save Cooper for round two. If you really want to pick fewer games then you want some of your SB team player going in round 3. This will push some of your round 3 picks into round 2. Is it possible to come up with a decent submission from just four teams across rounds 2-4 and use the best players from the remaining teams in round one? If you were come up short on good picks from 4 teams, you could sprinkle in players from a big favorite like Buffalo and Dallas in round 2 if you didn't have them in your final four teams.
Right, the idea is to take Flacco+Cooper in round 2, not in round 1. And to not take any DET or LAR players in round 2 or later, so that the result of the DET-LAR wild card game doesn't affect you.
My guess is that every entry is best off skipping at least 2 games (so picking at most 10 games).
How far can you push it? I think it could be viable to pick only 5 games and skip 7. You have 1 Super Bowl team, with 6 players from that team. Then one conference championship team from the other conference, with 4-6 players from that team and 0-2 from your Super Bowl team. Then 2 other teams in the divisional round, probably with 3 players from each. Something like this:
Super Bowl (6 SF): Purdy, CMC, Aiyuk, Deebo, Kittle, Niners D
Conference Championship (5 KC, 1 SF): Mahomes, Kelce, Rice, Pacheco, Chiefs D, Moody
Divisional (3 LAR, 3 BAL): Stafford, Nacua, Kyren Williams, Lamar, Likely, Flowers
Wild Card: whoever
So you're predicting a LAR wild card win, KC wild card & divisional wins, and SF Divisional & conf championship wins. There's about a 1/15 chance of those 5 games all going your way, and you could get as high as a 1/5 chance if you took all the biggest favorites.
In order to win, you also need some other things to go your way. You're hoping for there to be lots of upsets (which will thin your competition a lot) and one-sided games (such as a blowout Super Bowl win for San Francisco). e.g., If it's a SF-HOU Super Bowl then this roster might be in pretty great shape against a vastly narrowed field.
One worry is that you're digging a little deep on some of these players (e.g. two defenses & a kicker over the final 2 rounds). And if KC puts up big numbers in the conference championship game then probably they're in the Super Bowl, so why not just take some of their players then? Although the conference championship and Super Bowl multipliers aren't that different, so if Mahomes has a huge conference championship game and a mediocre Super Bowl game then it could be good that you're getting 3x his conference championship points instead of 5x his Super Bowl points. And if KC loses a conference championship shootout to Houston then you're golden.
So this isn't ideal in every way, but it does seem like there are plausible scenarios where it could come together well for this kind of roster. I'm not planning to push it quite this far, but this probably has a better shot at winning than a roster that picks all 12 games and is starting with a 1/200 chance of coming through unscathed rather than a 1/15 chance.
)
