Round 1: U/R Pyro Delver
Round 2: Tezzeret
Round 3: Elves
Round 4: U/R Guide Delver
Round 5: U/W/R Delver
I enjoyed playing this deck generally, although the clear lack of answers to an Insectile Aberration is a huge issue that needs to be rectified before the deck is ready for a tournament. In that context, I can see a good argument for a card like Ratchet Bomb, which blows away Insectile Aberration. Remember that flipped cards have no converted mana cost and "does not exist" counts as zero in Magic. So it's possible that Ratchet Bomb deserves consideration.
Phyrexian Metamorph was wholly unimpressive. It seems like a card that should be really good, but it's not. If you have a six-drop in play that isn't about to get killed (say, in response to a Phyrexian Metamorph), you're almost certainly winning. When you don't have something in play, Metamorph is not good. Since this a deck that wants to explode onto the board more, I would recommend cutting the Metamorphs for a card that is more proactive and that can disrupt Delver of Secrets.
I don't know what that card should be—Chalice of the Void and Trinisphere are both obvious considerations, but we definitely built our deck around not having to beat our own Chalice of the Void on one. Also, casting Mox Opal and Sensei's Divining Top in the same deck works. Adding Trinisphere makes all of that stop working. Similarly, Chalice of the Void largely invalidates a strategy that wants to gain midgame consistency via Faithless Looting, Goblin Welder, and Sensei's Divining Top. It's definitely not clear cut. It may just be the case that you want to add Ratchet Bomb to the maindeck.
The six-drops were great. While there were a few hands where we drew fully half of the top end of our curve, that's atypical, and we still could've won had we discarded Staff of Domination instead of Wurmcoil Engine to Faithless Looting. We also could've won had we not sacrificed our six-power lifelink creature against a board of six unblockable damage, so chalk that last game loss up to me, not the deck. Wurmcoil Engine and Steel Hellkite are great, and you'd be out of your mind to cut either.
I never really wanted something bigger than the six-drops. Karn Liberated is an inelegant solution to Delver of Secrets, especially given his vulnerability to Spell Pierce. Kuldotha Forgemaster may just be better than Phyrexian Metamorph, in which case the Faithless Lootings can probably hit the bricks for the rest of the Forgemaster package (which is really just a Blightsteel Colossus and four Forgemasters since we have two Lightning Greaves anyway).
The sideboard was pretty hit or miss. We never played against a spell-based combo deck, so I can't tell you if Tangle Wire is good. Phyrexian Revoker clearly earned its keep against the two non-Delver decks that I played against. Contagion Engine was a great ten-mana Marsh Casualties against Elves, but Ratchet Bomb may just be better. That's up to you.
Blood Moon was emphatically bad. It did too little, I wanted it basically never, and it comes in against decks with Spell Pierce. It's just not what the deck wants, and I would happily replace it with almost any other disruptive card. This deck should be focused on dominating the board, not on locking out someone's mana for the entire game. Even Tangle Wire, a card that pseudo Armageddons someone, is really just there to buy time until board control can be established.
There are clearly a lot of directions available for the deck. As several people pointed out on Tuesday, the Cloudpost / Glimmerpost / Vesuva package is a reasonable consideration, although it makes it a lot more likely that Metalworker ends up being Palladium Myr. Managing the artifact count in this deck is difficult, especially given how many sweet nonartifacts there are in Legacy that want to enable artifact-counting cards.
I want to make the informal Facebook likes poll a fixture of video article comment sections, but I don't have a great way to calibrate it. My goal is to provide a way to enable exclusionary audience selection on a list of choices. They won't always be "what color[s] should I play next week?" but they will be in a similar vein. My underlying philosophy is to take questions of content preference to you, the audience. From there you can give me a good idea of what you want more (and less) of.
Since I already know what I want to write about next week, this week's topic is going to be a little meta. For the next few weeks, the format of this particular feedback mechanism will evolve as I come to better understand how you want to express your preferences, but this question feels like a good starting point:
What types of choices do you want these polls to present you with?