MTG Salvation

The card is good against:

• Gx Tron
• Eldrazi Tron
• Ad Nauseum
• Storm
• Snapcaster
• Bloodbraid Elf
• Living End
• Hollow One decks
• Burn (if we get it fast enough)
• Bogles (although we're already pretty good here)
• Mardu Pyromancer
• Eggs
• Traverse Shadow decks
• Zoo
• Elves
• Amulet decks

...And I'm probably forgetting a few. It has a small, and seemingly insignificant, impact on us.

Quick question: how do you beat a game 1 opponent's Leyline of Sanctity?

I play 3 Shredder, 2 Bell, 1 Pyxis as my mill rocks. I typically have no problems with Leyline.

Quote from Pistallion »

Quick question: how do you beat a game 1 opponent's Leyline of Sanctity?

The only deck that main-decks (to my knowledge) Leyline of Sanctity is boggles.
... and that deck can literally not beat ensnaring bridge game 1. You will win with academy ruins.

Otherwise you beat it with Pyxis. Most lists play at least two Pyxis main.

Is Surgical Extraction used anymore? I haven't been using it and most lists I've seen doesnt, but I feel that it can be useful against decks that has a lot of problem cards, specifically RG decks with Tireless Tracker. I also have been having problems vs Through the Breach decks as well but idk if its just me or if its a bad mu

Quote from Pistallion »

Is Surgical Extraction used anymore? I haven't been using it and most lists I've seen doesnt, but I feel that it can be useful against decks that has a lot of problem cards, specifically RG decks with Tireless Tracker. I also have been having problems vs Through the Breach decks as well but idk if its just me or if its a bad mu

Sam Black did a tournament this weekend, and he was running Surgical. He posted it to the Lantern Facebook group. It can definitely has it's use still, you just need the correct meta for it.

Quote from Skitzafreak »

Quote from Pistallion »

Is Surgical Extraction used anymore? I haven't been using it and most lists I've seen doesnt, but I feel that it can be useful against decks that has a lot of problem cards, specifically RG decks with Tireless Tracker. I also have been having problems vs Through the Breach decks as well but idk if its just me or if its a bad mu

Sam Black did a tournament this weekend, and he was running Surgical. He posted it to the Lantern Facebook group. It can definitely has it's use still, you just need the correct meta for it.

know how he ended up?

Sam was second, although he conceded the finals. The list is here http://nerdragegaming.wpengine.com/modern-nrg-ct-mar-2018/

Quote from nogeist »

Sam was second, although he conceded the finals. The list is here http://nerdragegaming.wpengine.com/modern-nrg-ct-mar-2018/

Surprised to see Sam at a local tournament.
Was originally supposed to go to one but had to apartment look instead : /

So I just went through and started working a bit more on the spreadsheet, after figuring out some more stuff working on my Ux Tron version, and found an error in the function checking for when a Mox Opal is online immediately. After correcting the function, I found that my numbers were a bit off. Here's what I've got right now:

We've won 1125 of 1627 games where Opal wasn't online immediately (69.15%).
We've won 40 of 54 games where Opal was online immediately (74.07%).

This seems to imply that having Opal online immediately, turn one, does correlate with about a 5% increase in win percentages. The previous erroneous function showed it at a decreased correlating effect.

However, Mox Opal still correlated with a drop in win percentage - from a 70.13% (803 of 1145 games) win rate without one in the opener to 68.64% (324/472 games) with one to 59.38% (38/64) with two. So this made me wonder how this could be possible. If having an Opal online increased the win percentage in a deck full of 1-drop artifacts, how could having an Opal in the opener correlate with lower win percentage?

So what I did was make two additional data points.

The first data point I set was to check and compare games where Opal was a dead card at least until turn two (Opal in hand but not online on turn one). This shows a slight drop in win percentage, from 69.82% (879/1259 games) to 67.77% (286/422 games). That's not a heck of a drop, though.

The second data point I set was to check and compare games where Opal isn't online on turn one, but should come online turn two (I didn't check to see if an opponent somehow prevented this with a discard spell, Chalice of the Void, etc.). For this data point, the win rate went up from 69.02% (1132/1640) to 80.49% (33/41). That's quite a jump, so I'm guessing that there weren't many of the 41 game sample size in which an opponent prevented the Opal from being online on turn two. I had it specifically ignore hands which were already counted for in the "Opal online on turn one" data point, so this is specifically only when it was drawn, not online turn one, but able to be turned on on turn two.

So then I wondered, what if we compared the total number of games where Opal was online on either turn one or two? So I set it up to check that, and got 68.85% (1092/1586) when we didn't have an Opal online at all by turn two and an increase to 76.84% (73/95) when we had an Opal online by turn two, as soon as turn one, combined. EDIT: Fixed.

So I recognize that we need to be careful how we interpret the data, but it's very interesting to me. It appears that overall, having an Opal in the opener correlates to a drop in win percentage over a larger sample size. However, in the relatively small sample size in which Opal is turned on in the first two turns, the win rate does increase a decent amount. I suppose that brings us to a good discussion.

First, what's everyone's opinion on the reliability of that small sample size? I don't want to dismiss it entirely, as it does seem to show a trend, but I'm up for hearing everyone's reasoning and opinions on it.

Second, if the small sample size is accurate, we see that there is a generally negative effect of having an Opal in the opening hand, as the instances where we do have it online by turn two is relatively rare (5.6% of games total). But in those 5.6% of games, our win rate increases by 8%.

So, again, I'm interested in hearing everyone's thoughts on this. I understand that typical stock lists seen on MTGGoldfish run four, and prominent recognized players default to four, but I'm looking for independent and rational discussion concerning the evidence, rather than any appeal to authority or bandwagon effect. I'm interested in what each of you, as individuals, think about this development and why.

If anyone sees any more errors in the functions on the spreadsheet, please let me know! I just happened across this one, but having a few other sets of eyes might help.

Thanks!

EDIT: Link to sheet

EDIT 2: I think there's another error in the functions which count up the Turn 1, Turn 2, and Turn 1 or 2 Opal. I'm pretty sure there should be 95 games total (54 turn 1, 41 turn 2), but it's counting it up as 85 instead. Anyone see what might be wrong in the functions? I'm not seeing it (maybe I'm just tired). Fixed it.

Quote from thnkr »

So I just went through and started working a bit more on the spreadsheet, after figuring out some more stuff working on my Ux Tron version, and found an error in the function checking for when a Mox Opal is online immediately. After correcting the function, I found that my numbers were a bit off. Here's what I've got right now:

We've won 1125 of 1627 games where Opal wasn't online immediately (69.15%).
We've won 40 of 54 games where Opal was online immediately (74.07%).

This seems to imply that having Opal online immediately, turn one, does correlate with about a 5% increase in win percentages. The previous erroneous function showed it at a decreased correlating effect.

However, Mox Opal still correlated with a drop in win percentage - from a 70.13% (803 of 1145 games) win rate without one in the opener to 68.64% (324/472 games) with one to 59.38% (38/64) with two. So this made me wonder how this could be possible. If having an Opal online increased the win percentage in a deck full of 1-drop artifacts, how could having an Opal in the opener correlate with lower win percentage?

So what I did was make two additional data points.

The first data point I set was to check and compare games where Opal was a dead card at least until turn two (Opal in hand but not online on turn one). This shows a slight drop in win percentage, from 69.82% (879/1259 games) to 67.77% (286/422 games). That's not a heck of a drop, though.

The second data point I set was to check and compare games where Opal isn't online on turn one, but should come online turn two (I didn't check to see if an opponent somehow prevented this with a discard spell, Chalice of the Void, etc.). For this data point, the win rate went up from 69.02% (1132/1640) to 80.49% (33/41). That's quite a jump, so I'm guessing that there weren't many of the 41 game sample size in which an opponent prevented the Opal from being online on turn two. I had it specifically ignore hands which were already counted for in the "Opal online on turn one" data point, so this is specifically only when it was drawn, not online turn one, but able to be turned on on turn two.

So then I wondered, what if we compared the total number of games where Opal was online on either turn one or two? So I set it up to check that, and got 68.85% (1092/1586) when we didn't have an Opal online at all by turn two and an increase to 76.84% (73/95) when we had an Opal online by turn two, as soon as turn one, combined. EDIT: Fixed.

So I recognize that we need to be careful how we interpret the data, but it's very interesting to me. It appears that overall, having an Opal in the opener correlates to a drop in win percentage over a larger sample size. However, in the relatively small sample size in which Opal is turned on in the first two turns, the win rate does increase a decent amount. I suppose that brings us to a good discussion.

First, what's everyone's opinion on the reliability of that small sample size? I don't want to dismiss it entirely, as it does seem to show a trend, but I'm up for hearing everyone's reasoning and opinions on it.

Second, if the small sample size is accurate, we see that there is a generally negative effect of having an Opal in the opening hand, as the instances where we do have it online by turn two is relatively rare (5.6% of games total). But in those 5.6% of games, our win rate increases by 8%.

So, again, I'm interested in hearing everyone's thoughts on this. I understand that typical stock lists seen on MTGGoldfish run four, and prominent recognized players default to four, but I'm looking for independent and rational discussion concerning the evidence, rather than any appeal to authority or bandwagon effect. I'm interested in what each of you, as individuals, think about this development and why.

If anyone sees any more errors in the functions on the spreadsheet, please let me know! I just happened across this one, but having a few other sets of eyes might help.

Thanks!

EDIT: Link to sheet

EDIT 2: I think there's another error in the functions which count up the Turn 1, Turn 2, and Turn 1 or 2 Opal. I'm pretty sure there should be 95 games total (54 turn 1, 41 turn 2), but it's counting it up as 85 instead. Anyone see what might be wrong in the functions? I'm not seeing it (maybe I'm just tired). Fixed it.

1) It is a small sample size; however, it can still provide insight. Also the insight it does provide makes sense. If mox opal is a "dead card" in our hand its like we mulliganed.

2) One problem with this analysis: it does not take into account specific match-ups. It's an aggregated total. The increased win-rate could be substantially higher in certain matchups.

I can name two matchups that would theoretically benefit from a fast mox opal: Jund and any blood moon deck.
Jund because if we have mox opal turned on that quickly it means we were able to get under discard spells. Blood Moon decks because it allows us to actually cast our spells.

I suspect that if you broke out the statistics by deck archtype, the % win ratio with mox opal being turned on by turn 2 would differ radically.
I also suspect it would mirror the matches where we leave mox opal in and/or take it out.

So thnkr, I believe you may be on to something. Essentially the deck slots in Lantern are so tight, we need to heavily scrutinize any change from the normal list. At most, our 60 card deck has 1 flex slot, with most people deciding between Abrupt Decay, or the 4th Ensnaring Bridge. Doesn't give you a lot of wiggle room.

This being said, I believe it is correct to go down to 3 Mox Opals. And I am going to attempt to justify this through math.

Now before I start on taking about Opal specifically, we'll talk about how the difference between 3 and 4 copies of a card changes how often you start with multiples of them in a deck. To calculate chances of you drawing a specific card at a specific time, we use something called Hypergeometric Distribution. Honestly, there are a lot of online calculators that do a good job of explaining how the math works, so I won't go into it here. Suffice it to say, it gives us a lot of statistics off of just a few variables:

1) Population Size - This is the size of the deck
2) Number of Successes in Population - This is how many cards in the deck are the card we want.
3) Sample Size - This is how many cards we will draw from the deck
4) Number of Successes in Sample - For the card we are looking for, this is how many we wish to see in the number of cards drawn

So when making a deck in Magic, what is the difference between a 4-of or a 3-of with concerns to having a card in our opening hand?

Chances of having 1 or more copies of a 4-of in the opening hand: 39.9%
Chances of having 1 or more copies of a 3-of in the opening hand: 31.5%

As you can see, there is approximately an 8.4% difference in having a card in your opening hand if it is only a 3-of as opposed to a 4-of. Now what about the difference between having the card for turn 1? Magic is an interesting game for hypergeometric distribution because you don't draw a card on your first turn if you are going first. So if you are on the play, your chances of having a 4-of or 3-of card for turn 1 are the same as written above. However if you are on the draw, you draw a card on your first turn. This makes the sample size larger and changes our percentages just a bit:

Chances of having 1 or more copies of a 4-of in the opening hand on the draw: 44.4%
Chances of having 1 or more copies of a 3-of in the opening hand on the draw: 35.4%

As you can undoubtedly assume, all the numbers have gone up by a marginal amount. Now why do these percentages matter? In the short term, playing a few FNMs every week, you won't notice a difference between running 3 of a card, and 4 of a card. However at large events like GPs, you play enough games you will notice the difference. For the sake of argument, lets say you win a GP, but every single round you have played 3 games. How many games have you played? Over the course of a whole GP weekend, the maximum number of games of Magic you can play is 54. This is obviously assuming you don't get into some form of ridiculous loop where you and your opponent are drawing games in the un-timed Top 8. But 54 games are a LOT of games, and you WILL feel a noticeable difference in your deck numbers. This is why this discussion is important. For high level play, you play a LOT of magic, and even the slightest changes in variance can drastically change your win percentages.\

Now all of my above points seem to point towards wanting to have 4 Mox Opals in your deck as opposed to 3. So why did I start this off by saying that 3 is the better number? Well it's because Mox Opal is Legendary. This isn;t Chrome Mox, or a Lotus Petal where we can just slam as many as we draw onto the board without consequence. Every time we play our second one, we have to sacrifice the one that is already on the board. And the way the game mechanics work, we can't play a second Opal to turn on Metalcraft to tap both Opals for mana before sacrificing one. So having more than 1 is a detriment. So what are the chances of us having more than 1 Opal?

On the Play - 4 Opals
Chances of 2 or more Opals in Opening Hand/on Turn 1: 6.3%
Chances of 2 or more Opals on Turn 2: 8.2%

On the Draw - 4 Opals
Chances of 2 or more Opals in the Opening Hand: 6.3%
Chances of 2 or more Opals on Turn 1: 8.2%
Chances of 2 or more Opals on Turn 2: 10.3%

We are doing the math to Turn 2, since thnkr's post above discusses our success with being able to turn on Opal for mana by turn 2. At worst, once in every ten games we'll have to deal with having two Opals on Turn 2. At best, only slightly better than that.

On the Play - 3 Opals
Chances of 2 or more Opals in the Opening Hand/on Turn 1: 3.3%
Chances of 2 or more Opals on Turn 2: 4.4%

On the Draw - 3 Opals
Chances of 2 or more Opals in the Opening Hand: 3.3%
Chances of 2 or more Opals on Turn 1: 4.4%
Chances of 2 or more Opals on Turn 2: 5.6%

And here are the numbers that matter to me. Look at the differences between 4 Opals, and 3 Opals. The numbers are almost completely halved. Looking at these numbers I can fairly say that you are twice as likely to get 2 or more Opals in your hand by Turn 2 if you are running 4 instead of 3. For me, this is the deciding factor. Yes the numbers are small percentages. 10% VS 5%, 39.9% VS 31.5%. The question ultimately comes down to which percentages matter more to you. For me, I'll half my chance at drawing the extra Opal

I definitely agree with it coming down to which percentage matters more, because I'm leaning the other direction. You're giving up 8.4% worth of a chance for a turn 2 opal to remove 5% worth of a chance to have one (more) dead card in your hand regardless of if opal is turned on. Maybe I'm looking at this wrong, but I'd rather have more explosive starts with one dead card. Then again, I haven't touched the Whir version yet, so that could be less important now.

Explosive starts are just 1 factor. Opal helps a lot in other matchups like Blood Moon decks or Ponza, or anything with Field of Ruin. It helps to have 1 in grindy matchups to get to 4 mana for Invetor's Fair or Codex Shredder's 2nd ability a turn faster.

Quote from ciago92 »

I definitely agree with it coming down to which percentage matters more, because I'm leaning the other direction. You're giving up 8.4% worth of a chance for a turn 2 opal to remove 5% worth of a chance to have one (more) dead card in your hand regardless of if opal is turned on. Maybe I'm looking at this wrong, but I'd rather have more explosive starts with one dead card. Then again, I haven't touched the Whir version yet, so that could be less important now.

You see here is my situation, with every game of Lantern I have played Opal is terrible Turn 1. I find myself never wanting it in my opening hand, but I want to draw it by Turns 2 or 3. It's a really weird situation honestly.

This is why I personally opt for 3 Opals instead of 4. I think you will lose more games to drawing multiple Opals than you will to not drawing any Opals.

Quote from axman »

1) It is a small sample size; however, it can still provide insight. Also the insight it does provide makes sense. If mox opal is a "dead card" in our hand its like we mulliganed.

2) One problem with this analysis: it does not take into account specific match-ups. It's an aggregated total. The increased win-rate could be substantially higher in certain matchups.

I can name two matchups that would theoretically benefit from a fast mox opal: Jund and any blood moon deck.
Jund because if we have mox opal turned on that quickly it means we were able to get under discard spells. Blood Moon decks because it allows us to actually cast our spells.

I suspect that if you broke out the statistics by deck archtype, the % win ratio with mox opal being turned on by turn 2 would differ radically.
I also suspect it would mirror the matches where we leave mox opal in and/or take it out.

That seems to make sense. We are able to go back and look at the matchups where having an Opal online by turn two were wins and which were losses, and so on. I started to go through and list all of them, but there are quite a few in which it was with decks that are now banned or matchups which we rarely see (Emeria Control, etc.). Thus, I filtered these results to only show the ones where more than 10 total games were played. Here's what I've got:

Deck, Online W/L, Total W/L
Humans 1-0/10-1
8Rack 2-0/25-7
Affinity 2-1/52-21
Bant Eldrazi 1-0/11-3
BtL Scapeshift 2-0/10-6
Burn 4-1/54-37
Death and Taxes 0-1/25-8
Dredge 2-0/26-5
Eldrazi Tron 3-0/22-8
Elves 0-2/34-10
Esper Control 1-1/22-12
Grixis Control 1-0/7-18
Grixis Delver 1-1/29-11
Grixis Shadow 1-1/9-11
Gx Tron 3-2/52/35
Jeskai Control 4-0/28-6
Kiki-Chord 2-0/10-5
Living End 2-1/16-7
Merfolk 3-1/33-7
Stompy 1-0/18-6
Storm 2-0/16-5
UR Delver 1-0/11-4
UW Control 4-0/30-4
Ux Tron 2-2/28-11
Vizier Combo 1-0/11-4

Deck, Dead Opal Win/Loss, Total Win/Loss
Humans 3-0/10-1
8Rack 2-0/25-7
Abzan 2-1/12-2
Abzan Combo 3-1/11-5
Ad Nauseum 5-2/21-4
Affinity 13-6/52-21
Bant Eldrazi 4-2/11-3
Bogles 5-1/23-2
BtL Scapeshift 4-3/10-6
Burn 16-6/54-37
BW Tokens 7-2/18-6
Death and Taxes 6-2/25-8
Dredge 6-0/26-5
Eldrazi Tron 12-1/22-8
Elves 6-1/34-10
Esper Control 4-3/22-12
GriselBreach 3-2/12-3
Grixis Control 2-6/7-18
Grixis Delver 9-1/29-11
Grixis Shadow 2-3/9-11
Gx Tron 7-11/52-35
Infect 4-2/14-10
Jeskai Control 8-0/28-6
Jeskai Nahiri 0-2/11-6
Jund 8-11/31-32
Kiki-Chord 3-1/10-5
Living End 4-2/16-7
Merfolk 3-2/33-7
Naya Zoo 3-0/9-2
RG Titanshift 1-1/4-7
RUG Delver 1-3/6-5
Soul Sisters 4-1/13-3
Stompy 5-1/18-6
Storm 3-2/16-5
UB Fae 4-2/6-7
UB Mill 5-2/16-5
UR Delver 3-2/11-4
UW Control 14-1/29-4
Ux Tron 5-4/28-11
Vizier Combo 3-2/11-4
Zooicide 4-4/12-11

I do need to point out that those numbers on Gx Tron are a bit skewed. The reasoning is due to how lopsided the matchup was in our favor before World Breaker and Ulamog were printed and now in their favor after they were printed. If we only show results from after those cards were printed, we have only four games in which Opal was online by turn two. Two were wins and two were losses. We have 15 games in which we had a dead Opal after they were printed, six were wins and nine were losses.

EDIT: Forgot to post the total wins/losses for Gx Tron since the printing of World Breaker and Ulamog. It's 29 wins, 23 losses. Which strikes me as interesting, because that seems to show that the matchup is still slightly in our favor. I haven't checked the videos to verify how quickly Gx Tron adopted those two cards, when they became staples for the deck. I did take a hiatus from making videos shortly after they were printed, initiated by the Eldrazi Winter and then continued because of work/school/family/volunteer stuffs, and picked back up later that year.

EDIT 2: Well crap (sorry, cursed). Function for Turn 1 Opal was screwed up. Gah. Well, here's corrected numbers:

Turn 1 Opal online: 66.96% (75/112), compared to not: 69.47% (1090/1569).
Cumulative Turn 1/Turn 2 Opal: 70.59% (108/153), compared to not: 69.18% (1057/1528).

Will edit this post to update the rest of the information. Need to help daughter with her school project

EDIT 3:

Hello! This is a wonderful and thorough primer to lantern control. I was delving into your spreadsheet, and found an error in the "Overall Matchup Win %" page. As per your request I would like to help you fix it. While calculating Game Meta % and Match Meta % you seem to be sourcing cells that contain a 0. In fact they are very far down your spreadsheet. Honestly, I'm not sure what value you are looking to find. From the names I think you want: (games won/games played) and (sets won/sets played). If this is the case, I suggest creating two columns that sum matches played and sets played against each archetype. You can simply drag down your equation, and then hide the columns if you wish to keep the look of the current spreadsheet intact. Hope this helps, really appreciate all of the hard work you have put into this!

Thanks! Yeah, I think what happened was when I recently went through and fixed the errors I described in my last post, I found that I'd accidentally mispelled some of the deck names in "archtypes". When I fixed it, those respective cells on the tab you're looking at disappeared, because they'd merged with the correctly spelled versions. Thus, the cells the function is looking for no longer contains data, and I'd forgotten to update the function to correct it. The sum columns were columns H and I on that tab. Fixed it, thanks for pointing it out! If you happen to see anything else, or have any suggestions to make it better, I'm always listening

Greetings, one and all! I'm having some difficulty understanding why the conventional wisdom is to run only 2 Leyline of Sanctity in the sideboard. If we follow the math from the Mox Opal debate above, it seems to me like we need to be running at least 3 in order to have a reasonable chance of having it in our opening hand. Obviously, there's something I'm missing here; otherwise, I wouldn't be asking this question! Any insights and explanations this community can offer would be greatly appreciated.

Quote from sk00gle »

Greetings, one and all! I'm having some difficulty understanding why the conventional wisdom is to run only 2 Leyline of Sanctity in the sideboard. If we follow the math from the Mox Opal debate above, it seems to me like we need to be running at least 3 in order to have a reasonable chance of having it in our opening hand. Obviously, there's something I'm missing here; otherwise, I wouldn't be asking this question! Any insights and explanations this community can offer would be greatly appreciated.

You don't need Leyline on Turn 1. Yes in a lot of match ups it definitely helps to have it, but in some instances you want it to act as extra Witchbane Orbs later on. Add to that the fact that you can't sideboard too heavily with this deck since you don't want to be diluting it, and you have why there are only 2 Leylines in the sideboard.