The Upstart Theory

hobanLast week I talked about why I didn’t think that the 53-card Mermail deck had the theory behind it to consistently perform well. This week I want to take that theory one step further. Traditionally, decks like Dark World or any FTK strategy have run [ccProd]Upstart Goblin[/ccProd] in threes to boost consistency of otherwise inconsistent strategies. Today I’d like to purpose the idea of incorporating [ccProd]Upstart Goblin[/ccProd] in every deck, not just decks that need consistency boosts to have a chance at competing.

Last week I talked about why 40 cards are superior to any number over 40. I’m not going to go back into detail again about why 40 cards are better, but I’ll quickly summarize the points I made last week. First of all, running 40 cards gives you better chances of drawing the best card for a given scenario than if you had more than 41 cards in your deck. Second of all, it allows you to get to the cards you side deck more reliably. Lastly, it allows you to get to cards that are outs to what they’re siding. For example, if they’re [ccProd]Dimensional Fissure[/ccProd], you can get to MST or [ccProd]Dust Tornado[/ccProd] with a higher probability than if you played more cards in your deck.

For the purpose of this article, we are going to assume that the minimum is the best. Mathematically, 40 cards are best because it is the minimum. If the minimum were 50 cards, mathematically 50 cards would be the best. If the minimum were 37 cards, mathematically 37 cards would be the best. [ccProd]Upstart Goblin[/ccProd] essentially makes the minimum number of cards in your deck 37. If the minimum is 37, why play 40?

upstartUpstart’s Costs

Previously this year, I wrote an article on the cost of a card. In it, I talked about how cards have their obvious downsides such as giving your opponent 1000 life points, but I also talked about how cards cost 1 card. What I mean when I say that a card costs 1 card in addition to whatever costs the card has written on it, is that by choosing to play a card you’re forgoing all other cards that could take that spot. Why do you choose to play [ccProd]Mirror Force[/ccProd] over another [ccProd]Fiendish Chain[/ccProd] in the 40th spot? It’s likely that through testing you’ve found it to be more beneficial than [ccProd]Fiendish Chain[/ccProd]. That’s fine and it is entirely possible that this was the best card for that spot. Perhaps [ccProd]Mirror Force[/ccProd] is more beneficial than [ccProd]Fiendish Chain[/ccProd], but in running [ccProd]Mirror Force[/ccProd] in that spot, you’re giving up all the costs and benefits of [ccProd]Fiendish Chain[/ccProd].

If this is the case, what makes running [ccProd]Upstart Goblin[/ccProd] worth running over anything else that could possibly fill the 38th, 39th, and 40th, spot in your deck? This begs the question of whether or not [ccProd]Upstart Goblin[/ccProd] is actually worth running over conventional cards such as a [ccProd]Mirror Force[/ccProd] that might normally fill one of those spots. Realistically, a lot of the time they serve the same purpose. Both of them generally will allow you to safely get to the next card in your deck, but [ccProd]Mirror Force[/ccProd] comes with the added benefit of potentially protecting your field for a turn, but the additional downside of potentially being destroyed by a card like [ccProd]Mystical Space Typhoon[/ccProd] or [ccProd]Atlantean Marksman[/ccProd]. You can weigh the pros and cons of running one or the other in those crucial last spots, but I’d like to argue a different point. I’d like to argue that the cost of 1 card does not apply to [ccProd]Upstart Goblin[/ccProd]. As previously mentioned, playing [ccProd]Upstart Goblin[/ccProd] essentially lowers the number of cards you play. Instead of looking at [ccProd]Upstart Goblin[/ccProd] as taking up 3 spots in your 40 card deck, view it as taking up 0 spots in your 37 card deck. This eliminates the cost of it being 1 card. By viewing your deck as having 37 cards, you shouldn’t weigh the pros and cons of [ccProd]Mirror Force[/ccProd] vs Upstart, instead you should view the pros and cons of running above the minimum number of cards. This is similar to running 41 cards in a 40-card deck, which we have already judged to be inferior.

Because of this, we are able to eliminate the idea of Upstart costing 1 card. This, however, does not mean that we can eliminate other costs. Upstart still costs 1000 life points to activate. This cannot be ignored, but we should accurately evaluate the impact those additional life points have. Let’s take a quick look at the top decks and see how they win. Let’s start at the top with Mermails. They have the ability to throw plenty of damage onto the field at one time and even have an 8000 combo with Megalo Diva. This was the first thing that Frazier brought to my attention when I mentioned the idea of running 3 Upstart Goblin in Mermail to him. In order for 1000 life points to make a difference, you must

  1. Draw [ccProd]Upstart Goblin[/ccProd], before the combo (if you have the combo already, you can not play Upstart).
  2. Draw Megalo with Dragoons and another Water or Megalo with 2 other Waters with Diva in hand.
  3. Your opponent must not be able to stop the combo.
  4. Your opponent must not have taken more than 1000 life points, prior to the combo.

fire fist - bearRealistically, this doesn’t seem very likely to impact the likelihood of the combo.

Next let’s look at Fire. Fire is a deck that doesn’t win off large pushes for massive damage. It wins off advantage. In fact, Fire is generally incapable of making large pushes. In decks like these, there is a significant reduction in the importance of life points as games are usually decided on a control basis rather than simply life point totals.

Then there is Rabbit. This is a deck that is similar to Fire in that it has a large control aspect to it, but is also very important to establish a clock. Essentially, how many turns is it going to take to beat your opponent? With a [ccProd]Sabersaurus[/ccProd] and no life points, you’re going to have to attack your opponent 5 times before their life points reach 0. That’s 7600 on the 4th attack and 9500 on the 5th attack. This means that you can play 1 Upstart Goblin, and still keep that 5 turn clock. Conversely, with[ccProd]Kabazauls[/ccProd], it adds a turn to the clock counter, an important downside to be aware of.

Evilswarms once again share the control aspect of Fire and Dinos, but a clock can also be important. If you have an Ophion up, your opponent is on a 4 turn clock. This clock allows for you to play 2 [ccProd]Upstart Goblin[/ccProd]s and still be on the same 4 turn clock. Additionally, Heliotrope and Thunderbird (at 1950) give you a 5 turn clock, and allow for 1 Upstart to be played while keeping the same clock.

You’re going to have to decide if these 1000 life points per Upstart are going to make too large of a difference to make it worth playing. I’ve decided that in the current format, they don’t make a big enough difference.


This is an issue that I think must be handled delicately in order to effectively utilize this theory. The format is in a unique position where Water is the most played deck, but doesn’t play many trap cards, but every other deck in the format does play a fair amount. This one of the many paradoxes this format has. It essentially means that cards like [ccProd]Mystical Space Typhoon[/ccProd] and [ccProd]Forbidden Lance[/ccProd] are relatively subpar when you play against a deck like Water, but amazing against pretty much any other deck in the format. I think [ccProd]Upstart Goblin[/ccProd] actually helps alleviate this issue. It allows you to cut down on the number of 3 ofs in your main deck. Let’s say that instead of running 3 [ccProd]Forbidden Lance[/ccProd], you only run 2, but are playing 37 cards because of Lance. Because of this, you’ll only experience a slight drop in the number of times that you’ll see a single Lance in your hand, but you’ll see a significant drop in the number of times that you’ll see multiple Lances or combinations of cards like Lance and MST together. This drops in the number of times you’ll see multiples can help reduce the amount of clogging and allow for more consistent hands.

Similarly, you’re going to have to take measures to prevent clogging when you side deck. It’s not uncommon for Water decks to side deck 2 or 3 [ccProd]Dust Tornado[/ccProd]. Well if you side deck a full 3 copies, between 3 MST, [ccProd]Heavy Storm[/ccProd], and 3 [ccProd]Dust Tornado[/ccProd] WHILE you’re only playing 37 cards it’s likely that you’re going to see multiples quite often and have quite the issue with clogging. There are two possible solutions to this.

  1. You could side out Upstart Goblin. I don’t recommend this as this would involve playing more than the minimum number of cards, which I believe to be the best.
  2. You could side less Dust Tornados to decrease the chances of opening multiples. If instead of siding 3 Dust Tornado with 37 cards, you side 1 or 2 with 37 cards you’re going to draw multiples significantly less, but still be able to get to them the majority of the time.

I definitely think that the latter is the better solution. I think that to be the most effective with this, you should cut down on the number of 3 ofs you run in general and keep it to absolute staples only like Bear or Tenki, not [ccProd]Dimensional Prison[/ccProd] or [ccProd]Fiendish Chain[/ccProd]. This will allow for more consistent hands.


The last thing I want to talk about with [ccProd]Upstart Goblin[/ccProd] is the externalities of running the card. Let’s say for instance a factory creating things, but at the same time putting carbon dioxide into the air. The factory has certain costs like equipment or paying its workers that it must calculate, but there are additional costs that are the factory owners do not have to pay, but still impact society such as polluting. Similarly, [ccProd]Upstart Goblin[/ccProd] has these externalities associated with it. Sometimes externalities are a good thing. Let’s say that you’re running [ccProd]Summoner Monk[/ccProd]. While the main idea behind [ccProd]Upstart Goblin[/ccProd] is to be running the minimum number of cards, you it is still technically a spell card that you are able to pitch for [ccProd]Summoner Monk[/ccProd]. On the other hand, sometimes externalities are negative. When the new set is released, [ccProd]Spellbook of Judgment[/ccProd] will often be set to punish the opponent for using too many spell cards. You don’t want to give your opponent too much advantage off this, so playing Upstart against Spellbooks could have certain negative consequences. Additionally, if cards like [ccProd]Eradicator Epidemic Virus[/ccProd] become heavily played as a response to Spellbooks, [ccProd]Upstart Goblin[/ccProd] might get inherently weaker because it would become a prime target.

That about wraps up this week’s article. I’m still not quite sure what I’ll be playing this weekend in Jersey, but whatever it is, it’s very likely that it will have 3 copies of Upstart Goblin. Until next time everybody, play hard or go home!

Patrick Hoban

Patrick Hoban

Patrick Hoban

Latest posts by Patrick Hoban (see all)



  • fewer* Dust Tornadoes

  • Oscar Contreras

    Despite this being an old article, let me see if I have this correct, because I notice some people using it and when I saw Upstart in the article I think I had an idea on what it could be: Fill in the gaps with upstart or side them in when playing against curtain deck like Water that run minimal backrows etc… and instead of usual 3 Mst or Dust Tornado run 2 and something else like upstart if you want to to draw them and not clogged up your deck during or in between duels etc… I hopefully got that right. Also understand if Upstart can work in your deck and using it wisely.

  • BenRM

    Interesting piece.

  • Brian Pence

    You’re forgetting a number of reasons why you WOULDN’T need a minimum. Statistically, your logic is infallible. I can tell you right now though, that when it comes to deck thinning and the meta, 37 cards is not the best. First of all, any card that you need to run in three increases its probability of doubling up in your hand by 2%. Now, this doesn’t mean much to the untrained eye, but if you crunch the numbers through a hypergeometric distribution calculator, this means that every single card you run in three has a 35% higher chance than it would in a 40 card deck of doubling up in your hand. Not too big of a deal at such small percentages. Now say you have three seperate staple cards that you run in three (let’s say we’re talking about verz with infestation pandemic infection, heliolope and forbidden dress/safe zone/ imperial iron wall) you’ve just more than increased the chance of drawing two cards that do the same thing by about 70% over what it was. It gets worse the more cards you run in multiples. If you sport 4 sets of three cards (a feasible number for some decks) you’ll see 85% more duels with doubles in hand, which will usually leave you with a dead card. furthermore, your upstart theory increases the chance of a given set of doubles of a set of two to end in your hand by 30%. Now, most decks in the meta tech cards in sets of two by the buttload. To give you an idea, reducing your deck by only 3 cards will increase the total probability of grabbing doubles of sets of two by 45%. I can disprove your “theory” simply by sending you to Jason Grabher Meyer’s Upper Deck articles on hypergeometric probability (so you can do these calculations yourself) and the reasoning behind the 45 card fifth gadget deck. Though upstart makes sense in decks where multiples aren’t really a problem, taking 3 cards out of a game that was designed for 3 card set multiples in a 40 card deck will statistically kill your deck. Unless you want to start running your power cards in sets of two. that’s a whole different monster, and I can tell you that it also will fail you. This is why so many 41+ card decks win, they force hand diversity by decreasing the NEED for consistency and thusly the consistency itself. Throwing in 5 extra cards that do 5 different things will actually make your hands come out without multiples 42% more often, that means no double chains, no double helilope, no double tengu, no double vanilla. You’re preaching what duelists have been trying to assert for like 6 years. The Upperdeck masters already proved this theory wrong a loooooonnnggg time ago.

  • I usually side a few upstarts when I play antimeta decks and grind decks. They increase the odds of drawing into multiple fissures/cosmos to deal the mermail’s msts/storms.

    I prefer not to main em. In the first round, it feels that most of the cards in the decks I play have the same weight. There’s no trap better than the other. They always have some pros and cons which have equal weight in my decision to run em. If I need to though, I’ll side the upstarts in in the event that the cards I’m siding in for my opponent are better than the the cards I side out in that particular match.

  • To elaborate on the point on externalities, there are legitimate cards in the meta that can punish Upstart and make it become dead. Nat Beast, Shock Master, EEV. Even if you don’t run a deck that players would side EEV against, all you have to do is face one player who maindecks it and all the benefits you experience from 37 over 40 will be offset by that one play.

    In addition, there are decks that run 3 of cards not to maximize drawing them but because they actually need 3 of that card to do well in a grind. Last format’s Wind-Ups is my favorite example of this, and is also a deck I would not play 3 Upstarts in.