Small Earth Template Review and Guide - Better than Discovery’s Small Earth Walkthrough©
Kenghis Ghan (game theory, strategy analysis, case studies), Meth Damon (strategy analysis, case studies, additional notes), Phobos (classification, case studies)
Last updated May, 2025
(Use the sidebar for quick navigation through the sections.)
Abbreviations
SA = South America
AU = Australia
AF = Africa
EU = Europe
NA = North America
AS = Asia
NAf = North Africa
ME = Middle East
FTB = First turn bonus
Introduction
Small earth is a template that many can play adequately, but few master. This guide aims to cover everything there is to know in playing Small Earth effectively.[1] Before reading, it is assumed that readers possess a complete understanding of the mechanics of turn order, attacks and pick intel.
We will begin our discussion on the intuitive approach to playing the game. This is useful for new and intermediate players as it helps teach the game/template in a digestible manner. Later, we discuss optimization and finally the game theoretic approach. These parts are more suitable for advanced players since it introduces a different approach to thinking about Warlight. The best modern chess players are all computers, and the goal is that with some of these ideas, you too can begin to think and play like a machine.
The Intuitive Approach
First, the basics of income. In small earth it is paramount to not let your opponent get an income advantage for any prolonged period of time. Unless you are much more skilled and/or have an overwhelming stack, you will lose. Below is a list of possible incomes after various turns:
Turn: Income
1: 7 (Australia or South America)
2: 9 (South America & Australia)
2: 10 (Europe)
2: 10 (South America & Africa)
2: 10 (North America with adjacent start)
3: 10 (North America with single start)
3: 12 (Europe & (South America or Australia))
3: 12 (Asia with adjacent start)
4: 12 (Asia with single start)
The defining feature of a small earth board is the wasteland. We can therefore simplify our analysis by categorizing a board by its wasteland. They are the following, in order from most common to least common:
1. Asia wasteland
2. North America wasteland
3. Europe wasteland
4. Africa wasteland
5. South America wasteland
6. Australia wasteland
7. North Africa wasteland
8. Siam wasteland
(1-4) “The default game”
These boards share similar playstyles. The majority of games will have picks 1 and 2 focused on a combination of SA or AU with EU, AF and/or NA constituting the rest. The goal is to secure SA or AU without allowing your opponent to hold EU/AF/NA/AS (or transposing into a position where the opponent has safe expansion into one of the aforementioned territories). Games between skilled players usually involve the player in SA maintaining a stack in NAf and the player in AU maintaining a stack in the ME. The goal is for both players to slowly expand around their respective bonuses while preventing EU/AF in the short term. These bonuses almost always tend to be NA for the player starting in SA and AS for the player in AU. Players may also attempt to break the opponent’s “safe” bonus by routing through NA for SA or AS for AU. While we call these boards the default game, understand that the exact locations of starts will produce a lot of variation in how games are played. We try to showcase examples of each category in the following case studies.
Case Studies
This game features a bit of brawling. I get South America turn 1, Vindude threatens with stack and slow expands Australia. I secure North Africa, and threaten Australia. Vindude seizes Europe which I insta-break. I consolidate a stack which heavily threatens Australia. Won by keeping the pressure up and eventually taking Europe.
Here NA is not as strong due to the W. EU pick. This game demonstrates some attacking play. huddyj miscalculated AS expansion and did not take Kamchatka with enough armies to hold it, otherwise the game would have been winning.
This is a standard NA expansion. Very little can be done for DemHunt in this situation. These boards occur commonly and are usually a winning game for the player with the first pick.
This game features an AS and NA rush.
This game features a 3 turn AS. The opponent presumably did not attack ME because they feared a large stack attack and also wanted to counter NA. This was not the right decision. Never underestimate the power of AS.
Here I take Europe in 2 turns while Miroslavg grabs South America + Africa. We both break each other on turn 3. Miroslavg then plays a predict turn 4 which lets him retake Africa despite my best efforts to delay sufficiently. If you want to know the value of delay pay attention to this turn. From there he correctly predicts every move I make resulting in my defeat.
I counter Siam counter and got Australia late. A stack standoff in Africa ensues as I slow expand Asia. Camille moves toward Australia through North America. Camille commits too much in Asia allowing me to break stack and seize Africa. Camille tries to take Asia, but it's too late.
Here is a game that features a NAf and EU pick. Beginners will often take the opportunity to block off NAf and slowly expand in EU. From an income point of view, EU is a more efficient bonus than SA/NA and 5>4. I do not recommend that you go for this approach however, which is demonstrated by the game. EU is simply too risky and perfect predictions are unlikely.
The start in this is rather unique, since both SA and AU are unsafe due to adjacent bonuses. I decided that SA was more valuable due to future NA expansion, and it paid off.
Unsafe SA allowed me to continuously put pressure. The optimal move for Fizzer would be to block off Brazil and slowly expand into AF.
Here is a tournament of games between primarily SE players for further studying. Here are some more tournaments.
(5) South America wasteland
This takes away the most powerful starting bonus in the game but creates a new threat: safe NA. If you see a wasteland in Brazil or Venezuela, NA usually becomes the first pick with the EU as the second. Even though it takes 3 turns, the Europe/North America combo can get 10 income while denying opponents more than 7 income. It usually wins unless you play badly or your opponent counters it exceptionally well.
Case Studies
In this game NA is very powerful since after it gets taken in 3 turns it is very hard to counter. Both players realized this. Take a look at Ozijs’ picks, they are structured in such a way that any reasonable pick set by the opponent will not produce a completely lost game (still losing though).
In this game the opponent missed the power of a safe NA, which leads to a 3 turn capture and break. While this is not a guaranteed win for me, The inefficiency of the AF and AU bonuses and my position helped. Here is a game with a similar board. I expanded to both NA and EU just in case there was an AS counter.
I go for 3 turn north America while breaking Africa. Despite having no bonus before then, Ilpegg can't break North America and loses in the long run.
In this game I took W. EU with a 3 to make it look like I was expanding in EU. If I did not encounter him there then I would have expanded in EU instead, since the assumption would be that he picked AS or SA. The opponent fell for the EU expansion and stack attacked instead of immediately grabbing AF.
(6) Australia wasteland
Compared to the SA wasteland, games of this type are more involved to analyse. Picks are split between EU rush and SA + NA/AF. AS may also come into play. These games are rare, and we will provide a few case studies to see some general ideas. In both (5) and this one, correct picks are absolutely critical. Pick in such a way that you do not get in a completely lost situation given any set of reasonable opponent picks.
Case Studies
In this game Ozijs realizes that the game is lost without breaking SA, so he converges on NAf with both starts. It ended up going his way.
This board has the threat of two turn NA. I structured my picks so that I would either get two turn NA or I would be in a position to counter it by stack attacking SA and taking AF. The opponent chose an inferior pickset; it was losing from turn 1 due to how far away the NA armies are. (Stack advantages are crucial in the beginning of the game.) After I noticed the opponent expanding in NA, I immediately took AF and began to threaten EU or SA stack attack.
In this game the only feasible bonus after SA was EU, so I countered it.
(7) North Africa wasteland
North Africa wasteland creates a similar game dynamic. These games usually involve a fight over SA and NA. It also creates a relatively safe 2-turn Europe. This is a scenario in which it may become worthwhile to break a wasteland for positional gain as it often surprises the opponent and can let you break EU.
Case Studies
This game has a stack up between NA and EU. In order to change the game dynamics, Phobos broke the NAf wasteland which allowed a surprise attack on NA. The board was winning for Alex, but they failed to take advantage of the counter.
The location of the other picks makes EU safe, so I made that the first and second pick. The opponent had different ideas for picks and lost.
(8) Siam wasteland
Last but not least is Siam wasteland. The wasteland traps remaining armies from AU expansion, making them only useful as delays. This leads to a rough early game for the AU player and makes AF/EU more viable since there is no counter coming from the east.
In this game I could take AU as a FTB, resulting in 18 armies in EU at the end of 3 turns; or in two turns, resulting in 19/18 armies at the end of 3 turns; or in three turns, resulting in 19 armies at the end of 3 turns. I went for the second option which was suboptimal since due the board it was strictly dominated by the first. My opponent played rather poorly — they decided to not take the SA FTB despite having an EU counter and I was fortunate enough to have an easy game.
Wilson picks SA counters, and it becomes a guessing game with FTB south america conferring an advantage.
Taking Australia as a FTB is not optimal given an immediate NAf attack since it takes time before the FTB pays off. Instead, the best options would be to either take AU as an FTB and not move EU or take NAf with 9 armies and go for 2 or 3 turn AU depending on the opponent’s response. Phobos was able to use the army advantage in SA/AF to put pressure and win.
Other Small Earth Templates
SEAD
Optimal Small Earth Auto Dist play requires a guide of its own and will not be covered here. As a general rule, remember to always counter the greatest threat. See our discussion of minmaximizing here. It may also be advantageous to disguise bonus captures so that the opponent counters the wrong threat. Some combinatorial and probability based work may also be required. Also, information tends to be more valuable in SEAD than in SE1W and it is often worthwhile to spend extra armies on expansion to get information. See Information Minimization for more details.
SE NLC 1ASG
Also known as the old Small Earth template, expansion is made far more costly in this template due to the one army stand guard setting (taking neutrals removes 2 armies from your usable income instead of 1), and thus it is necessary to avoid overexpansion and only take one territory per turn in order to receive card pieces or to counter a bonus. Exceptions apply, for example if starting in a FTB or expanding in a safe bonus.
No-luck cycle is another significant setting that should be exploited for a winning game. Most boards in small earth have unequal starts (for example if getting adjacent 2nd and 3rd picks leads to a safe bonus, but are dangerous to pick alone). For this reason, it is best to pick and commit immediately if being first in the pick order is advantageous or to wait as long as possible and allow autocommit to activate if the second pick order is more desirable. No-luck cycle tracks the amount of time between clicking the start button and committing, therefore it is best to click start as soon as the game begins and to allow autocommit to commit for you when the boot timer expires. If this is what you are going for and you see that your opponent has already committed, please be courteous and commit since autocommit will not be necessary.
This mode of play typically leads to either long games of attrition (in which both players have stacks of 50 or more and are trying to complete NA/AS) or short games in which one opponent loses immediately on picks due to a safe combo, a wasteland in SA/AU or a counter.
SE COMBOMB
The commander and bomb cards make some new strategies possible, but the same general advice in this discussion applies. Specifically, the commander makes breaking wastelands more feasible and allows for more traps and tricks. EU is also better in the default game due to the commander. There is not much more we can say other than to keep track of the board state at all times and use your commander to its maximum potential while keeping it safe. Also, it is advisable to not use the bomb card when it is not necessary. There are often better opportunities in later turns.
Notes on Strategy
Army Optimization
Every army in the board is valuable. In games on other templates it is not as important to optimize army count since positional or information advantages matter more, but in SE your army count should be the main focus. Because of this, bonus efficiency is good to keep in mind. In decreasing order, they are AS (12/7), EU (7/5), NA (9/5), AF/SA/AU (2/1).[3] Before deciding to take a neutral for information or to threaten bonus completion, try to look ahead a few moves and see whether or not it is worth it.[4] For example, it does not make sense to spend 5 armies taking a bonus for 3 armies if you know that you can’t hold it for more than a turn. In general, it is good advice to take as few territories as possible. Avoid expanding into neutral territories in bonuses that you do not intend to complete unless it is required for a card piece. Instead, let the opponent expand for you and then take over their territories. With that being said, do make sure that you get a card piece every turn. Each card piece is worth 4/3 armies, which is greater than the 1 army cost of taking a territory. It is often a good idea to hold cards in situations where playing a card would not confer a stack advantage or is not necessary for defense/countering, but you do not want to be behind on card timing since it can lead to games like this.
It is also good to think about army transfers. Look ahead 2-3 moves and think about the best order to take neutrals in to minimize the amount of time required to move remaining armies to the main stack. You should also consider grouping armies to be transferred whenever possible due to turn order — a first order transfer of 6 armies is very different from a first order transfer of 2 armies and two second order transfers of 2 armies. Sometimes you may prefer the reverse, and it can be a good idea to deliberately split remainders to maximize the amount of delays without deploying additional armies. These small optimizations can sometimes decide games.
One more tip: you may have heard of the timi blockade (or the lesser known arrow blockade). The core idea is that all attack/transfer orders reserve all of the armies in the order and prevents their use in future orders.[5] The transfer only setting is not turned on in SE, so we cannot use the same trick, but it is certainly possible to reserve a single army by attacking a neutral of size 2 with a 2 (or making any attack designed to fail, for example 4v3). This may be useful in rare situations involving large amounts of delays or for trick prevention.
Information Minimization
In SE, some boards will not give your opponent immediate pick/turn order information. In order to maximize this advantage, you should minimize the amount of information exposed to your opponent. For example, you can take territories at crucial borders at the last allowable moment. Or in situations with unclear turn order, mix up nonessential orders involving territories adjacent to the opponent so that you do not get two in a row or have the first visible attack/transfer. Note that deployment orders follow the same rules, and you should deploy away from borders first. Both players should also utilize zero taps to gain intel on turn order.
There are also ways to feign expansion (see this game). In general, the more uncertainty you give your opponent, the better. This game is a great example of this, my opponent was doing some very strange things with his stack, which forced me to expand to Mexico and divert my stack.
South America or Australia
When choosing between South America and Australia for the first pick, it is generally advisable to pick whichever one has a FTB as that results in a greater number of total armies after 2 turns. There are exceptions to this rule, for example SA can be a stronger counter. If both picks are safe FTBs, then SA is generally better due to the fact that the 6 leftovers from first turn expansion can be transferred immediately and used to either attack North Africa or to simply defend Brazil and race for North America. AU is weaker for this exact reason. It takes far longer to counter NA/get Asia from AU, and the 6 remaining armies are more difficult to utilize.
SE is one of the rare templates where you can count and know where every single one of the opponent's armies are or could be due to the small pool of picks. We recommend that readers do so and to be careful about picks, since a 2 turn EU or NA or a 3 turn AS expansion can quickly end the game if not countered.
Stack Advantage
Under standard kill rates and straight round, we say that an attack is efficient for a attackers and d defenders whenever
To avoid having to do casework, we can lose a bit of precision in our conclusion and consider the simpler case of weighted random where attacks are efficient whenever
Notice that this condition depends only on the ratio of a and d, and if they both increase by the same quantity (that is, assuming equal deploys), the ratio will tend towards 1, which is less than the efficient ratio. In SE, this means that army advantages are most pronounced during the beginning of the game, when both a and d are relatively small. This is when it is most important to consider transfers and choice of picks. In this board for example, AF would not be the optimal 4th pick since the early stack advantage given by the SA FTB results in a disadvantaged game. Instead, it is better to pick EU since it gives more time to minimize the ratio between a and d.
Counting and Delays
It is necessary to keep track of the quantity and location of all of the opponent’s armies at all times. If the opponent is seen deploying less than their income, figure out where it is going. If the opponent does not play a reinforcement card, remember that. With good enough board vision, it is possible to count delays. For example, your opponent can get in a situation in which their attack succeeds if and only if they use a significant proportion of their income on delays or if they stack attack without delays. With complete information on the board state you can capitalize on this information.
A note about delays: if you go first on a certain turn, you should aim to have an odd number of delays. Conversely (and contrapositively), if you go second, you should aim to have an even number of delays. The reasoning is obvious if you consider turn order.
Breaking Wastelands
We said earlier that the game of SE is determined by the wasteland. This is not entirely true. In longer games, the wastelanded bonus can come into play when one player breaks it for income or positional advantage. This can be the result of a unique and calculated position where the opponent neglects to counter a bonus due to the wasteland (thus making it safe and viable for long term income) or out of desperation.
Case Studies
In this game, Thermidor sees that Alaric has moved the main stack into NA and out of the main theatre. By breaking the wasteland, Thermidor sacrifices armies in the short term to threaten long term AF.
huddyj v. (ง︡'-'︠)ง let's fight!!
Huddy has an army advantage but an income disadvantage after AF is broken. So the logical step is to break the wasteland since SA can be held with the main stack.
Tricks (Predicts)[6]
When playing with an army disadvantage (also known as a losing game), certain opponents will try tricks instead of surrendering. These usually involve exploiting turn order or recaptures. Sometimes players offer an efficient attack on a small stack in exchange for positional gain (think poisoned pawn).
Case Studies
Turn 6 I made an error with turn order and believed that I was able to transfer my stack of 10 to the main one in ME first order (I couldn't). Instead, I attacked NAf, hoping that the opponent would hit ME with the entire stack in order to maximize the amount of armies killed.
Same idea as the previous one. Also note the importance of denying reinforcement card pieces.
This one involves a disadvantaged EU, AF start. Note how predictions completely changed the game.
This game should have been an easy win for huddyj. At one point he had a 14 army advantage!
There are some more examples in this tournament if you are interested.
The best advice we can give of avoiding tricks is to play passively and attack with the minimum necessary amount of armies to complete your objective. It also helps to know the board state at all times. If you have significantly more armies than your opponent, do not be afraid to fragment your stack in order to safely attack.
Luck
There is a degree of luck to any standard SE game. Take for example the board of this game. My picks were optimal (taking AU, NA as second pick would very quickly lead to two turn AF and taking NA, AF and secretly expanding in AF would lead to a losing game), and due to turn order I was immediately put into a losing game. There is no good way to resolve this (see the section on NLC), and even with perfect play, some boards strictly advantage the player with a certain turn order. The optimizations and strategy that we discuss here can only take you so far, and sometimes moves by nature will determine the winner. In the next section we will take all of these ideas into account and discuss some mathematical abstractions that can be applied to all games of Warlight.
Optimal Play and Abstractions
Game Theoretic Discussion (incomplete)
It is an unfortunate fact of life that the lessons of game theory are rarely applicable. There are ten thousand papers on auction theory but no auction theorist can give a bidder optimal advice in a real auction. This is because people in the real world are rarely rational, have incorrect and/or private information and do not play equilibrium strategies. Warlight players are no different, and even in duels it is often hard to perform good mathematical analysis. FFA games are even more complicated to analyse due to collusion possibilities.[7] Despite this, we can use the lessons of game theory to help inform decisions. In the next few paragraphs we will provide some preliminaries to game theory and its applications in Warlight[8]:
Fundamentally, Warlight is a strictly competitive dynamic game of imperfect information. Using subgame perfect Bayesian equilibrium (the Bayesian version of SPNE) as our solution concept, we find that the only subgame NE are cartesian products of the stage NE. This is because Warlight is an elimination game and each stage game is zero sum. Intuitively, this means that there is no bargaining — it makes no sense to threaten your opponent in chat to not do a certain move. The optimal overall strategy must therefore be optimal in every stage given types. In-zero sum environments the equilibrium payoff is unique: they are the minimax equilibria or mixtures over those equilibria. Now, in finite zero-sum games with perfect information the minimax theorem guarantees that there is a pure strategy equilibrium (we will not consider the extremely rare case of drawn games and assume that all games are finite). Chess, for example, is guaranteed to have a pure strategy.
Unfortunately, we cannot apply this result to Warlight. The reason is that we have simultaneous moves and fog (even in games with light/no fog simultaneous moves present a problem). In exceptionally rare cases there exist saddle points in which we have a pair of pure strategies where each strategy is a best response to the other. (I have observed this in games where after attacking, the income gained is exactly cancelled out by the armies lost on both sides and further bonus acquisition is suboptimal for the attacker.) In nearly all other cases, we will almost surely have a unique or infinite number of mixed equilibria. We will define actions in WZ to be a list of moves, so our equilibrium(a) will consist of players choosing a strategy that guarantees the game’s value over the expectation of the opponent’s types. The takeaway from this is that even a perfect player cannot win every game, and optimal strategies are history-independent.[9]
Now let us get some intuition about finding these minimax equilibria.
Calculation and Searching
We know that Warlight is a zero-sum game, or in other words, your opponent’s loss is your gain. We first present some definitions and an ordinarily dry proof result about games of perfect information that have analogies in games of imperfect information:
In plain English, a maxminimizing action is one that maximizes the payoff that the player can guarantee. Similarly, a minmaximizer is one that minimizes the maximum payoff for the opponent. The lemma shows that for strictly competitive games, these are the exact same. With respect to Warlight, this means that the actions that minimize your opponents’ best situation are also the ones that maximize your worst-case payoff.
How do we apply these ideas to playing good moves in actual games? Well, the lemma tells us that we should search for the best actions that the opponent can play and counter the largest threat (minimize their best situation). Equivalently, we should play the actions that maximize our position in the worst possible case. As mentioned before, there is rarely a pure strategy that you can play all of the time, so each of these minimax equilibria are going to be sets of actions with a probability distribution assigned to them depending on the opponents’ payoffs. Depending on whether the board state is degenerate, there can be a single unique minimax equilibrium or a convex combination of strategies. In either case, it is possible to compute these equilibria.
We understand that this provides little direct guidance in games, however it is a good general concept to keep in mind. In small templates such as Small Earth, it is occasionally possible to find these optimal moves.[10] (In winning boards, it is possible that most moves are optimal.) Generally, it is advisable to calculate as much as possible and use heuristics only to prune the game tree. We define calculating here to be simulating all feasible opponent moves based on available information and constructing the best response. Just as how chess masters can defeat amateurs by calculating lines and viewing possibilities that the amateur cannot see, and chess engines can beat chess masters by calculating variations that the chess master cannot see, Warlight players can better their odds against an opponent by calculating more than their opponent.
[1]This guide primarily discusses Small Earth 1 Wasteland, however the lessons here can be applied to all mainstream templates that use the Small Earth map. The game theoretic discussion at the end is applicable to all games in general and is tailored towards advanced players.
[2]I have noticed across many games that my opponent tends to play QM with a 6 coin wager. An unrelated fact is that the rake to entry fee ratio is greatest for 6 coins at 1/3. (To be maximally tax efficient, create games with entry fees that are multiples of 50.)
[3]Because of this, NA is generally preferred over AF+SA/AU. It is also harder to break.
[4]Keep in mind the sure thing principle.
[5]This is not the case in multiattack games. Armies in failed attacks can be used in future multiattack orders.
[6]Information regarding how to effectively set up tricks will not be divulged in this guide. Tricks are for kids.
(You should have a general idea just by looking at the games.)
[7]The folk theorems tell us that in such situations, almost anything goes.
[8]If you do not understand what some of these terms mean, refer to Osborne and Rubinstein’s A Course in Game Theory Ch. I-III.
[9]These strategies are not easy to find. Look at Colonel Blotto games to see how even a simple variant of this problem can be hard to solve.
[10]I have a suspicion that multi attack templates are easier to calculate. I can’t prove this and have no evidence for this (if anything, intuition should point us in the opposite direction), but actions in multiattack games “feel” more forcing, which may help in terms of searching.