In the past few months, I had an opportunity to play more games than I usually do, and I've been gravitating towards a mix of team-based strategy and tactical games. It seems that there are a lot of parallels between good decision-making in games and other areas of life - I don't think this is a novel observation, but I do think there are under-explored lessons, like reviewing your own replays to improve. So here is a list of ideas I've compiled focusing on the more strategic aspects of running a startup.
Identify your win conditions: It's important to establish a theory of victory, and then construct a strategy that matches it. This includes identifying how and when you are strong relative to your rivals, and how you will convert that into wins. While this might come off as "have a plan," I think it is slightly more than that: it should describe the internal logic of your strategy and why it will make your victory inevitable and your opponents give up.
A common beginner-level trap is to simply "play not to lose," in which case you can easily be put on the defensive by an opponent with a clear plan and initiative. But a more subtle trap is to come up with a strategy - aligning the ways and means with the ends - without making explicit the links between them, or why it will work. This makes it harder to improvise when things inevitably go off plan, because it becomes easy to lose sight of the main ideas behind why a particular strategy works, and therefore harder to adapt or salvage in the face of a determined foe.
If you have an edge, exploit it ruthlessly. It's not enough to simply create advantages, you must set up and create situations where you get to use them. Aggressively confronting your opponents when you have an advantage pressures them and takes away resources and wins they could otherwise get uncontested. This lets you claim more resources overall and snowball faster.
In contrast, if you are behind, avoid attacking into the enemy's strengths. Avoid direct confrontations and play to catch up away from your opponent - e.g., on the other side of the map or in a slightly different vertical / customer base.
In general, if you're ahead: you want to speed up the game and make more plays, so you can maximize the number of times you get to use your winning position. As a silly trivial example: you want to bet on as many coin tosses as possible when you know the coin is 60%/40% in your favor.
A corollary to this is that you have to use your resources to help you win. Cash on the balance sheet, hired but idle units, or finished but unlaunched features are examples of resources that aren't doing productive work. All else being equal, the player with the more efficient resource utilization converts them into wins more often.
Play around your winners: It's usually a mistake to focus on your weaknesses to try and become more well-rounded. Because winning gets you advantages that let you make better use of additional resources, it's better to support your winning players and teams so they can continue to snowball. This lets them extend their lead, and can carry the whole organization.
For the same reason, it's much better to be great at one thing than to be just "good" at all things. It's too easy to be outperformed by a team of specialists. This applies recursively, from large complex teams down to the individual level. Double down on the star players, the functions that are dominating in their roles, the winning flank in the battle, etc, and the momentum is self-reinforcing.
Don't gamble away leads: When ahead, take fewer risks to avoid throwing your lead. For the opposite reason: when behind you should eventually try to make bigger, riskier, plays. Since you by definition want more volatility, it is advantageous to force the opponent to respond to you. Playing conservative from behind is usually a mistake as it simply perpetuates the (losing) status quo.
Reflect on probabilities, not outcomes. It's a mistake to judge whether a decision was good or not based solely on the outcome. For example, you usually wouldn't congratulate a lottery winner for making a good financial decision to buy lottery tickets, because it's important to consider the prior odds of success, and not just whether the result was good or bad. This encourages good decision making over time that adds value, instead of being misled by random events outside your control. On a related note, it's only of limited use to analyze situations with information that was only available in hindsight, because in similar future situations we won't have the benefit of hindsight.
Use information to your advantage. Gather and analyze information about adversaries to predict their movements and counter accordingly. Withhold information or actively deceive them so they can't do the same to you.
Think in terms of threats and counter threats: It's a mistake to defend against every threat or try to compete for every objective that is contested. As in chess, sometimes a better move is to credibly threaten something else, or take something valuable the opponent has left open by committing resources elsewhere. Sometimes threatening to trade rooks is the right play.
Think in terms of timing windows. Anticipate when you or your opponents will be strong or weak, and in what ways, and use it to your advantage to get something you normally wouldn't be able to get. Sometimes this happens unexpectedly so it is important always be prepared to take advantage of opportunities that may present themselves. The obvious timing windows for startups include product launches and fundraising announcements that can be amplified to find new customers or used in recruiting, but may also include industry/competitor developments that allow you to insert yourself as the obvious solution or follow up for some event.
Win or lose as one. The most important principle in working as a team is to prioritize the team's goals over the individual's. This often means interrupting in-progress activities (and at the expense of personal advancement) to go help the wider team. This can be frustrating and feel slow and inefficient, but is often optimal because well-coordinated mutually supporting efforts can be much more effective than numbers alone might suggest. For example, 3 well-coordinated people is not 50% better than 2 people, since numerical superiority could result in a binary swing in outcomes (e.g. winning instead of losing some engagement or objective).