FRC Alliance Selection Strategy: How Picks, Scouting, and Playoffs Work

Alliance selection is where events are actually won

You can build a fast, reliable robot and still go home early if you misplay the three minutes that decide the playoffs: alliance selection. In FIRST Robotics Competition, no single team wins an event alone. The playoff bracket is contested by alliances of three robots, and those alliances are assembled in a live, on-field draft after qualification matches end. The team that scouts well, builds a smart pick list, and understands the draft math will consistently reach — and win — playoffs above what its own robot alone could earn.

This guide covers the whole picture: how qualification ranking decides who gets to pick, the exact mechanics of the serpentine draft (verified against the official Game Manual), how to think about your first versus second pick, the scouting inputs that feed a pick list, and how the double-elimination playoff format should shape every decision you make on selection day.

Everything here is written to survive a game change. Specific point values and ranking-point (RP) criteria change every season, so treat any numeric example as illustrative and always confirm the current year's details in the official manual.

The structure that makes selection matter

At essentially all official FRC events, the playoffs use an 8-alliance tournament. Smaller official events (24 teams or fewer) also form fewer than 8 alliances under the Game Manual's Small Event Exceptions, and off-season events sometimes run other formats too — but the 8-alliance bracket is the standard you should plan around.

Here is the chain of events on the final day of a competition:

1. Qualification matches finish. Every team has a ranking based on how it performed across its ~8–12 randomly assigned matches (played with and against everyone else).
2. The top 8 ranked teams become Alliance Leads (commonly called alliance captains). They are designated Alliance 1 through Alliance 8 in ranking order.
3. Each captain drafts two more teams, in a specific order, until eight alliances of three robots each are formed.
4. Those eight alliances play a double-elimination bracket to decide the event champion.

So your qualification performance does two different jobs. If you finish in the top 8, it determines whether you pick or get picked and how early. If you finish outside the top 8, your qualification record and — far more importantly — your demonstrated robot capability determine whether a captain wants you.

How qualification ranking decides who captains

Teams are ranked by a season-long metric usually called Ranking Score — your average ranking points per match. Averaging (rather than summing) matters because it keeps rankings fair when teams play an unequal number of matches or have a "surrogate" match that doesn't count.

Ranking points themselves are awarded by game-specific objectives. Across recent seasons the pattern has been stable even as the details change:

• Winning a qualification match is worth the most RP (a win has been worth 3 RP, a tie 1, a loss 0 in recent games).
• Two or three "bonus" RP are available for hitting game-specific thresholds (scoring a certain amount in a particular way, completing an endgame task, or a cooperation objective with the other alliance).

When two teams have the same Ranking Score, a defined list of tiebreakers breaks the tie in order. The exact criteria are game-specific, but as a concrete example, the 2024 season (CRESCENDO) sorted tied teams by, in order: average cooperation-bonus points, then average alliance match points excluding fouls, then average auto/leave points, then average endgame points, and finally a random FMS sort. The takeaway that carries year to year: rankings reward consistent, foul-free scoring, and the deep tiebreakers mean a couple of clean, high-scoring matches can move you up a seed.

Why should a team that only cares about getting picked bother understanding this? Because captains read rankings the same way you do — they know that a team ranked 11th with a strong robot that got unlucky partners is often a better pick than a team ranked 6th that ranked high on tiebreakers. Rankings start the conversation; scouting data finishes it.

The alliance selection ceremony, step by step

Selection happens live on the field. Each of the eight captains sends one student representative to the field to make invitations out loud, one at a time. Invited teams also send a student rep to accept or decline. The whole thing is fast and public, which is exactly why you must walk in with decisions already made.

The draft runs as a serpentine (a.k.a. snake) draft, verified against the Game Manual's tournament section:

• Round 1: Alliance 1 invites first, then Alliance 2, and so on down to Alliance 8. Each captain adds one team.
• Round 2: The order reverses — Alliance 8 picks first, then 7, 6, down to Alliance 1 picking last.

After two rounds, you have eight alliances of three teams each. (The FIRST Championship divisions historically use a third selection round to build four-robot alliances, but at regionals and district events, three robots per alliance play and the serpentine runs two rounds.)

Two rules that quietly shape everything

A captain can pick another captain. The top 8 teams are not off-limits — Alliance 1 can invite the team currently sitting as Alliance 4. If that team accepts, something important happens: every lower captain moves up one spot, and the highest-ranked team not yet on an alliance is promoted into the Alliance 8 captain slot. This is why a strong team ranked 9th–12th can still end up captaining an alliance even though they started outside the top 8.

A team that declines is (mostly) out. Per the manual (rule T606 in recent years), an alliance captain may not invite a team that has already declined another alliance's invitation. Practically: if you are an alliance captain and you decline an invitation to join a higher alliance, you keep your own captaincy — but if you are not a captain and you decline, you have taken yourself out of the draft entirely and will not play in the playoffs. Declining is therefore a serious, rarely correct move for a non-captain. It only makes sense when a team is confident it will be a captain itself, or genuinely cannot play.

Because accepting-up cascades the captain list, the person representing you on the field must be empowered to make real-time decisions, not just read a script. If your #1 target gets invited by a higher alliance before your turn, you need your rep to jump instantly to the next name on the list.

First-pick strategy: raise the ceiling

Your first pick is the most consequential decision a captain makes. The governing idea:

A first pick should raise your alliance's ceiling — the best possible outcome — not just patch a weakness.

In most games, that means selecting the best available robot that also complements yours. Two robots that both do the same primary scoring task well can still be a fantastic pairing if the game rewards raw output. But the strongest first picks are usually the ones that let your alliance score in more ways or faster than your opponents — a partner that unlocks a bonus RP condition, covers a scoring location you can't reach, or simply out-scores everyone left on the board.

Concrete first-pick priorities that generalize across games:

• Scoring ceiling and speed — how many points per match can this robot add, and how quickly does it cycle?
• Complementary capability — does it do something your robot can't, or free your robot to specialize?
• Autonomous strength — a reliable, high-value auto routine that doesn't collide with yours is worth a lot, because auto points are uncontested.
• Reliability under pressure — a slightly lower-scoring robot that never breaks often beats a higher-scoring one that dies every third match.

The classic mistake is picking the highest-ranked available team instead of the highest-capability available team. Ranking is polluted by partner luck; your scouting data isn't.

Second-pick strategy: fit the role, not the ceiling

By the time your second pick comes around (in the reversed round), the elite robots are gone. The second pick is about completing a functioning three-robot machine, and the right choice depends entirely on what your alliance's first two robots already do.

Common second-pick roles:

• The specialist / role player — a robot that does one narrow but valuable job extremely well (a dedicated feeder, a fast cycler on a single scoring element, an auto specialist).
• The defender — in games where defense is legal and impactful, a maneuverable, sturdy robot whose job is to disrupt the opposing alliance's best scorer. A good defensive second pick can neutralize an opponent's first pick, effectively winning you a match without scoring a point.
• The reliable floor-raiser — a robot that isn't flashy but consistently does its job, so your alliance's worst case stays high.

This is the heart of the "second-robot theory": the value of a second pick is measured by how well it fits the gap left by the captain and first pick, not by its standalone scoring. A mediocre scorer that plays disciplined defense or feeds your cyclers perfectly can be worth more to your specific alliance than a better all-around robot that duplicates what you already have.

If you're on the bubble: make yourself pickable

Most teams will be picked, not picking. The goal for those teams is to be the obvious next name on as many pick lists as possible. You influence that far more than you might think:

• Be reliable. Captains fear robots that die. A team that scores modestly but never breaks down or gets called for fouls is a safer, more attractive pick than a boom-or-bust robot.
• Do one thing conspicuously well. A clear specialty — great defense, a fast single-task cycle, a strong auto — gets you onto lists as a targeted role player.
• Play clean. Foul-heavy robots hurt their alliance and show up in scouting notes immediately.
• Talk to scouts. Alliance-selection conversations happen in the pits before the ceremony. If a captain's scouts stop by, be honest and specific about what your robot can and can't do. Overselling gets you crossed off; a crisp, accurate pitch gets you picked.
• Have a driver who executes. Two identical robots are separated by driver skill. Consistent, calm driving is visible on the field and is a real tiebreaker for captains.

The scouting inputs behind a real pick list

A pick list is only as good as the data behind it. Serious teams combine quantitative and qualitative inputs.

Quantitative — the numbers:

• Match scouting data: your own team's recorded per-match performance for every robot — points scored by category, cycle counts, auto success, endgame success, penalties, and reliability (did it move the whole match?). This is the backbone; nothing replaces watching every robot every match.
• Calculated metrics (OPR, DPR, CCWM): Offensive Power Rating estimates a robot's average point contribution; Defensive Power Rating estimates points it suppresses; Calculated Contribution to Winning Margin combines both. These are statistical estimates derived from alliance scores and are widely published on The Blue Alliance and Statbotics. Treat them as a sanity check and a way to spot robots you under-scouted — not as gospel, because they can be skewed by partners and schedule.
• Rankings and RP breakdowns: useful context, but always cross-checked against capability.

Qualitative — the judgment:

• Pit scouting: talking to teams about their robot's design, capabilities, and known issues. Critical for finding robots whose match data undersells them (e.g., a great robot stuck with weak partners).
• Reliability and robustness notes: did it tip, brown out, or lose a mechanism? Written observations from your scouts matter as much as any number.
• Driver skill and driveteam composure: hard to quantify, easy to see.
• Do-not-pick reasons: honest notes on teams to avoid (chronic fouls, unreliability, uncooperative driveteam).

Turning inputs into a board. Most competitive teams build a tiered, ranked pick list: robots grouped into tiers (e.g., "elite first-pick," "strong role players," "defense," "reliable depth") and ranked within each tier, plus an explicit do-not-pick list. During the ceremony, you cross off names as they're taken and always invite the top remaining team that fits your current need. The tiering matters because the serpentine draft moves fast and targets vanish — you need the next name ready, not a re-debate.

The playoff format your picks are feeding into

Since 2023, FRC events use a double-elimination bracket for the eight alliances. Understanding it changes how you value robots on selection day.

How it works (verified against the Game Manual):

• Every alliance starts in the Upper bracket. Win an upper-bracket match and you stay upper; lose and you drop to the Lower bracket.
• In the Lower bracket, one more loss eliminates you. So an alliance is out of the tournament after two losses total.
• The full bracket is 13 matches plus the finals.
• The finals are a best-of-three: the first alliance to win 2 matches becomes the event champion. (If a finals match ties, it doesn't count as a win for either alliance, and additional matches are played until one alliance reaches two wins.)

Two strategic consequences for selection:

1. Depth and reliability beat single-match brilliance. Because a run to the championship can require winning five or more matches across a couple of hours, robots that break down or foul are far more dangerous partners than they look in a single quals match. A double-elimination bracket punishes fragility twice over.
2. You get a second chance, so ceiling still matters. Losing once doesn't end you. That slightly de-risks an aggressive first pick — a high-ceiling robot that occasionally stumbles can lose in the upper bracket and still fight back through the lower bracket. Balance this against point 1.

The backup team rule. During the playoffs, an alliance may invoke a one-time option to replace one of its robots with a backup team — the highest-ranked team not already on an alliance. The alliance is then recorded as four teams, but still only three robots play per match. In recent manuals (rule T608), an alliance cannot request a backup until after it has played its first playoff match, and the alliance may then field any 3 of its 4 teams' robots in subsequent lineups, including the originally replaced one. The FIRST Championship's Einstein round has historically had no backup-team provision. The practical implication: your third robot going down isn't automatically fatal, but the replacement pool is thin and the timing is restricted — so reliability at selection time is still your best insurance.

Common alliance-selection mistakes

• Picking by ranking instead of capability. Rankings are distorted by partner luck. Your scouting data isn't.
• Not scouting your own potential partners. Teams obsess over opponents and forget they might pick — or be picked by — half the field.
• No do-not-pick list. Inviting a robot with a hidden reliability or foul problem can sink your whole playoff run.
• A rep who can't improvise. If your on-field student can only read a fixed script, one early surprise pick derails your entire plan. Empower them and give them a tiered board.
• Duplicating your own robot. Two robots that do the identical primary task and nothing else leave obvious gaps for opponents to exploit — especially on the defensive second pick.
• Ignoring the playoff format. Optimizing for one perfect match instead of a reliable five-match run is how strong alliances lose to steadier ones.

Selection-day checklist

• Finalize a tiered, ranked pick list plus an explicit do-not-pick list before the ceremony.
• Confirm the current season's RP and tiebreaker rules so you read rankings correctly.
• Know the serpentine order and rehearse the "if my target is taken, who's next" flow.
• Choose a student rep who can make live decisions, and brief them on the promotion/decline rules.
• Weight reliability and role-fit appropriately — you're building for a multi-match double-elimination run, not a single highlight.
• Have your first pick raise the ceiling and your second pick fill the gap.

Alliance selection rewards preparation more than almost any other three minutes in FRC. The robots are already built; on that final afternoon, the teams that win are the ones who scouted honestly, understood the draft math, and walked onto the field knowing exactly whom they wanted and why.

Always verify current-season specifics — ranking-point criteria, tiebreakers, alliance size, and playoff details — in the official FIRST Robotics Competition Game Manual for your year, as these change season to season.