A reader asked for advice about how to run a league for Padel – a racquet sport that’s somewhat akin to tennis or squash. He expects to have about 30 teams, and wants to play the league in three months, at one match per week for each player.
First, a side observation about the request. I find that I’m more likely to be asked about games and sports I haven’t heard of than about the ones that I’m already familiar with. Perhaps that’s because the well-established sports also have established traditions for how their tourneys are run. It’s the new sports, or at least the less common ones, for which organizers seek the help of the likes of tourneygeek. And so it’s these less common sports that are likely to generate interesting new ideas about how tourneys should be run.
The reader suggested a creative format: Start with a group stage, with groups of four. Then use the group stage to seed a double-elimination bracket thus: group winners go to round 2 of the upper bracket; group runners-up go to round 1 of the upper bracket; third-place teams go to round 2 of the lower bracket; and fourth-place teams go to round 1 of the lower bracket. Will this work? Is there a better way?
The groups-into-double-elimination idea is an interesting one, but it’s not going to work without some adjustments. Let’s say there are 8 groups. If the 8 runners-up go to the first round of the upper bracket, there will be 4 round 1 winners. But then if the 8 group winners get seeded into round two, that round has 12 teams. And, with no more teams dropping into that bracket, you’re going to have to have some late-round byes to get the bracket back to an even power of two.
So, while I generally approve of the idea of seeding a bracket with some earned byes from previous play, it’s not going to work here without some fairly significant changes.
What you can do is seed the group winners and the group runners-up into the first round of an upper bracket, and the third- and fourth-place teams into the first round of a lower bracket. The upper bracket is, then, a conventional 16 bracket. The lower bracket looks like this: Padel lower
I like that bracket better as a consolation bracket rather than a full double elimination bracket because of the extreme separation of adjacent drops from the end of the upper bracket. The loser of the upper final drops to round 7 of the lower bracket, while the two losers of the upper semi-final drop to round 4. If the loser of the upper final doesn’t drop, then that’s avoided.
So, there’s a way to implement at least a modified form of my reader’s suggestion. But is there a better way?
The suggestion yields a short round-robin group stage that seeds a long bracketed playoff. The standard approach to this sort of league problem is just the reverse: a long round-robin group stage leading to a short playoff.
So, with 32 teams and 12 play dates (or rounds), I’d prefer to do this: draw the teams into four groups of six. Then play a double round robin within each group for the first 10 play dates. After that, have the four group winners play off for the league championship.
That’s going to lead to more participation – more rounds of play for the teams that don’t do very well. In the more conventional league, each team gets at least 10 matches, whereas in the other league some teams get as few as four. If this were a professional league, where spectacle is valued more highly than participation, you might prefer to bring on the high-stakes elimination matches sooner. But I’m assuming that this Padel league is being played more for the participants than the spectators.
Can one do better than this? I think so, though it would be at the cost of a good deal of additional complexity. Instead of running four groups of six in a double round-robin, run a 10-round Swiss for the entire group. That way, each team would play the same number of rounds, but they’d meet more different opponents, and those opponents would, on average, then to be of more equal skill. To dress the tourney up a little further, I think I’d take the top four finishers from the Swiss into a four-team page playoff rather than a simple knockout.
tourneygeek.com 
Oh hey, new post.
As an organizer, I come partially from an esports background, so I’m somewhat used to seeing weird groups-into-double-elimination formats. The reader’s proposed format (if tweaked to actually work) is downright tame when compared to some things I’ve seen in the past.
As for the “long groups vs. long playoff” argument, that’s purely down to a question of how much participation is valued against fairness. As this seems to be a fairly casual league, I’d definitely choose the side that gives the competitors as many games as possible – the Swiss.
LikeLike
“So, with 32 teams and 12 play dates (or rounds), I’d prefer to do this: draw the teams into four groups of six.”
So… what happens to the (up to) 8 players that aren’t drawn?
Let’s start with 4 pools of 8 in a single Round-Robin. This yields 7 weeks of competition. Since Padel appears to be an outdoor sport, I’d want to allow 2 “rain” weeks for the full season, which results in 3 weeks for the final competition. Which, unfortunately, means that only the Top 2 in each pool would advance to the Main Tourney, the next Two would advance to a Consolation Tourney, and 5th thru 8th would be relegated to Rounds-Robin with the other players finishing at the same rank. The 2 tourneys would have “consolation” and “placement” games, such that each player would have 3 games during the tournament period.
However, if no rain dates were used during pool play, then the tourney could be expanded to 4 rounds (still reserving 1 rain date). While I would keep the main Tourney to 3 rounds, the extra round would be used to drop the losers of the Quarter Finals into the 3rd-4th consolation in its 2nd round. If Padel is an indoor sport, then the number of “Rain Dates” required would be reduced to 1, allowing the Consolation to be expanded from the start.
The 4 pools of 8 setup should “pay” 5 “winners”. If the Consolation round was expanded, then all Proper Semifinalists and the Consolation Winner would get “Paid”. If not, then the 2 Finalists, the winner of the 3rd place match, the winner of the “5th place” match, and the Consolation Winner would all get “paid”, though if the 2nd rain date wasn’t used, I’d allow for the loser of the 3rd place match to challenge the winner of the 5th place match for the latter’s prize if the match wasn’t already done.
As James suggests, an 8-round Swiss followed by a short Tourney would probably be better than that. In fact, a Swiss Tourney where there are more rounds available than the log-base-2 (rounded up) of the number of entrants is extremely flexible, as the number of “rain dates” allowable is the difference between the number of Swiss rounds scheduled and that log-base-2:entrant result. (Again, I’d still allow 1 “Rain Date” for the Final Tourney.) So I’d have 8 Swiss Rounds SCHEDULED, (with the understanding that up to 2 Swiss Rounds could be dropped if site/staff-complications arose,) followed by a 4-team Page-McIntyre. (If this pays more than 2, then more pairs could be added: 5th & 6th Swiss if 3-5 pays, best these 2 can do is 3rd, 7th & 8th Swiss if 6-7 pats, best these 2 can do is 5th, etc.)
But let’s go back to the OP’s initial suggestion: 8 pools of 4 (3 rounds). With 2 “Rain Dates”, that leaves 7 rounds. As a side note, OP’s suggestion is quite like how the World Cup is, at least up to 2022: a 3 round Group Stage followed by a 4 round Tourney, with 16 teams eliminated (OP puts these into a Consolation). (In 2026, the Group Stage is still technically 3 rounds but each team gets a Bye, followed by a 5 round Tourney, again with 16 teams eliminated)
Here is where I would make things interesting: I would add a 2nd Group Stage based on the results of the first. Let’s call the original pools A, B, C, D, E, F, G, & H. I’d put A1, B1, C2, and D2 in the same pool (call it I), and C1, D1, A2, and B2 in another pool (call if J). Do the same for pools E-H (call them K & L) and also do the same for the 3rd and 4th places in each pool (M, N, O, & P). This takes 3 more rounds leaving 4 for the Tourney Proper.
As a side note, my original idea was to combine the 1st and 2nd in 4 pools into a single 8-team pool and the 3rd and 4th into another 8-team pool, with the match between each of the initial pool’s members standing, i.e. a “continuation Pool”. However, that amounted to 6 rounds, leaving only 1 or 2 rounds for the Tourney Proper, meaning that the initial 2nd places didn’t have any real hope of playing in the Final. Splitting up the Continuation Pool as above cut the number of rounds in half, allowing for a proper Final:
The players of the 4 “1-2” pools (I, J, K, & L) with the best TOTAL record (from both group stages) would then get byes into the quarterfinals of the main tourney, as per OP’s suggestion. The players from those 4 pools with the 2nd best total record and the players from pools M, N, O, & P with the best overall records would enter the 1st round of the main tourney. 3rd place from I, J, K, & L and 2nd Place from M, N, O, & P would all be inserted into a Consolation Tourney with the losers from the 1st Round of the Tourney Proper dropping in. If no rain dates were used in the Group Stages, then the losers of the Quarterfinals would also drop; otherwise they’d play their own (“5th Place”) Consolation. The “pay” structure would be similar to the 4×8 case. The 3rds from M, N, O, and P and the 4ths from all pools would all play a 3-game Round Robin among their equivalent finishers.
Or, mix it up: 8 pools of 4, followed by a Swiss with the pool results programmed in, then a Page-McIntyre based on the overall record.
Byes would be inserted, in order, into pools H, D, F, B, G, E, C. Then again, if there are 26-30 players, then the 8-round Swiss would be preferable. If Swiss-plus-Page-McIntyre is too bold, then 5 pools of 6 should be employed, with a group stage of 5 weeks followed by a 5-round Tourney: Pool Winners in the quarters as per OP, 2nd placers & best 3rd in Round 1, and 9 remaining 3rd & 4th places in a Consolation (the best of these 9 gets a Bye) with losers of Round 1 & dropping to the consolation (again, at least 5 “payouts”).
LikeLike
Lots of ways to make this work, really interesting challenge.
One of the earliest tournaments I ran had 30 players, I did two pool phases into a short bracket:
– First phase: five RR pools of six, top two teams/pool advance (to different Second phase pools)
– Second phase: two RR pools of five, top two teams/pool advance (to different semifinals)
– Bracket: 4-team single-elim.
It’s not a sophisticated bracket or anything, but it would certainly work for this timeframe (need 12 weeks to do it), scales from 25-36 without any major changes, and no rematches possible until the semifinals. You would have to design consolation pools based on the actual size of the competition, but they can be smaller in size and have a different format.
On the flip side, the most recent league I finished was a little closer to the request: 10 teams, 7 weeks of league play followed by 4 weeks of bracket. I’ve noticed that 7 weeks is a pretty good length for partial-RR in what we do, and I have a 32DE that has exactly 10 teams remaining (2-3 winners/7-8 losers) with 4 rounds left to go. The two ideas worked well together.
Honestly, I might go with the latter…I think having players who start at the bottom still have a fighting chance to win the league/tournament is a worthwhile goal in casual/social competitions, but make them REALLY have to fight for it. Bad seeding, bad matchups, longer paths, etc…just try to avoid dead rubbers as much as possible. I don’t think many DE’s can be split in quarters that way, though…you might be looking at a partial-TE or giving byes to the top X of the winner’s side.
LikeLike