How do I manage time for a double-elimination bracket? (TOing)

[Juggleguy]

As the Smash community continues to grow at a record rate, many TOs have been caught biting off more than they can chew, often overloading their own tournaments with multiple games and events without the appropriate preparations. This creates a logistical mess and an unpleasant experience for attendees. Before even thinking about hosting a multi-game tournament, TOs should learn how to manage time for a one-game tournament: these are simple and mostly boil down to analyzing the time required to complete each isolated event.

In the first of this two-part article, I'll discuss how to manage time specifically for a double-elimination bracket. This is the most common format at Smash tournaments, or at least the most common endgame format when preceded by round robin pools (we'll talk about how to estimate time required for round robin pools in the next article). A double-elimination bracket offers two chances for players to lose before being eliminated; while this sounds great taken at face value, the logistics must be tackled with efficiency in order for the tournament to run on time.

Think about some of the following expected values for your double-elimination bracket. The answers to these questions will serve as the parameters for your time estimate.

1) Entrants: How many entrants do I expect to participate?
2) Setups: How many full setups do I expect to be available?
3) Set Length: How many minutes do I expect the average set to require?

--

Here's an example.

1) Entrants: 16 entrants for singles
2) Setups: 16 setups available throughout
3) Set Length: 10 minutes per set

Our parameters are as follows:
16 entrants
16 setups
10 minutes per set

A double-elimination bracket can be deceptive in that the time required to complete it is fairly fixed. It doesn't matter if you have a million setups or the best TOs in the world running things; the queueing nature of the double-elimination bracket lends itself to a minimum running time that plateaus after you hit a "critical value" of setup resources. If you're unfamiliar with basic queueing theory, look closely at a double-elimination bracket and ask yourself, "can this losers bracket set happen before this winners bracket set happens?" for several different parts of the bracket.

This example is pretty much as simple as it gets. Note that the number of sets available to play at any given time does not exceed the number of setups, so you can assume all sets in each round can be played concurrently. In fact, there are at least 8 unused setups at any given time in this case. Also, the total number of players involved is very low, which offers a relatively easy time managing brackets and downtime; you can probably get away with assuming a number as low as 10 minutes per set here, if this is a typical local with typical Smashers and typical TO staff manpower.

Break it down by round:

Round | # of Concurrent Sets | Time Required
WB1 | 8 sets | 10 minutes
WQ + LB1 | 8 sets | 10 minutes
WS + LB2 | 6 sets | 10 minutes
WF + LB3 | 3 sets | 10 minutes
LQ | 2 sets | 10 minutes
LS | 1 set | 10 minutes
LF | 1 set | 10 minutes
GF first set | 1 set | 10 minutes
GF second set | 1 set | 10 minutes

Assuming our parameters of 16 entrants, 16 available setups, and 10 minutes per set, the estimated time required for this event is 90 minutes, or 1 hour 30 minutes.

--

Let's do another example.

1) Entrants: 64 teams for doubles
2) Setups: 16 setups available throughout
3) Set Length: 15 minutes per set

Our parameters are as follows:
64 entrants
16 setups
15 minutes per set

Things get trickier this time.

The first thing you should notice is the number of available sets to play exceeds the number of available setups during the early stages of the bracket. Because of this, you have to break down the overloaded round into smaller chunks and treat each as a separate round. In this case, break down WB1 into two halves as indicated below.

The second thing you should notice is that the sheer volume of players compared to the previous example will almost inevitably create longer sets. Not only are there 64 entrants, but those 64 entrants represent 128 total people because it's a doubles event. That's 8x as many as the previous example. More time will inevitably be spent finding players, reporting results, counterpicking stages, and so on. As you host more tournaments, you'll find that adjustments must always be made to scale to your turnout. In this case, I've upped the set length parameter to 15 minutes per set to account for the increase in moving parts.

Again, break it down by round:

Round | # of Concurrent Sets | Time Required
WB1 first half | 16 sets | 15 minutes
WB1 second half | 16 sets | 15 minutes
WB2 | 16 sets | 15 minutes
LB1 | 16 sets | 15 minutes
LB2 | 16 sets | 15 minutes
WB3 + LB3 | 16 sets | 15 minutes
WQ + LB4 | 12 sets | 15 minutes
WS + LB5 | 6 sets | 15 minutes
WF + LB6 | 5 sets | 15 minutes
LB7 | 2 sets | 15 minutes
LQ | 2 sets | 15 minutes
LS | 1 set | 15 minutes
LF | 1 set | 15 minutes
GF first set | 1 set | 15 minutes
GF second set | 1 set | 15 minutes

Assuming our parameters of 64 entrants, 16 available setups, and 15 minutes per set, the estimated time required for this event is 225 minutes, or 3 hours 45 minutes.

--

Food for thought:

* Of course, there are ways to estimate running time in a quick-and-dirty fashion, aka a formula that uses the three mentioned parameters as an input and produces the estimated running time as an output. I've purposely excluded any kind of formula like that here because I feel that there are too many intangibles within a double-elimination bracket to take a generic shortcut like that. Every Smash tourney is different, whether it's the venue or geographic region or composition of turnout, and as a TO you should account for all those factors independently.

* Another reason against a formula: other considerations definitely come into play with respect to time management. You might want to record all of the late winners bracket on a single streaming setup; if so, then it's not necessary to play late winners bracket rounds concurrently as illustrated above. You might want to upgrade the final few sets to a best-of-5; if so, then increase the set length of later rounds accordingly. This article is intended to provide a foundation where you can make those adjustments yourself.

* All the above calculations assume zero downtime between sets. This is obviously unrealistic when it comes to Smash tournaments, which are often filled with players who don't report to their matches on time or take unpredictable breaks between sets. So before proceeding, it's advisable to add a few minutes to your set length parameter to account for the overall timeliness of your attendees and any outside circumstances surrounding your tournament.

* Setups are everything. In the above example, if the number of available setups was reduced from 16 setups to just 8 setups, the bracket would require an extra 105 minutes (7 extra rounds * 15 minutes per round) to complete given the same assumptions/parameters. Hopefully this gives you a better idea about why TOs are always so adamant about attendees bringing setups; every one counts.

* Well, every setup counts up until you reach the "critical value," or the point at which extra setups no longer help the bracket run faster. This is typically half the number of entrants in the bracket. For example, in a top-32 bracket, 16 setups is the critical value; it accomplishes maximum efficiency because the number of available sets to play at any given time will never exceed it. This number should be your setups goal during the planning phases of your tournament.

* A lot of variance is involved with set length, or average time required per set. In my experience, and for the sake of simplifying things for Smash tournaments, it's fair to assume a typical bell curve distribution for set lengths. In other words, unusually long sets will be balanced out by unusually short sets at the end of the day. Focus on obtaining extra setups to reduce the effect of bottlenecks.

* It should come as no surprise that the number one tournament-killer is downtime between sets. This is usually what separates tournaments that run on time from those that run behind schedule. If you have the staff and resources available, focus on reducing downtime to as close to zero as possible. After a tournament begins, the setup count and attendees count is pretty much out of your control, but you have plenty of control over how to reduce downtime.

In the second of this two-part article, I'll discuss how to manage time for round robin pools.

--

Juggleguy is a national tournament organizer, Melee It On Me team member, and Smashboards contributor. You can follow him on Twitter: @JuggleRob

 

[TechMage299]

This is some great stuff.

 

[KrymsonSkyz]

Informative as ****

Kill 'em TO's

 

[the_irish_dude]

Good job breaking things down for people.

 

[Wasp]

if only the TO's themself will read this... would be very helpfull

 

[Morbi]

I suppose this article was necessary after MLG Anaheim and CEO, huh?

 

[SphericalCrusher]

Awesome write up! Thank you very much. I'm actually hosting a 30-40 player tournament (my first big SSB tourney) this Saturday, so this article came right on time!

 

[Jaxel]

Set lengths should always be at 15 minutes. Even if the match only takes 10 minutes, you need to account for player setup (hand warming), as well as TOs having to run around and organize the players.

 

[Jaxel]

Also, keep in mind that running a tournament isn't just about setting up a proper schedule in advance, its also about having the right people on the floor at the time of the actual tournament to make sure everything runs smoothly.

Below is a screenshot of a small part of the schedule I put together for APEX 2013 (I had nothing to do with the schedule for APEX 2014, I won't take the blame for that disaster). APEX 2013 was ran fantastically, and in fact, it was way ahead of schedule. This probably in part due to it being my first Smash Bros tournament I ever scheduled, and I accounted for 20 minutes per match, when I probably could have gotten away with 15... but I erred on the side of caution.

You'll notice that we had every single pool, every single match scheduled down to the minute. We also had a giant clock in the main hall so that everyone in the venue could see the current time and DQs would be managed fairly. Everyone knew what time their matches would be and had no one else to blame but themselves if they got DQ'd.

But even with a well-though-out schedule; its irrelevant if the people running the tournament that day aren't on the ball. Maria and her team did a fantastic job on the floor making sure all the TOs were moving according to the schedule. If people had to be DQ'd, they showed no favoritism towards top players.

The major stall on most major tournaments and the prime reason they run late is because the TO team on the floor can't handle it. I actually did a pretty long video a while back about this:

 

[Juggleguy]

Set lengths should always be at 15 minutes. Even if the match only takes 10 minutes, you need to account for player setup (hand warming), as well as TOs having to run around and organize the players.

It depends on a variety of factors. First few that come to mind are: size of tourney, game being played, quality of TO staff.

I typically go 10 minutes at my locals because the tourney is of a very manageable size, it's usually the relatively fast-paced Melee, and my TO staff support is quite good. But obviously I wouldn't necessarily recommend 10 minutes to everyone else... it should be adjusted based on the circumstances.

 

[Skull_Kid]

I find this an excellent read, may i translate it and use it for Smash Bros italia? I'll obviously quote the OP and SB.

 

[Grime.]

This is a very informative post, thanks!

 

[Juggleguy]

I find this an excellent read, may i translate it and use it for Smash Bros italia? I'll obviously quote the OP and SB.

Sure, no problem.

 

[0RLY]

I was really hoping for a formula or equation we could use. Doing this by hand is dreadful

 

[~Firefly~]

One minor correction: In your second example, you have LQ listed as 1 set, when it should be 2 sets. It doesn't actually add to the running time though, so it doesn't really matter too much.

Great article.

 

[Juggleguy]

I was really hoping for a formula or equation we could use. Doing this by hand is dreadful

I understand. But I excluded a formula intentionally. Just addressed this in the food for thought section:

* Of course, there are ways to estimate running time in a quick-and-dirty fashion, aka a formula that uses the three mentioned parameters as an input and produces the estimated running time as an output. I've purposely excluded any kind of formula like that here because I feel that there are too many intangibles within a double-elimination bracket to take a generic shortcut like that. Every Smash tourney is different, whether it's the venue or geographic region or composition of turnout, and as a TO you should account for all those factors independently.

 

[0RLY]

I understand. But I excluded a formula intentionally. Just addressed this in the food for thought section:

* Of course, there are ways to estimate running time in a quick-and-dirty fashion, aka a formula that uses the three mentioned parameters as an input and produces the estimated running time as an output. I've purposely excluded any kind of formula like that here because I feel that there are too many intangibles within a double-elimination bracket to take a generic shortcut like that. Every Smash tourney is different, whether it's the venue or geographic region or composition of turnout, and as a TO you should account for all those factors independently.

Of course I would plan an extra 30 minutes or an hour in addition to these variables. However, having such a formula would be great for a minimum time estimate. If it's an overestimate, then it's likely that my prediction for the set length was overestimated. Which isn't even a problem! Extra time leftover can still be used for friendlies/MMs and clean up.

 

[TOGOpuff]

Thanks for this Juggleguy. I am starting to TO in my region and it's been a pain because we are really low on setups (specially tvs) and we're trying to start a inclusive scenario (not melee only). So having 3 tvs and one game per tv with 16 entrants each game is absurd to handle. We thought on doing 16 entrants each game 3 setups per game, each game at a time, which would lower each game time a lot but it makes as casual show-ups never look upon the games they're not inscribed (a melee player will leave the venue when melee's over) and we don't know how to try and deal with this yet.
Another big problem we're facing is lack of controllers. Most people interested in showing up doesn't have their own controllers and it's hard to find players with a spare one or a old usable one they're willing to lend (which is completely understandable imo considering its really hard to buy a controller in brazil).

 

[MrJeff]

Can't thank you enough for this. I am in the early planning stages of a tournament, and this was something I was worried about. I have been to a tournament where time management was something of an issue, so this was on my mind.

 

[Jaxel]

I was really hoping for a formula or equation we could use. Doing this by hand is dreadful

2n-1

This is the formula for the number of matches in a double elimination tournament. With a 32 person tournament, you need 2(32)-1 matches to run the tournament... so 63.

 

[0RLY]

2n-1

This is the formula for the number of matches in a double elimination tournament. With a 32 person tournament, you need 2(32)-1 matches to run the tournament... so 63.

Do I then divide the result by the number of setups and multiply by set length?

 

[Jaxel]

Do I then divide the result by the number of setups and multiply by set length?

Possibly... unfortunately, its not that simple... because there is a point in the tournament where you can no longer use every setup because you dont have enough matches.

For instance, during the last 2 rounds of losers, you can only run on 1 station at a time. During the last 3-4 rounds of losers, you can only run on 2 stations at a time. During the last 5-6 rounds of losers, you can only run on 4 stations at a time... and so forth, in that manner.

 

[Shoto]

Thanks for this Juggleguy. I am starting to TO in my region and it's been a pain because we are really low on setups (specially tvs) and we're trying to start a inclusive scenario (not melee only). So having 3 tvs and one game per tv with 16 entrants each game is absurd to handle. We thought on doing 16 entrants each game 3 setups per game, each game at a time, which would lower each game time a lot but it makes as casual show-ups never look upon the games they're not inscribed (a melee player will leave the venue when melee's over) and we don't know how to try and deal with this yet.
Another big problem we're facing is lack of controllers. Most people interested in showing up doesn't have their own controllers and it's hard to find players with a spare one or a old usable one they're willing to lend (which is completely understandable imo considering its really hard to buy a controller in brazil).

I've been on Craigslist, and sometimes you can find people selling wii's for 30-40 bucks. Goodwill and other thrift stores sell TV's for about 5-10, so check that out man. Maybe at the next tournament ask everyone to pitch in some money so you can acquire another setup at your next event.

 

[TOGOpuff]

I've been on Craigslist, and sometimes you can find people selling wii's for 30-40 bucks. Goodwill and other thrift stores sell TV's for about 5-10, so check that out man. Maybe at the next tournament ask everyone to pitch in some money so you can acquire another setup at your next event.

this is i'm from brazil. So no used tv under $50. People are greedy. My wii costed $200 used lol

 

[Shoto]

this is i'm from brazil. So no used tv under $50. People are greedy. My wii costed $200 used lol

Holy ****, get a ticket to the US, and buy them here hahaha

 

[Atalaya]

Can someone please help me with something, I'm running a bit of a grand scale tournament (about 96 players) and our plan is to cut the first 64 by round robin, with 12 pools of 8, I will have 16 set ups available, planning to use them all simultaneously and with as much efficiency as possible. I did the calculations with a formula a user posted on Reddit, and with 16 setups, i could run 4 pools at the same time (4 set ups for each pool of 8, as 1 setup=2 people. 2x4=8) Anyway, it gave me almost 5 hours to run all pools and narrow it down to 32 people.

The remaining 32 I plan to end the tournament by double elimination, counting with 16 setups, the entire 32 could play at the same time? And according to this, 16 setups with 16 entrants takes an hour and a half? I'm confused because I understand 1 setup=2 people, so I'm assuming they're using 8 setups, and leaving 8 setups unused. And that would take 1 hour and 30 minutes?

What about 32 participants, with 16 setups, all running with a slight inefficiency factor, lets say 15 mins per set. That would also take me 1 hour and a half? Thanks

 

[Atalaya]

Possibly... unfortunately, its not that simple... because there is a point in the tournament where you can no longer use every setup because you dont have enough matches.

For instance, during the last 2 rounds of losers, you can only run on 1 station at a time. During the last 3-4 rounds of losers, you can only run on 2 stations at a time. During the last 5-6 rounds of losers, you can only run on 4 stations at a time... and so forth, in that manner.

I'm sorry i don't understand, why is it that i can't run on all my stations at the same time? I will have 16 setups for a 32 man tournament with double elimination, and i hope to conclude it in maximum, and hour and a half. Why can't I use all my setups, not enough "matches" available? I don't understand that concept. Can't I do something like:

all 32 on 16 setups, narrowed down to 16 winners and 16 losers

1 setup=two people, so:

8 setups for the 16 winners
8 setups for the 16 losers

And it keeps going, assuming every set lasts 15 minutes, with a slight inefficiency factor included. I'm not sure if you get the idea, but can you shed some light on the matter? Remember:

32 participants
16 setups

Can it be done in under an hour and a half? If all setups are used consistently and simultaneously?

 

[Jaxel]

I'm sorry i don't understand, why is it that i can't run on all my stations at the same time? I will have 16 setups for a 32 man tournament with double elimination, and i hope to conclude it in maximum, and hour and a half. Why can't I use all my setups, not enough "matches" available? I don't understand that concept. Can't I do something like:

all 32 on 16 setups, narrowed down to 16 winners and 16 losers

1 setup=two people, so:

8 setups for the 16 winners
8 setups for the 16 losers

And it keeps going, assuming every set lasts 15 minutes, with a slight inefficiency factor included. I'm not sure if you get the idea, but can you shed some light on the matter? Remember:

32 participants
16 setups

Can it be done in under an hour and a half? If all setups are used consistently and simultaneously?

You CANT run every setup towards the end of the tournament. For instance, during losers finals, and grand finals, you literally can only run one setup at a time... because you dont have other matches to run.

 

[plalor]

Great post! Just wondering how you guys normally deal with 4 player game finals, 2 from winners bracket and 2 from losers bracket? When it's only two players as described in this post, one from winners and one from losers, there can be two final games, if the winner of the losers bracket wins the first game. Is there a similar way to play the finals of a 4 player game?