Quarters and semis and finals - oh my

[ProgressNotPerfection]

How do competitions decide how many couples to advance from quarters to semis and from semis to finals? It isn’t a consistent number; i.e., it’s not true that ALL final rounds consist of 6 couples, or that the next round will always have exactly half as many couples as danced in the previous round. So I’m wondering what factors go into making those decisions.

For example, the results from Manhattan Dance Championships 2025 showed these numbers:

L-B1 singles in closed silver smooth
Pre Silver - straight final with 8 couples
Intermediate Silver - semifinal with 13 couples; final with 8 couples
Full Silver - semifinal with 12 couples; final with 6 couples

Why did they recall 8 couples to the intermediate silver finals, but only 6 for the full silver finals - especially when the full silver semi had fewer couples to begin with? Why not recall the same number (whether 6, 7, or 8) in both levels?

L-C Championships - semifinal with 12 couples; final with 6 couples

L-C Scholarships - quarterfinal with 17 couples; semifinal with 10 couples; final with 6 couples

Why did they recall only 10 couples to the scholarship semifinal instead of 11 or 12?

Is it something that’s decided in advance (“these specific events will have X number of couples in the final rounds”) or is it determined based on what happens that day at the comp, maybe depending on how many couples actually show up (vs how many registered in advance), or how close the judges’ marks are (“these couples both received the same number of marks so we will recall both of them to the next round, even though that means we will have a larger-than-usual semifinal”), or what?

Note: I’m not criticizing any of the callback decisions, or suggesting that there are “correct” or “incorrect” heat sizes - I’m just genuinely curious about how it all works. Would love to hear your insights on it, especially those of you who have worked behind the scenes at NDCA comps.

 

[j_alexandra]

 

[Mengu]

Largely chairperson discretion, but there are some rules (and exceptions allowed). Largest final allowed is 8. Callbacks must be at least 50%. However, when numbers are close, exceptions are allowed so they are not creating an extra round for one or two couples. For instance, in the example you have, the 17 couple round could have been a semi, with 8 or 9 recalled, to eliminate a round. It's a discretionary decision at that point. It's risky to do it as the initial round, because based on judges results you may have to insert another round into the schedule, which messes things up a lot. So it's easier to plan for more rounds, and cut them if needed.

Again for your 17 couple quarterfinal example, I suspect the reason they recalled 10 is the natural break in judges scores. The scrutineer probably went to the chairperson with some numbers, like we can recall 15 or 10, which do you want? And 10 was picked. (I feel like this sort of question is the most common one I've heard.)

The target numbers are 24/12/6, but natural breaks in judges scores have to be taken into account. Scrutineers can usually make the decision that abides by the 50% rule, so they don't have to go to chairperson too frequently. But when the numbers get weird, then they need a decision.

 

[Mengu]

Also to add one other interesting tidbit, you've possibly seen events where they run a semi final, and then they have to rerun it, with maybe just 1 or 2 fewer couples. What happens there is they start with something like 13 couples, results come back, there is a clump of couples with the same number of recalls, and looking at the natural breaks, you have to callback either 12, or 5... Calling back 5 out of 13 is too far below the 50% threshold, so they will rerun the semi with 12. (It probably doesn't feel that great when you're the 13th couple, but that's competition.)

In the final there are a lot of ways to break ties, but in the prelims ties can happen a fair amount, which creates these unusual situations.

 

[Joe]

Some places (or Chairs) also rule that as long as a previous round had the correct number of couples, they can cut more than 50% later, e.g. 24-14-6 if that's the way the recalls broke.

 

[ProgressNotPerfection]

Thanks for all your helpful replies! It reinforced what I thought, which is that there are guidelines (called rules, but they have so many exceptions and so much discretion is allowed that they don't feel like rules to me), but ultimately it depends on what happens at that comp on that day for that event.

 

[Quickandslow]

And in single dances when there aren’t usually as many judges, it gets really tricky. I’ve had a semi kind of turn into a quarter because there were too many tied for recall so they had to cut a few and then do another round before the final.

 

[Arun Garg]

Can confirm from the backend, yes it varies depending on amount of couples entered and the desires of the Organizer and Chairperson. Most Organizers would rather a busy floor and have more happy clients so they are likely to (for Pro-Am) recall 7-8 if it is a large semi. If it is the smallest size semi (9 couples), they will probably call back 6, so that 1-2 people aren't being feeling 'singled out' in a pro-am event; apparently 3 couples cut is enough to make it seem reasonable. Same logic applies for quarters to semis. Another factor is how much time is available in the schedule (fewer rounds/heats means less time spent on a specific event). The other built in factors are tie breaking rules for tied amount of recalls where the Chairperson makes the call on how many get recalled into a final (see Skating System Rules and Mengu's post above). Also note: Pro-Pro events cannot have larger than a 7 couple final. If there are 8 couples, they will recall 6 from a semi final.

*GENERALLY* speaking, Chairpersons and Organizers will err on the side of generosity for Pro-Am, efficiency of schedule and logistics for Pro-Pro when it comes to recalls.

 

[Warren J. Dew]

How do competitions decide how many couples to advance from quarters to semis and from semis to finals? It isn’t a consistent number; i.e., it’s not true that ALL final rounds consist of 6 couples, or that the next round will always have exactly half as many couples as danced in the previous round. So I’m wondering what factors go into making those decisions.

6 is what they aim for in the final, but it may be larger, or very rarely smaller, for a variety of reasons.

Note that the difficulty of judging a final goes up considerably more than linearly with the number of couples. Other than that, the primary consideration for the chairman is keeping the competition running on time.

For example, the results from Manhattan Dance Championships 2025 showed these numbers:

L-B1 singles in closed silver smooth
Pre Silver - straight final with 8 couples
Intermediate Silver - semifinal with 13 couples; final with 8 couples
Full Silver - semifinal with 12 couples; final with 6 couples

Why did they recall 8 couples to the intermediate silver finals, but only 6 for the full silver finals - especially when the full silver semi had fewer couples to begin with? Why not recall the same number (whether 6, 7, or 8) in both levels?

A semi of 12 is ideal to be cut in half for a final of 6.

With a first round of 13 in the intermediate silver, they could cut it in half and round up to 7, or they could make it a quarterfinal and aim for 6. A quarterfinal would have taken too long, so they aimed for 7. There was a tie for 7th (or maybe a three way tie for 6th or some such), so they brought back 8.

L-C Championships - semifinal with 12 couples; final with 6 couples

L-C Scholarships - quarterfinal with 17 couples; semifinal with 10 couples; final with 6 couples

Why did they recall only 10 couples to the scholarship semifinal instead of 11 or 12?

It’s preferable, and based on what j_alex shows, now mandatory in the US, to bring back half or more. There is an outside chance that in the semifinal, there will be a four way tie for sixth, giving a choice between 5 or 9 for the next round. With a semifinal of 10, 5 is half and legal. With a semifinal of 11 or 12, you’d have to call back 9, which is really too big for a final, so you’d likely have to add a round, blowing the schedule. Since the quarter was only 17, a semi of 10 is comfortably more than half, so it isn’t too deep a cut.

There may also have been ties, so a semi of 11 or 12 might not have been choices.

Is it something that’s decided in advance (“these specific events will have X number of couples in the final rounds”) or is it determined based on what happens that day at the comp, maybe depending on how many couples actually show up (vs how many registered in advance), or how close the judges’ marks are (“these couples both received the same number of marks so we will recall both of them to the next round, even though that means we will have a larger-than-usual semifinal”), or what?

For cases like these, it’s decided on the day of the competition.

Note: I’m not criticizing any of the callback decisions, or suggesting that there are “correct” or “incorrect” heat sizes - I’m just genuinely curious about how it all works. Would love to hear your insights on it, especially those of you who have worked behind the scenes at NDCA comps.

Have not worked at an NDCA competition, but things are not much different between circuits, except with respect to competition size (collegiate competitions regularly have events with three digit numbers of entrants in a single event; this is extremely rare in NDCA competitions).

 

[Warren J. Dew]

The scrutineer probably went to the chairperson with some numbers, like we can recall 15 or 10, which do you want? And 10 was picked. (I feel like this sort of question is the most common one I've heard.)

There’s also another reason for this. The scrutineer actually knows which couples will be affected. The chairman doesn’t unless they ask, which they never do. This insulates the chairman from accusations of favoritism that might be leveled at the scrutineer if the scrutineer made the decision directly - not to mention the scrutineer may not even have judging qualifications. (Back in the days of manual scrutineering, the chairman had to be both a qualified judge and also a qualified scrutineer, and there weren’t that many people that were both; not sure what present requirements are.)

 

[Warren J. Dew]

Some places (or Chairs) also rule that as long as a previous round had the correct number of couples, they can cut more than 50% later, e.g. 24-14-6 if that's the way the recalls broke.

The NDCA rules used to allow this, by specifying the minimum number of rounds that had to be run as a function of number of competitors, rather than what fraction had to be recalled. I’m mildly surprised that has changed, though this comes up more often in very large events and so may be more relevant to college competitions.

 

[Joe]

There’s also another reason for this. The scrutineer actually knows which couples will be affected. The chairman doesn’t unless they ask, which they never do. This insulates the chairman from accusations of favoritism that might be leveled at the scrutineer if the scrutineer made the decision directly - not to mention the scrutineer may not even have judging qualifications. (Back in the days of manual scrutineering, the chairman had to be both a qualified judge and also a qualified scrutineer, and there weren’t that many people that were both; not sure what present requirements are.)

These days, the scrutineer doesn't actually know which couples will be affected either, unless they actively go looking after the software tells them there's a recall issue. I've certainly never done so myself, and I can't imagine anyone I know doing do—nobody cares, and there's really just no time.

The Chair still needs to be a qualified scrutineer—at least having passed the exam, though they will likely never have actually practiced as one.

 

[Warren J. Dew]

These days, the scrutineer doesn't actually know which couples will be affected either, unless they actively go looking after the software tells them there's a recall issue. I've certainly never done so myself, and I can't imagine anyone I know doing do—nobody cares, and there's really just no time.

The scrutineer doesn’t see the couple numbers?

Not saying any scrutineer would actually do this, but couples looking for someone to blame may make accusations that are not true. Most people do not understand the time pressures involved at the head table.

The Chair still needs to be a qualified scrutineer—at least having passed the exam, though they will likely never have actually practiced as one.

Probably a good thing that the requirement is still there.

 

[j_alexandra]

The Chair still needs to be a qualified scrutineer—at least having passed the exam, though they will likely never have actually practiced as one.

IIRC, the Chair has to be qualified to fill any position.

 

[Joe]

The scrutineer doesn’t see the couple numbers?

The scrutineer can see all the couple numbers. The scrutineer can also see the marks, but they have to make additional effort to do so. But when actually working the heat, zero scrutineers care about which couple number the cutoff point is going to be when they ask the Chair to select 11 or 14 couples into the semi.