[Discussion] Cumulative Cutoffs

[Ranzha]

Hello everyone! I've been wanting to discuss this subject for a long time.

Cumulative cutoffs! What are they, how are they used, and should they be included in the regulations (and if so, how)?

What they are: For a combined round with a cumulative cutoff, the competitor's cumulative solve time during the first phase determines if they advance to the next phase. The time elapsed in a DNF result counts towards the cumulative cutoff. This is currently not allowed for combined rounds.
How they have been used: For a long time in the San Francisco Bay Area, cumulative cutoffs were used for 3x3 Blindfolded instead of using a cumulative time limit, usually with something like 15:00 cumulative cutoff. So after each of the first two attempts, the competitor's elapsed time would be assessed, and if they had not yet exceeded 15:00 of solve time, they would be eligible for the next attempt. During the next attempt, the judge would allow the competitor to solve past 15:00 elapsed solve time if needed, and would not stop the competitor. After an attempt where elapsed solve time exceeded 15:00, they would not receive any result for subsequent attempts.
The way it should have been done, according to the Regulations: While there is no cumulative cutoff in the regulations, there is the option to impose a cumulative time limit. In rounds with a cumulative time limit, once the competitor reaches the time limit, the judge must stop the attempt and record the result (and subsequent results) as DNF.

I think cumulative time limits are unnecessarily harsh to the competitor for a few reasons:

1) Cumulative time limits do not distinguish between DNFs due to not solving the puzzle in an attempt and DNFs due to being disqualified from future attempts.
I am of the belief that if a competitor is ineligible for an attempt, they should not be penalised as if they were eligible for the attempt. Combined rounds do not give competitors DNFs for not advancing to the second phase (i.e. no penalty for not being eligible for future attempts). This model accurately reflects what actually happened.

2) Cumulative time limits stop competitors, which can incentivise DNFing. I understand having time limits for individual attempts is necessary--there must be some way to keep competitions moving forward. However, knowing that there's an "expiration time" to my round, if I have a puzzle defect, I might choose to take a DNF so I allocate more time for future attempts in the round knowing that helps minimise the chance I get another DNF. With a cumulative cutoff, I don't have to worry about the chance of getting another DNF (or even one DNF) since attempts I don't get aren't penalised.

3) Cumulative time limits can hurt competitor morale. This more has to do with time limits in general, but especially during BLD, if I get stopped during an attempt, all my focus goes out the window. It lacks the closure of stopping the timer and assessing the result. It negatively affects how well I can focus in future attempts, more than if I actually didn't solve the puzzle. During BLD (esp MultiBLD), I have had to stop competitors' attempts, and they're never glad that I've had to stop them. With cumulative time limits, competitors can be discouraged from even undergoing an attempt if they know they have little-to-no chance of finishing the attempt uninterrupted.

These considered, cumulative time limits are very useful to organisers/delegates for planning schedules and keeping competitions on-track. But I think the balance between the benefit to organisers/delegates and the detriment to competitors is better satisfied with an option for cumulative cutoffs.

Discuss!

 

[JustinTimeCuber]

Wouldn't that still incentivise DNFing? Say I'm competing in Megaminx and there's a cumulative cutoff of 4:00. If I finish both attempts, I won't make it, so I decide to DNF my second attempt and my results are as follows:
2:31.09, DNF(0.03)
The time elapsed is 2:31.12, so do I make the cutoff, or am I misunderstanding?
A1c would, I guess, make that not be allowed, but what if my Megaminx pops and I decide not to fix it? Same idea,
2:31.09 DNF(1:28.90)
Again, the time elapsed is 3:59.99, so would that make the cutoff? How do you tell if a competitor intentionally pops a puzzle to exploit this? Do you ban all "intentional" DNFs to get around this? If so, I'm interested in how that would work. That would also imply that if someone is using an SS 6x6 and gets one of those big explosions that happen once in a while, and stops the timer, they are disqualified from the event or something?

 

[sqAree]

I've been told I'm allowed to DNF a BLD attempt if I think the scramble is bad and I have better chances with the 3rd one, regarding cumulative time limit.

 

[Ranzha]

Wouldn't that still incentivise DNFing? Say I'm competing in Megaminx and there's a cumulative cutoff of 4:00. If I finish both attempts, I won't make it, so I decide to DNF my second attempt and my results are as follows:
2:31.09, DNF(0.03)
The time elapsed is 2:31.12, so do I make the cutoff, or am I misunderstanding?
A1c would, I guess, make that not be allowed, but what if my Megaminx pops and I decide not to fix it? Same idea,
2:31.09 DNF(1:28.90)
Again, the time elapsed is 3:59.99, so would that make the cutoff? How do you tell if a competitor intentionally pops a puzzle to exploit this? Do you ban all "intentional" DNFs to get around this? If so, I'm interested in how that would work. That would also imply that if someone is using an SS 6x6 and gets one of those big explosions that happen once in a while, and stops the timer, they are disqualified from the event or something?

I really wasn't clear...sorry.
Cumulative cutoffs would be an alternative to cumulative time limits and function in much the same way (i.e. across all attempts), except:
- Competitors would not be stopped mid-attempt, and be allowed to finish the attempt during which they exceed the cutoff
- The competitor is not penalised with DNFs for not being eligible for future attempts

It wouldn't function like most cutoffs today, where phases are usually Best of 2 -> Average of 5 or Best of 1 -> Mean of 3. The first phase would be across all attempts until exceeded, at which point the second phase starts. (This means the second phase might have 4 attempts, or it might have 0 attempts.) So everybody gets the same amount of elapsed solve time to work with without being penalised for multiple attempts for going over on one attempt.

 

[mark49152]

Cumulative time limit is a clean way of conserving competition resources and personally I don't object to being stopped. If I attempt a 3BLD with 2 minutes' time limit remaining, I know I have to get it done inside 2 minutes, will pace it accordingly, and no complaints if I don't make it.

I think a stronger argument for this idea is ensuring consistency and fairness. Some judges will be too nervous or shy to stop the competitor, or won't have the experience to be sure they should. That's not fair to those who do get stopped, so it should then be up to the score taker to apply the maths and DNF any results that shouldn't have been recorded. But does that always happen?

 

[guysensei1]

If you get a bad BLD scramble and you intentionally DNF to give yourself more time for the remaining 2 attempts, does it go against A1c?

 

[Laura O]

So what is the problem with being stopped when you reach the cumulative time limit? You are also stopped when you reach the "normal" time limit: "A1a4) The competitor must end each solve within the time limit. If a competitor reaches the time limit for a solve/round, the judge stops the attempt immediately and records the result as DNF. [...]".

And well, I actually do not understand the reasons mentioned in your post. You completely miss one big advantage cumulative time limits have: slow competitors are eligible to do a solve.
To give an example: compare a round with a time limit of 5 minutes and a round with a cumulative time limit of 15 minutes.
A slow solver, averaging 6 minutes would be stopped at every single attempt with the time limit. With the cumulative time limit, he has the chance to do two solves. So you call that harsh?

 

[Goosly]

To give an example: compare a round with a time limit of 5 minutes and a round with a cumulative time limit of 15 minutes.
A slow solver, averaging 6 minutes would be stopped at every single attempt with the time limit. With the cumulative time limit, he has the chance to do two solves. So you call that harsh?

+ regarding OP's third argument (being stopped during an attempts hurts morale): If you already used 12 of your 15 minutes, simply don't start your third attempt.

Anyway, at my last comp, I got a 5/12 (58:xy) multi, which felt even worse than being stopped during an attempt. A few hours later, I got 3BLD single & mean NR. So, OP - no offence - but if you lose morale because of bad solves, the regulations shouldn't fix that, you should fix that yourself.

 

[Chree]

Just a quick anecdote on Point #2 from the OP.

During a competition last summer, Rose City 2016, we used cumulative cutoffs for BLD. During my first attempt, I made it through my entire (very slow) memo before I realized a missed a Corner. I was already 4 minutes into my attempt, and knew that re-memorizing would cost me almost as much time. So rather than double the amount of time for an individual attempt, where the scramble was really bad and I knew it, I opted for the chance at a completely fresh start, and purposefully DNF'ed.

Looking back on it now, I'm not sure how this meshes with the current regulations. I think it's still OK, since I already knew the attempt would not be my best, and was hoping for a chance at something better.

 

[mark49152]

I have done the same thing, abandoned a BLD solve once I realized my memo was so slow as to prevent a good result. I don't consider that an "intentionally poor result" according to the spirit of the regs. My intention was always to get a good result, it just didn't happen.

 

[JustinTimeCuber]

I am still a bit confused. What happens exactly when you hit the cumulative cutoff?

Sent from my MotoE2(4G-LTE) using Tapatalk

 

[Dene]

I find it interesting that you say you've been doing it wrongly where you're based.

But I don't get why people would get a DNF for future solves they're ineligible for. As far as I'm aware, if you never start an attempt, it's DNS. So if there is a cumulative time limit of 10 minutes, and your first solve is 9 minutes and you decide not to have any more attempts, your results should be 9, DNS, DNS.

 

[Ranzha]

So what is the problem with being stopped when you reach the cumulative time limit? You are also stopped when you reach the "normal" time limit: "A1a4) The competitor must end each solve within the time limit. If a competitor reaches the time limit for a solve/round, the judge stops the attempt immediately and records the result as DNF. [...]".

I understand having time limits for individual attempts is necessary--there must be some way to keep competitions moving forward.

--

You completely miss one big advantage cumulative time limits have: slow competitors are eligible to do a solve.

This is a shared advantage of cumulative cutoffs--I'm not sure how that was unclear.

--

...A slow solver, averaging 6 minutes would be stopped at every single attempt with the time limit. With the cumulative time limit, he has the chance to do two solves. So you call that harsh?

Allowing slower solvers two attempts is great! And allowing them to undergo their third attempt uninterrupted is better than stopping them mid-attempt and DNFing them.

--

I think a stronger argument for this idea is ensuring consistency and fairness. Some judges will be too nervous or shy to stop the competitor, or won't have the experience to be sure they should. That's not fair to those who do get stopped, so it should then be up to the score taker to apply the maths and DNF any results that shouldn't have been recorded. But does that always happen?

As a score taker, I have had to enter DNF for competitors' attempts since the judge did not enforce the cumulative time limit.

--

+ regarding OP's third argument (being stopped during an attempts hurts morale): If you already used 12 of your 15 minutes, simply don't start your third attempt.

With a cumulative cutoff, you don't have to do this! You get your entire third attempt.

--

if you lose morale because of bad solves, the regulations shouldn't fix that, you should fix that yourself.

more fun

Stopping people is less fun than letting them continue. (The information that they will get no more attempts is constant.)

--

I am still a bit confused. What happens exactly when you hit the cumulative cutoff?

Example:
BLD Round 1 (cumulative cutoff 10:00, time limit 15:00 per attempt)
Competitor: You
Your first attempt: DNF(3:45.00)
You are allowed a second attempt since you have not yet reached the cumulative cutoff.
Your second attempt: 4:30.00 (elapsed: 8:15.00)
You are allowed a third attempt since you have not yet reached the cumulative cutoff.
Your third attempt: 3:30.00.
Results entered as: DNF, 4:30.00, 3:30.00

Competitor: Me
My first attempt: 6:25.00
I am allowed a second attempt since I have not yet reached the cumulative cutoff.
My second attempt: 3:36.00 (elapsed: 10:01.00)
My second attempt is valid as 3:36.00, but I am not allowed a third attempt since I have reached the cumulative cutoff.
Results entered as: 6:25.00, 3:36.00
Here there is no result for the third attempt since it never existed, like in a Combined round.

Competitor: Slowy McWristTurn
Slowy's first attempt: 11:00
Slowy's attempt is valid as 11:00, but is not allowed any additional attempts since Slowy reached the cumulative cutoff.
Slowy's results entered as: 11:00
Here there is no result for the second or third attempts since they never existed, like in a Combined round.

Another example with the same attempts, but a cumulative time limit instead of cutoff:
BLD Round 1 (cumulative time limit 10:00)
Competitor: You
Your first attempt: DNF(3:45.00)
You are allowed a second attempt since you have not yet reached the cumulative limit.
Your second attempt: 4:30.00 (elapsed: 8:15.00)
You are allowed a third attempt since you have not yet reached the cumulative limit.
Your third attempt: When the timer reaches 1:45.00, you reach the cumulative limit. You are stopped and receive a DNF for your last result.
Results entered as: DNF, 4:30.00, DNF

Competitor: Me
My first attempt: 6:25.00
I am allowed a second attempt since I have not yet reached the cumulative limit.
My second attempt: When the timer reaches 3:35.00, I reach the cumulative limit. I am stopped and receive a DNF for my second result. I am not allowed a third attempt since I have reached the cumulative limit.
Results entered as: 6:25.00, DNF, DNS (thanks, Dene)

Competitor: Slowy McWristTurn
Slow's first attempt: When the timer shuts off at 10:00, Slowy reaches the cumulative limit. Slowy is stopped and receives a DNF for the first result. Slowy is not allowed any more attempts since Slowy reached the cumulative limit.
Results entered as: DNF, DNS, DNS

 

[1973486]

I find it interesting that you say you've been doing it wrongly where you're based.

But I don't get why people would get a DNF for future solves they're ineligible for. As far as I'm aware, if you never start an attempt, it's DNS. So if there is a cumulative time limit of 10 minutes, and your first solve is 9 minutes and you decide not to have any more attempts, your results should be 9, DNS, DNS.

Are you saying you don't understand it or that you think it should be changed? Because it's quite clear why you get a DNF right now.

EDIT: Lucas mentioned A1a2+++ which seems like it conradicts 9f5. If the competitor is ineligible why are they given a DNS?

 

[JustinTimeCuber]

OK, so I like this idea. Competitors shouldn't have results for attempts they are ineligible for. That would be like having everyone who didn't make r2 of an event get results of DNS for it.

Sent from my MotoE2(4G-LTE) using Tapatalk

 

[Lucas Garron]

HHow they have been used: For a long time in the San Francisco Bay Area, cumulative cutoffs were used for 3x3 Blindfolded... During the next attempt, the judge would allow the competitor to solve past 15:00 elapsed solve time if needed, and would not stop the competitor.

‽ I don't recall seeing this, and would have pointed out that it is against the Regualtions

1) Cumulative time limits do not distinguish between DNFs due to not solving the puzzle in an attempt and DNFs due to being disqualified from future attempts.

We've cleared up that this is false due to A1a2+++, right?

2) Cumulative time limits stop competitors, which can incentivise DNFing.

Based on your paragraph following this, it seems like your concern is not actually the DNF result. Rather, it seems you're concerned that the choice to DNF can allow competitors to take up more of the time limit.

But the time limit is the time limit – by setting it to a certain value, you are giving the competitor license to take up that much time. And I think your proposal would generally let them take even *more* time, right?

3) Cumulative time limits can hurt competitor morale.

This may be, but it doesn't seem important enough to me on its own. Competitors will always be sad to hit time limits, and this is just one case.


I think that it is not a good idea to introduce "cumulative cutoffs". There is already a lot of confusion and inconsistent terminology around time limits, cutoffs, "soft cutoffs", and "hard cutoffs" (and, you claim, incorrectly applied "cumulative cutoffs" in the past). This proposal would add a new kind of continuous combined round, which I don't think we actually need. Are there any organizers or competitors from other regions who would like to add this?

(It also wouldn't be a good idea to *replace* the existing cumulative time limit system entirely, since much of the world is already using it properly. The switch would cause massive confusion, probably resulting in many mistakes against the favor of competitors.)

 

[Ranzha]

‽ I don't recall seeing this, and would have pointed out that it is against the Regualtions

I thought cumulative cutoffs were standard operating procedure at Berkeley, so I enforced them for BASC until I was corrected by Kit Clement when he came to delegate BASC 5.

We've cleared up that this is false due to A1a2+++, right?

Yes!

Based on your paragraph following this, it seems like your concern is not actually the DNF result. Rather, it seems you're concerned that the choice to DNF can allow competitors to take up more of the time limit.

I am concerned about both--competitors being incentivised to DNF, and the result itself.

But the time limit is the time limit – by setting it to a certain value, you are giving the competitor license to take up that much time. And I think your proposal would generally let them take even *more* time, right?

Yes.

 

[Dene]

Are you saying you don't understand it or that you think it should be changed? Because it's quite clear why you get a DNF right now.

9f doesn't directly address this circumstance. 9f4 doesn't apply because it requires the attempt to have been started (or disqualified). 9f5 requires a competitor to be eligible.

A1a2+++ is the correct place to look for this scenario.

 

[One Wheel]

The examples here are for Blind, and it seems to me to make a lot of sense. Would it work for speedsolving, especially large cubes? If, for example, instead of a 5:00 cutoff, there was a 10:00 cumulative cutoff, then if my first two were 5:50 and 4:09 I could get a mean (for 6, 7, or feet), and if my first solve was 9:55 I could get a second solve? I'm really liking the sound of that. It seems that it would make cutoffs much less frustrating, because in a sense close might actually be good enough.

 

[newtonbase]

Initially I agreed with this idea but time limits are there for a reason. Here in the UK I have always been made when it looks likely that I won't complete my next attempt which is good practice as it gives the competitor the choice whether to risk a DNF or not.