833 Round Robin

Members and associates of ATF Professionals hold many of the exclusive-by-market vanity toll-free programs in existence today. Programs such as 1-800-Home-Care, 1-800-Pavement, 1-800-Injured, 1-800-800-Cars, 1-800-Laywers, and 1-800-Attorney are examples of exclusive-by-market licensing.

We fully support the Commissions effort to make 833 “numbers available on an equitable basis” using an off-line process. In a 2010 submission by Vanity International, a founding member, we outlined how the majority of RespOrgs—and thereby subscribers— were disadvantaged by the high-powered, specialized automation held by a few. Further, we predicted that a 100-per-day allocation did not ensure equal access to 855 toll-free numbers, and later 844 numbers, as it did nothing to “avert those with the most power connections from getting first pick of the best available numbers.” We hold today, as recommended then, that code openings are an “anomaly” and best allocated by an off-line, randomized round robin (RRR) process.

First, though, we’d like to address a troubling reference in the April 21, 2017 order to assignment by “auction or lottery.” While we are recommending a form of lottery (RRR), we do not support any form of auction for the following three reasons. We believe each rationale is fully aligned with the Commission’s historical position and statutory mandate:

  1. Implied Property Value: A cash auction implies that all toll-free numbers now have property value and can, therefore, be resold. Time and again, the FCC has held that the sale of toll-free numbers is contrary to the public interest. Further, I distinctly recall Les Seltzer stating in his presentation at the Forum on Toll Free Administration, March 4, 2002, that “it would take an act of congress” to instill property value in toll-free numbers. Compelling users to buy numbers, at any price, would set a precedent that assignees have the right to sell.
  2. International Concerns: During the 888 replication (circa 1998), the idea of auctioning off the 888 number pool was proposed by Senator McCain, but ultimately dismissed, once they recognized that the North American Numbering Plan covers dozens of countries, including Canada and many in the Caribbean. With that, what right does US, or its agencies, have to unilaterally benefit from an auction? Alternately, Somos is fully funded by the toll-free community and in no need of additional revenues derived from the community they serve.
  3. Discrimination: Even if property value and international concerns were dismissed, imposing an auction on the public would defeat the statutory mandate to make “numbers available on an equitable basis,” as only well-funded corporations and the well-healed need apply. While 833 numbers are in demand by all subscribers, an auction would render them out of reach to anyone but the very wealthy.


Common Per-Number Lottery

In the alternate, holding a “lottery” in its common form is just as troubling. A per-number lottery creates an enticement for mass speculation and number brokering. Put another way, when lottery tickets are free— zero cost— how many chances do you take? Answer: “All I can get.”

In a per-number lottery, there is zero cost to submit the maximum number of requests (currently 2,000), so why not speculate? For instance, if a given RespOrg only needs 1,200 numbers there is no disincentive to speculate on 800 more, as each additional request is just as free. A per-number lottery means that each submission adds one new chance for reward, with zero downside. Should such a lottery be open to public, the speculation would only grow exponentially.

Further, since a lottery on a per-number basis results in mass speculation, the best numbers are far more likely to fall into the hands of non-users, who secured them only for resale or to leverage their assignment to the prospects that really want them. The FCC already has experience with failed lotteries[1], when it allocated cellular telephone licenses in the 1980’s. This open process “resulted in hundreds of thousands of applications,” where most were hoping to “win a license purely for resale.”

Randomized Round Robin

These perverse incentives are precisely addressed by the Randomized Round Robin (RRR) lottery. The randomization, in effect, creates a cost for speculation without imposing any actual monetary value. Rather, a RRR imposes a trade-off where each additional submission diminishes the odds of getting the other numbers on the list.

Before we explore the steps, I’d like to point out that the RRR is just a common lottery with this additional feature. If equal-sized, speculative lists (2,000 let's say) from the pre-assignment round were randomized and drawn in random order, the results would be the same as in a common lottery. For instance, of five RespOrgs requesting the same number, one of them would be drawn first, just as if they were drawn from a hat. The difference is that RRR creates an advantage to those who restrain from speculation, where the common lottery does not.  

Here is a summary of the steps:

§  Participating RespOrgs submit a request list with up to 2,000 selections[2]

§  The list order is randomizes, thus, there is no advantage to seeding the best numbers first, nor will all the best numbers be assigned in the first rounds.

§  The draw order is randomized, thus, all RespOrgs have any equal chance to being the first assignment, second, third, and so forth.

§  The Round Robin draw proceeds, one request per round, from the first list to last, from the first request to last one, where both the draw order and list order are now randomized.

§  Numbers are assigned only if available at the time of the request (“Assigned”). If the requested number is unavailable (“taken”), no assignment is made.

§  The Round Robin proceeds for 2,000 rounds, where one request from each RespOrg is either assigned or denied, or until the longest request lists are fully processed.

The Randomized Round Robin (RRR) format was presented in its entirety in the September 2, 2010 comments and has already been out for public comment. While this RRR assignment as simple enough to be processed with spreadsheet, the registry, now Somos, has in-house programming expertise to automate the process and to share results in near real-time. Transparency will foster trust in the process.

Equity is maintained in a RRR lottery by allocating chance over the entire list, forcing each participating RespOrg to trade-off certainty of assignment for quantity of numbers requested. Using the above example, the RespOrg speculating on 800 numbers will, by design, suffer reduced certainly of getting any one of 1,200 they really need. Rather, by submitting a list of just the 1,200 needed numbers, they will improve their odds of getting every one of them[3]. Yet, because the lists are randomized, every number a RespOrg submits has an equal chance of being assigned in the first round and each round beyond that. 

The only known objection was made by the Toll-Free Number Coalition in their September 9, 2010 reply comments and addressed by Vanity International in their September 9, 2010 reply and continuing comments on September 13, 2010. The Coalition objected to randomized list order and, rather, suggested that lists should be seeded: submitted in the “customer’s order of preference” so that their first choices will be assigned before the others, thus improving the odds of those assignment. While this argument seems appealing, without randomization there is no disincentive to speculate. RespOrgs would pay no price for submitting speculative requests. Rather, our model RespOrg would simply submit the 1,200 numbers they need first, and then speculate on 800 more below that. What’s worse is that list seeding invites discrimination, as carriers will certainly submit requests from their biggest and more important customers first, seeding the less important ones for end, or leaving them off all together.

Finally, if any RespOrg asserts they are disadvantaged by this method of equitable allocation, it is essential to note that “real customers” are in no way disadvantaged; they can place their order with any RespOrg they like.

Cap and Resubmit Request Lists

Regardless of auction or lottery type, RespOrgs must be allowed to reduce their May, 2017 list size to allow for cancellations, subscribers that no longer desire numbers placed on the original list and have moved on. All submitted lists are only accurate on the day of submission. The best practice, we believe, is to require all participating RespOrgs to resubmit a list once a final order has been issued. In the interim, all will be free to work with newcomers and work around lost orders within the constraints of their original, May 2017 submission, as detailed below.

We raise a strong objection to submitting the final request lists prior to knowing the ramification of list size or the future status of customer orders, which cannot be known in May, 2017. Specifically, because the lottery format was not defined in time for the submissions, we respectfully request that lists submitted in May only be used to determine unique assignments and “mutually exclusive” set-asides, and not deemed the official request list for this as-yet-defined “auction or lottery.”

The April 22, 2017 order has no disincentives in place to constrain speculation or right-size the submission list. Just like the per-number lottery discussed above, a longer lists assure more chances for pre-assignments at zero cost. Yet, we believe the May submissions were of great value, as they immediately removed and pre-assigned all non-contested numbers and, in effect, crowd sourced the numbers that were destined for the mutually exclusive pool. A final submission list will not only depend on the type of lottery used but, as discussed below, customers are certain to come and go over the time.

The final question is whether the lottery should be open or closed, open to newcomers or restricted to the just the original, contested submissions. We don’t have a fast position on which way this should be decided but, rather, believe the FCC has the best vantage point to make such a determination. Time is enemy. 

Surely, in the intervening weeks or months, there will be new customers demanding access to these “mutually exclusive” numbers. The counter argument is that these numbers we set aside for those who requested them at the pre-assignment round— and everyone else just missed the deadline. A further reality is that existing orders are likely to be canceled by customers who had hopes of getting numbers in the pre-assignment, but have since moved on to more immediate options. The longer this decision languishes, the more pressure will mount in all directions.

Had the lottery format been included in the order, then the submission would be the closed by design. As time goes by, however, pressure will build to open it to new customers. We believe there is one good way to resolve this dilemma that aligns with all of the above concerns.

The FCC could require RespOrgs to resubmit a final list, up to the number of mutually exclusives on their original list, once the lottery format is defined. That is to say, for example, if a RespOrg submitted a list of 1,500 with 500 pre-assignments, they would then be eligible to submit a revised list not to exceed 1,000 or, at their discretion, simply resubmit the original list, less the 833 numbers pre-assigned at the opening. Another RespOrg submitting 2,000 numbers with 300 mutually exclusive would be allowed a list of just 300. The rule: Participating RespOrgs can ask for no more numbers then they requested in May 2017. The FCC could further constrain the resubmitted lists to include only those numbers initially requested, at its discretion, i.e. each RespOrg can ask for no other numbers then those originally requested.

This rule respects both the integrity of the May, 2017 deadline and effectively bars participating RespOrgs from expanding their influence at a later date and time, and non-participating RespOrgs from having a “second bite at the apple.”

The magic here is that RespOrgs can concurrently add new customers as they replace customers lost through attrition, all while maintaining the integrity of the process. While participating RespOrgs would be constrained by their May, 2017 submissions, it is essential to understand that the public would not barred from participation. They can work with the participating RespOrg of their choice. There will likely be dozens, if not hundreds, to choose from.

Respectfully submitted,

Loren Stocker


Founder, Association of Toll Free Professionals















participation if RespOrg are not allowed to revise and resubmit their request list.


non-participing RespOrgs by containing the total requests to no more than requested them,

missed the deadline

didn’t participate

did not ask for 2,000

Wholesale level.

The magic here is that RespOrgs can

real customers can add/omit

attrition or addition.

Slots used for n


The FCC willthe revised list could be closed to mutually exclusives on the original list, if decided at once; open to revision if this decision lingers.

The final issue is should it be limited to just those number included on the list.


We fear, though, that many weeks and even months will go by the process is defined.




It’s a fact that any given RespOrg




[1] See bottom of page 3: http://faculty.smu.edu/tsalmon/fccchapter.pdf

[2] As we outline below, we recommend that the FCC cap the list size at the quantity of numbers submitted in the in the pre-assignment round in May, 2017, minus the quantity of numbers pre-assigned. For example, a RespOrg submitting 1,500 requests where 500 were pre-assigned, would then be eligible to revised and re-submit up to 1,000 requests. The only way any RespOrg would be eligible for list size of 2,000 is to have submitted 2,000 in May, 2017 and none were pre-assigned.

[3] If this is not obvious, here’s how this works: Submitting 2,000 numbers will seed the 1,200 needed ones over 2,000 rounds of draw, as the lists are randomized. In contrast, submitting just the 1,200 needed numbers will seed them over the first 1,200 rounds, thus the 1,200 are more likely to be assigned by eliminating 800 rounds of chance.