While @W1GRD has been wandering the graveyard searching for inspiration I have been doing some low level statistics analysis of activators and activations.
I’ll be writing this up in more detail, but just a few observations that have dramatically changed my understanding of how the program works.
Ranking activators by number of activations:
The top 1% account for 25.8% of all activations.
the top 5% account for 55.7% of all activations.
the top 10% account for 71.6% of all activations.
the top 25% account for just under 90% of all activations
20% of all activators have just one activation. More than half have 6 or fewer.
Very interesting stats. They bring up many questions.
When you say “activation”, do you mean a POTA “successful” activation or all attempts?
I would expect the 20% with one activation to have a much higher chance of also having unsuccessful attempts and also a higher chance of one activation years ago but really being inactive. My own first (failed) activation was followed by 6 months of health problems.
Would be interesting to see stats for mode. My guess is that ops with CW have higher success rate or higher number of activations.
I don’t see anywhere on the POTA sites to find the data for ops other than myself.
Those statistics are gleaned from the “top activators” list, which counts only ‘valid’ or ‘successful’ activations. I don’t much care for that terminology nor the taxonomy, but that’s what the program says and I’m just going along with it.
There’s no aggregated data that I know of that includes how many ‘failed’ activations a participant has.
Anecdotally, it does seem that many activators have one or more ‘failed’ attempts before their first ‘successful’ activation but the caveat here is the usual aphorism that ‘anecdotes are not data’. Note that many ‘failed’ attempts are things like ‘forgot the battery’ or ‘forgot coax’ and result in no log being uploaded, and thus would escape the statistics no matter what.
It would make sense that as an activator gains experience they improve the odds on each activation being successful, but I can tell you that in my experience, no matter how good you are if you keep doing activations eventually you will get skunked, if only by bad propagation condx.
I’ve not looked at how the stats break down by mode in detail; that’s a good suggestion and I’ll take a look in the coming week.
My gut tells me the numbers are mode-sensitive, so there will be 5% of the CW ops running 90% of the CW activations.
There’s always the question of overall size. I’ve seen 65k total POTA participants tossed around but no definition of particiapnt. The Facebook POTA group has about 35k members and seems pretty active, but again, how does pretty active translate to population size?
My back-of-the-envelope scratching use 100k as a nice round population size.
Stats I’d love to see:
Dwell time between activations
Time since last activation
Time between first and most recent activation
Median frequency of activations
Activation diversity (do people just sit in one park or spread out?)
Pota.app lists 15416 callsigns that have done at least one activation in 2025 so far. I would use that as a reasonable first order estimate of the number of ‘active’ activators.
Pota.app lists 36085 callsigns that have hunted at least 10 activations in 2025 so far. Again, I’d use that as a reasonable first order estimate of the number of ‘active’ hunters.
When I wave my hands over my head, stand on my left foot and hop up and down, and use my finely calibrated bullshit generator I come down with a rough estimate of something on the order of 50K callsigns are ‘active’ as hunters, activators, or both. I have a high degree confidence that it’s more than 5k and a high degree of confidence that it’s less than 500k.
The obvious point is that I’m pretty confident that the total number is also less than the 100K on Facebook. My first order take on that difference is that I don’t believe anything Facebook/Meta tells you, and on top of that i observe that while there are a few very POTA active hams on the various POTA social media sites, my guess is that the vast majority of participants on social media are not really hugely active, because given a choice between getting on Facebook, or going out to a park or getting on the radio to hunt, they’re picking Facebook.
Bottom line: how big is POTA? In one very real sense, it’s the biggest thing in ham radio, and in another very real sense, it’s not really all that big in absolute terms.
How far back do the spot posts go in #pota-spots? That’d give you some really fine-grained data on who and when and what and even where. Why, that’s a question we’re still asking.
It would be nice if POTA.app exposed more of their data stream and their historical database; it would be a treasure trove of empirical data about operating patterns, propagation, &c.
Sadly, a) they seem to want to play the game with their cards pretty close to their chest, and b) I get the feeling that they are strapped for development resources and the entire thing is held together with spit and baling wire. Neither of those does much to ease my concerns that the entire thing will eventually fold up like a cheap card table at a big potluck dinner.
I do understand that they want to shape things so that it is cooperative and doesn’t just turn into a cutthroat competition, and I am foursquare with that. But I also think the best path forward is to be ruthlessly transparent with the data so that outside innovation can expand the game.
Heh, can’t argue with that. I sent a note to POTA Prez Rick Parent with a heads-up that I was planning a new awards program, his response was that he’d never heard of me or PN&R and felt since it wasn’t official POTA business he didn’t have a position.
I noted that he’s got about 30 activations total, and it sounds like a lot of those were early days. Whatever.
I was thinking of the Discord#pota-spots channel. There’s a ton of data there and it shouldn’t be hard to write a bit of code that ingests that feed and tracks stats almost in real time. I added it to my planning list, I think it os low effort unless you get to it first.
Based on my experience serving on the boards of various non-profits, I can tell you that the skill set needed to shepherd an organization/effort along and the skill set needed to participate in whatever activity the org is promoting are frequently orthogonal. It helps to have some understanding of what’s going on but my observation is that some of the most accomplished participants in POTA absolutely lack the interpersonal skill set needed to do management.
So I would not spend much time fretting about that.
Scraping the POTA spots channel would work, of course, but it would be so much simpler if POTA just exposed a feed of activations. Presumably there’s code somewhere that’s scraping the POTA spots page and feeding the spots channel. You can make inferences based on spots (e.g. it’s highly likely that if I spot a W1GRD activation as “529 in WA CN97bq TU” that’ I’m in your log, but of course you can respot someone even if you tried but failed to get in their log. Maybe that functionality is exposed at some informal level, I don’t know.
It took an astounding number of steps to synthesize POTAmap’s per-mode QSO counts. All I wanted to do was to use PQL queries like ?mine:0 mode:cw max:0 maxdist:25 to find parks I haven’t activated that are close and have zero CW QSOs logged. Easily a third of the batch work at night on my side is updating those counts.
It’s funny I’m of two minds on ‘official’ data. Of course it would make life easier for me, and a lot of other folks, but I’ve synthesized data that isn’t available via an official API call, and I don’t want to lose my edge :^)
n-Fers, too, there’s a job that runs quarterly to infer n-fers from accumulated data. It’s my N-ference Engine.
As far as scraping #pota-spots, I’m thinking more of capturing meta data, so that on the PN&R home page I could have a counter: “# Active Right Now” by mode.
Oh, F me, I have that data already, POTAmap snags the activation feeds every six minutes. I just need to add a step to parse out the meta stuff.
@w7pfb not to hand you any work, but as we are thinking about awards it’d be interesting to know some median numbers in order to set a challenge correctly. For example I arbitrarily picked 50 QSOs for a few award ideas, but I don’t know if 50 is a one-year or a one-week thing.
Just something to think about when you are wading around in data.
All time, half of all activators have six or fewer activations.
in 2025, half of all activators have five or fewer activations.
In 2025, half of all activators have 139 or fewer QSOs during activations.
At this point I’m not so much wading around purposefully in the data. It’s more like I’m restfully floating on my back and doing a very leisurely backstroke while considering whether to paddle over to the swim-up bar and get a cocktail, and whether I should go for a Sazerac or a Maker’s Mark manhattan when I do that.
I always forget that a lot of ops are there for the numbers…my activation style is A Dozen and Done and back on the trail. It’s a huge blind spot for me.
Yeah, that’s 139 QSOs total across all of their activations so far in 2025.
So if they had, say, 5 activations, then they averaged 139/5 = 27.8 Q’s per activation.
When you look at the range of QSO’s per activation across all activators, you find some very unusual things. Example: CY9C has done a total of 12 activations and only ever activated one park. Across those 12 activations of that one single park, CY9C has logged 112758 QSO’s, for an average of 9396.5 QSOs per activation.
I’m guessing that this is a DXPedition, but still. Each activation can only span a single UTC day, so that works out to better than 6 QSOs/minute, which is a helluva rate.
At the other end of the range you have VA2JQT with 39 activations in 39 different parks, and a total of 390 QSOs, so exactly 10 per activation.
The thing that makes the cocktail choice difficult is that making a really good Sazerac is non-trivial and you only really want one that’s well done, and the odds of getting that at a swim up bar are poor. So, high probability of low satisfaction but small chance of high satisfaction. The key to getting an acceptable MM manhattan is the vermouth choice, and they’ll probably have MM, Antica vermouth, and some decent orange bitters even at a swim up bar. Would I say that will be an excellent manhattan? No, but it’s a swim up bar. The key to avoiding disappointment here is to have realistic expectations. (Side note that @W1GRD will no doubt appreciate: a small group of my friends has run a years long project to find the best combinations to build a manhattan. We call this the Manhattan Project.)