
The trade you are making
Paper ballots cost next to nothing and work in a field with no signal. Phone voting needs a connection and a plan for the people it leaves out, but it gets you out of the counting room. Nearly everything else in this comparison is a footnote to that trade, and which side wins comes down to how many ballots you expect and how long your crowd will wait for trophies.
| Paper ballots | QR voting | |
|---|---|---|
| Up-front cost | Printing, pencils, boxes | Depends on the tool |
| Needs a signal | No | Yes, on the voter’s phone |
| Who gets left out | People who miss the ballot table | People without a phone, or with one in the truck |
| Crew | Ballot table, plus counters | A volunteer to help at a tablet or kiosk |
| Time from close to results | 20 minutes to 90 minutes or more | Minutes |
| Worst day | You run out of ballots, or the count eats your awards slot | The signal drops or location checks fail |
| Bloc voting by a club | Still happens | Still happens |
What a paper count takes
The most useful account we’ve found is a 2008 British Car Forum thread where organizers compared notes. One show gives every participant a booklet of color-coded ballots, one per class, with a matching colored box near the announcer. It keeps one counter per class, pulls ballots 45 minutes before voting ends, pulls them again at the close, and aims to start awards within 30 to 45 minutes. That is a well-drilled club with 34 volunteers. Another organizer in the same thread has eight people counting in batches and still needs about an hour and a half.
Small shows do better, since the math is kind to them. An art league’s committee procedure (yes, an art show, but a ballot box is a ballot box) plans on 200 to 250 ballots, four people handing them out, two more adding them up, and at least 15 to 20 minutes of tabulating. The 2024 Model T Homecoming gave itself 30 minutes between its 1:30 ballot deadline and its 2 PM announcement. At the top end, Carlisle’s 2025 GM Nationals took participant ballots until 2 PM and posted winners at 6 PM, which works for a three-day national and would empty a Saturday show in a church lot.

Volume is what breaks paper. A Texas show, the Heartland Cruisers, reported close to 3,500 ballots in 2009 and ran out for a while. Our own reference point is the 2026 Labor Day Car Show in Fairfax, which ran on Car Show HQ: 3,574 votes across fifteen categories, with results final four minutes after voting closed. Researchers at Rice measured people taking 11.9 minutes to count one race on 120 pre-sorted ballots. Run the Labor Day count at that pace and it’s nearly six hours for one person, or about 45 minutes for eight, if none of them stops for a sandwich. (That’s our arithmetic, and a lab pace, so treat it as a sketch.)
Hand counts make mistakes
Nobody at a car show wants to hear this, but tired volunteers miscount. A Rice University study of election hand counts found error rates of half a percent to 1 percent when one person reads and another marks, and up to 2 percent with the sort-and-stack method. Researcher Michael Byrne’s summary: “It is probably impossible to completely eliminate errors in hand counting of ballots.” An earlier paper from the same group had 40 people count races on 120 ballots, a stack about the size of a small show’s, and 57.5 percent of the counts produced the correct result.
That was a lab, with ballots built to test counters, so don’t read it as a verdict on your treasurer. The practical version is smaller. A 1 percent error on 250 ballots is two or three votes, and a class at a small show can turn on that many. If you count by hand, have two people count the top two cars in every class separately. It adds five minutes and it’s the one step we’d never skip.
What QR voting asks of you
It asks for a signal, first of all. Car Show Pro’s voting documentation states the requirement plainly, a cellular-enabled phone or tablet, and its page on the Location Unknown message lists what goes wrong: location services switched off, permission denied, and poor coverage. Those are vendor docs, and they’d know.
A fairground on a Saturday isn’t the fairground on a Tuesday, either. Hundreds of phones share the same tower, so test from the back row during a comparable crowd if you can, not in the empty lot on a weekday. If there’s any doubt, keep a box of paper on hand. Car Show Pro offers an optional paper ticket for exactly this, tallied by hand, and we think that’s the honest setup for any tool. Car Show HQ has no offline mode, and a show on a thin signal should plan around that.

The second cost is people. Some of your crowd won’t vote on a phone: it’s dead, it’s in the truck, or the small print defeats them. We wrote a whole post on voting for people without a phone, because a volunteer with a tablet solves most of it, but you do have to staff that volunteer. A QR setup does save the ballot table crew, and it can’t run out of ballots. Spectators also vote where they stand instead of queuing at a box. The Labor Day show took more than 1,100 votes in its busiest hour, which would have been a long line.
Cheating, on both systems
Paper’s reputation here is earned. A 2015 CorvetteForum thread describes one club member filling out every ballot for the group before registration closed. Shows answer with tighter ballots: a Shelby club’s Pittsburgh show numbered each entrant’s voting tickets to the entry and threw out duplicates in any envelope, and the organizer reported little stuffing that year.
QR voting ends the stack of photocopied ballots. Car Show Pro limits each user to one vote per car, we limit each person to one per category, and a location check keeps voting on the grounds. It does nothing about the club that agrees over coffee to all vote for Dave’s truck, or the kids talking their parents into the loudest paint at the show. No format fixes friends, and we’d be suspicious of anyone who claims otherwise. Our post on voting formats goes further into who each one is fair to.
Picking one for your show
| Your show | We would use | Why |
|---|---|---|
| Under 75 cars, one or two awards | Paper | A 20-minute count fits, and there’s nothing to set up |
| Participants vote, spectators don’t | Paper | One ballot handed over at check-in is hard to abuse |
| 75 to 200 cars, several classes | QR, with a paper box as backup | Class-by-class counting is where paper gets slow |
| Over 200 cars, or ten or more categories | QR | The count alone would outlast your crowd |
| Rural field with one bar of signal | Paper | A location check you can’t trust is worse than a box |
If you try QR voting on Car Show HQ
We build a QR voting tool, so take our table with that in mind. On Car Show HQ spectators scan a code at the show, vote on their own phones, and the tally runs as votes arrive, so there’s no count at the end. It’s on-site only, and a staffed tablet takes ballots for anyone who can’t or won’t use a phone. A phone that can’t get a location fix isn’t turned away: it’s let through and counted as unverified, or under the strict setting a staff member scans a code and vouches for the person standing there. It’s in the free plan. The voting doc walks through setup.
Before your next planning meeting, multiply your expected voters by your categories, and divide by what two volunteers can count in the gap before awards. If that number scares you, go stand at the far end of your venue with your phone and see how many bars you get.


