$64.3 million changed hands in online auctions last week, across 1,377 cars. We ran the numbers through GoAskSam again, and while there were some pretty impressive cars changing hands, the stories that caught my attention weren't necessarily the most expensive ones.

The typical sale came in at $22,500, down slightly from $23,425 the previous week, while total sales dropped 3.9% from $66.9 million.

As I mentioned last time, we're now keeping the live auction houses separate, so these numbers are strictly for online auctions. Bring a Trailer accounted for 69% of the money and, rather impressively, every single car in the top ten.

We only had one car break the million-dollar mark this week, compared with three the week before, and that was a 2020 Lamborghini Aventador SVJ with 7k miles, which sold for $1,055,555.

It was followed by a 2020 Ford GT Carbon Series at $950,000, another 2005 Ford GT at $815,000, and a 2023 Ferrari 812 GTS at $801,000.

I also loved seeing a 2008 Spyker C8 Laviolette in the mix. Just 171 kilometres on the clock, and it brought $761,000. Not exactly a car you come across every day!

Three Ford GTs made the top ten, which together accounted for $7.3 million. So even in a slightly quieter week, there's clearly still plenty of money chasing the right cars.

But enough about the expensive stuff, because the next set of numbers is where things got particularly interesting.

You might remember that I've started looking at cars where the seller previously turned down a bid, only to come back to auction and eventually sell.

I find this fascinating because we're constantly debating whether sellers should accept the money in front of them or hold out for more. And for the third week running, we've got some interesting evidence.

Last week, 30 cars sold that had previously been to auction without selling. For 28 of them, we could identify the actual bid that was rejected, and here's the interesting part.

23 of those 28 eventually sold for more than the bid their seller had turned down.

Of the 11 cars that moved to a different auction platform, ten beat their previous bid. Even among the 17 that returned to the same platform, 13 managed to do better.

And these sellers weren't necessarily sitting around for months waiting for the market to change. The typical wait was just 1 1/2 months.

Those who came out ahead got roughly 13% more, while the five who didn't lost around 11% compared with the bids they'd rejected.

A few examples really bring this to life.

Take this 1958 Chevrolet Delray. It was bid to $23,250 on BaT back in July, but the seller said no. Last Wednesday it returned and sold for $42,001. That's an 81% improvement, and quite a reward for waiting.

Then there's this 2015 Mercedes-Benz SLS AMG GT Final Edition. It had been bid to $360,000 on Cars & Bids in June, but didn't sell. It subsequently appeared on BaT and brought $505,000. That's another $145,000, or 40%, compared with the rejected bid. Don’t think it was just going that way, see below of some examples going the other way.

Of course, not everyone got rewarded for holding out.

A 1969 Camaro Z/28 rejected $133,969 before eventually selling for $117,000, while a 1990 BMW 325is went from a rejected $18,000 to a completed sale at $14,750.

So there's definitely a risk in saying no, but here's the number I keep coming back to.

Over the three weeks we've now tracked, 70 out of 92 cars with identifiable rejected bids eventually sold for more. That's roughly three quarters.

I'm not suggesting sellers should automatically reject bids and try again. Far from it. We're only looking at cars that eventually sold, so this doesn't tell us what happened to those that never found a buyer.

But it does make me question the idea that a failed auction necessarily tells you what the market is willing to pay.

Sometimes it might simply mean you haven't found the right buyer yet.

This naturally got me looking at another group of cars in GoAskSam, those that had previously sold on one platform and then turned up on another.

There were 13 of them last week, with six selling for more and seven for less. So it's certainly not a case of moving platforms and automatically making more money.

Interestingly, for the second week running, the most common move was from Bring a Trailer to Cars & Bids. Seven cars made that journey, four selling higher and three lower.

But a couple of the moves in the opposite direction caught my attention.

A 2003 Acura NSX-T had sold for $145,000 on Cars & Bids in 2023. It appeared on BaTast Monday and made $175,500, a 21% increase.

A 2017 Mercedes-AMG GT S went the other way, having previously sold for $63,000 on BaT before bringing $68,500 on Cars & Bids.

And then there was a 1985 Mercedes-Benz 300CD Turbo that had sold for $25,750 on Cars & Bids in September, only to sell for $20,000 on BaT.

What I find interesting here is how differently these cars can perform. There are obviously lots of variables, including condition, presentation, timing and who's actually bidding, so we can't put every price change down to the platform.

But it's another reminder of why I think sellers should look at the actual results for cars like theirs before deciding where to list.

It's one of the things we've been trying to make easier with the GoAskSam Seller Intelligence tool, rather than expecting everyone to spend hours digging through old auction results.

While we're on the subject of cars coming back to auction, we also looked at the ones that had actually sold before, rather than simply failed to meet reserve.

There were 114 of these last week, and the results were pretty evenly divided at first glance. Of those, 48 sold for more, 65 for less and one for exactly the same amount.

The typical car was 2.2% down on its previous sale, having been away from the auction market for around 20 months.

That's actually an improvement on the previous week, when the typical result was down 5.3%.

But what I really wanted to understand was whether the amount of time between sales made a difference.

And this is where the quick flips continue to stand out.

Cars that returned to auction within six months were down 4.3% on their previous sale. That's now three consecutive weeks where this group has lost money.

The first week they were down 9%, the second week 6.6%, and now 4.3%.

The longer-held cars are much less consistent. Some groups are slightly up, others slightly down, and this week even the cars held for more than three years were down 4.7%.

I find the quick-flip numbers particularly interesting because people often assume that a desirable car can simply be bought, relisted and sold for more a few weeks later.

Sometimes that works brilliantly, but the numbers we've seen so far suggest it's not quite that easy.

And remember, these are sale-price comparisons. Once you add the costs of buying, selling, transporting and preparing a car, the actual economics of a quick flip can be considerably worse.

Now, three weeks certainly isn't enough to declare that quick flips are a bad idea. Last week the longer-held cars were the problem, and this week it's different again.

That's actually why I want to keep doing this. Rather than taking one week of results and building a grand theory around it, let's see what the numbers look like after a few months.

One last thing before I let you get on with your day.

We had two 2005 Ford GTs sell on Bring a Trailer just a day apart, and the difference between them was pretty remarkable.

The first had just 2,000 miles and brought $815,000. The second was a one-owner car with 45,000 miles and sold for $480,000.

Same model, same year, one day apart, but $335,000 between them.

Now, 43,000 miles is an enormous difference on a car like this, so I'm not suggesting these two should have sold for anything close to the same money.

But I love examples like these because they show why looking at a single headline sale can be so misleading.

It was the same story further down the price ladder.

A 1995 Acura NSX-T with 40,000 miles made $143,000, while another with 194,000 miles sold two days later for $77,000.

And a 2002 BMW M5 with 112,000 miles brought $37,000, compared with $15,600 for another showing 245,700 miles.

We tend to talk about what a particular model is selling for, but the reality is that two cars that look almost identical in a headline can be very different propositions once you get into the details.

And that's really what much of this week's analysis comes down to.

The same car can fail to sell and then beat its old bid a few weeks later. It can change platforms and do better, or worse. It can come back to auction just months after selling and find an entirely different result.

The market isn't always giving us one clear answer, and that's what makes following the actual sales so interesting.

We'll keep running the numbers through GoAskSam each week, particularly the cars that have come back to auction, and see whether some of these patterns hold up over time.

There's still a lot to learn, and I'm looking forward to seeing what next week's results tell us.