gb.uk

[trendmedia]

gb.uk ... interesting 2 character drop at DL auction...

What could it be worth?

4 to 5 figures surely? But in a domainer auction? who knows?

 

[Opsuk]

I missed this by 0.2 milliseconds (0.002 seconds) on my drop catch server. Very close. I had been hoping to get it. It went for £15,501. Apparently the previous owner got it for £50.

 

[GradeZero]

I missed this by 0.2 milliseconds (0.002 seconds) on my drop catch server. Very close. I had been hoping to get it. It went for £15,501. Apparently the previous owner got it for £50.

Unlucky. Reading posts like yours makes me want to get my own tag.

 

[alanalan is verified member.]

I missed [gb.uk] by 0.2 milliseconds (0.002 seconds) on my drop catch server. Very close. I had been hoping to get it. It went for £15,501. Apparently the previous owner got it for £50.

You mention on another thread:

I have missed a few recently, mostly by a few milliseconds. gb.uk was one I was particularly keen on. I missed it by 0.006 seconds (0.6 of a millisecond). Very close.

@Opsuk I'm curious, what exactly do the 2ms and 6ms values represent, and how have you worked out that you missed the catch by those margins?

(0.002 s = 2 ms, and 0.2 ms = 200 µs, so which measurement is it?)

 

[SamSam is verified member.]

I missed this by 0.2 milliseconds (0.002 seconds) on my drop catch server. Very close. I had been hoping to get it. It went for £15,501. Apparently the previous owner got it for £50.

Where did you here that they didn’t buy it for £50

 

[deedee is verified member.]

You mention on another thread:


@Opsuk I'm curious, what exactly do the 2ms and 6ms values represent, and how have you worked out that you missed the catch by those margins?

(0.002 s = 2 ms, and 0.2 ms = 200 µs, so which measurement is it?)

You're a dropcatcher . You know the difference, the round trip ... and the moving target lol

 

[Opsuk]

You mention on another thread:


@Opsuk I'm curious, what exactly do the 2ms and 6ms values represent, and how have you worked out that you missed the catch by those margins?

(0.002 s = 2 ms, and 0.2 ms = 200 µs, so which measurement is it?)

 

[Opsuk]

Sorry, correction, I missed by 0.0006 seconds which is 0.6 of a million second.

I have an automated drop catch system which fires requests 6 times at around 10ms intervals after the due drop time. I record the time it fires and I can see the time it was registered by the successful registrar. I am often a few milliseconds out. This time I was much closer.

Timing Analysis​

ATTEMPTS SPAN RECORDED REGISTRATION

The recorded registration timestamp is used for comparative timing analysis. It should not be treated as the exact instant another registrar's EPP CREATE request reached Nominet.

Registration delta from expected drop+60.779ms
Earliest attempt vs registration-38.569ms
Nearest attempt vs registration-0.600ms
Latest attempt vs registration+9.404ms
Attempts before registration5
Attempts after registration1

So yes, it really was that close. It was sold for £15k. Gutted!

You mention on another thread:


@Opsuk I'm curious, what exactly do the 2ms and 6ms values represent, and how have you worked out that you missed the catch by those margins?

(0.002 s = 2 ms, and 0.2 ms = 200 µs, so which measurement is it?)

 

[Opsuk]

Where did you here that they didn’t buy it for £50

AI. I checked it again and it tells me it can’t find the source. It thinks it may have been confusing the listing price on domainlore. Sounds reasonable.

 

[alanalan is verified member.]

Sorry, correction, I missed by 0.0006 seconds which is 0.6 of a million second.

I have an automated drop catch system which fires requests 6 times at around 10ms intervals after the due drop time. I record the time it fires and I can see the time it was registered by the successful registrar. I am often a few milliseconds out. This time I was much closer.

Timing Analysis​

ATTEMPTS SPAN RECORDED REGISTRATION

The recorded registration timestamp is used for comparative timing analysis. It should not be treated as the exact instant another registrar's EPP CREATE request reached Nominet.

Registration delta from expected drop+60.779ms
Earliest attempt vs registration-38.569ms
Nearest attempt vs registration-0.600ms
Latest attempt vs registration+9.404ms
Attempts before registration5
Attempts after registration1

So yes, it really was that close. It was sold for £15k. Gutted!

Thanks for sharing. I asked because I suspected you were measuring the wrong thing. You haven't accounted for round-trip time, which includes network and registry latency. Comparing your request timestamps with the registration timestamp won't tell you anything meaningful.

 

[Opsuk]

Where did you here that they didn’t buy it for £50

Sorry, my mistake. I just found it. It wasn't £50 it was £70. Someone bid Jon DomainLore.co.uk, £70 to drop catch gb.uk. Therefore it cost them £70 but they managed to resell it for £15k.

DomainLore: .uk Domain Sales and Auctions

DomainLore: .uk Domain Sales and Auctions, Domain Names Aftermarket, Backorder Expiring Domains, UK Droplist

domainlore.uk

 

[SamSam is verified member.]

Sorry, my mistake. I just found it. It wasn't £50 it was £70. Someone bid Jon DomainLore.co.uk, £70 to drop catch gb.uk. Therefore it cost them £70 but they managed to resell it for £15k.

DomainLore: .uk Domain Sales and Auctions

DomainLore: .uk Domain Sales and Auctions, Domain Names Aftermarket, Backorder Expiring Domains, UK Droplist

domainlore.uk

That’s a proxy bid of £70, not an indication that it would sell for £70. It was always going to fetch a good price.

 

[Opsuk]

Thanks for sharing. I asked because I suspected you were measuring the wrong thing. You haven't accounted for round-trip time, which includes network and registry latency. Comparing your request timestamps with the registration timestamp won't tell you anything meaningful.

Thank you. Surely if I time the moment I request it and compare the time it was actually registered, that would tell me how close I was. I can see that there would still be some latency in there but surely the time registered - the time requested would show the maximum difference. It my difference was 0.6ms, then I must have been even closer from my request to the actual drop.

Any advice you have on increasing my accuracy is much appreciated. I was catching 10-20% of the names I requested although than has decreased a bit recently.

 

[deedee is verified member.]

Sorry, my mistake. I just found it. It wasn't £50 it was £70. Someone bid Jon DomainLore.co.uk, £70 to drop catch gb.uk. Therefore it cost them £70 but they managed to resell it for £15k.

DomainLore: .uk Domain Sales and Auctions

DomainLore: .uk Domain Sales and Auctions, Domain Names Aftermarket, Backorder Expiring Domains, UK Droplist

domainlore.uk

That's not what it will have gone for... that was pre bid probably to push for an auction if caught.

 

[alanalan is verified member.]

Thank you. Surely if I time the moment I request it and compare the time it was actually registered, that would tell me how close I was. I can see that there would still be some latency in there but surely the time registered - the time requested would show the maximum difference. It my difference was 0.6ms, then I must have been even closer from my request to the actual drop.

Any advice you have on increasing my accuracy is much appreciated. I was catching 10-20% of the names I requested although than has decreased a bit recently.

It's a competitive space, so I'll keep the advice fairly general.

Your request timestamp records when you initiated the request, not when Nominet received or processed it. There are several factors to consider, including local execution time, network latency, jitter, registry queuing and processing time.

You can measure your round-trip response times to gain some insight into these delays, although you can't easily isolate each component. Registry latency can be highly variable, particularly around competitive drops.
For gb.uk specifically, I observed response times increased to over a second during the drop, with some exceeding 2 seconds. That's not unexpected for an LL .co.uk domain.

The registration timestamp doesn't tell you when the winning request was received or processed, so the 0.6ms difference doesn't mean you were within 0.6ms of catching it.

And that's assuming your clock is properly synchronised. There are plenty of other factors at play too.

 

[Opsuk]

I know. The customer would have paid Domainlore to catch it. It will have cost them £70. DomainLore would have paid £3.50 to Nominet for the caught domain. So effectively, the customer would have paid £70 for it.

 

[DomainPortfolio.ukDomainPortfolio.uk is verified member.]

I know. The customer would have paid Domainlore to catch it. It will have cost them £70. DomainLore would have paid £3.50 to Nominet for the caught domain. So effectively, the customer would have paid £70 for it.

That's not how it works, the prices you see are just pre drop bids that people place, it would never sell for £70. If there's a pre drop bid for any domain and it's caught by Denys or someone else using his catching script, it goes to auction, at no point is a pre drop bid an indication that the domain will sell for that price.

People often place bids so that they get a reminder of the auction when they're outbid.

 

[Opsuk]

I realise that. I am just saying that it cost the buyer who resold it, £70. It cost Denys £3.50.

 

[deedee is verified member.]

I realise that. I am just saying that it cost the buyer who resold it, £70. It cost Denys £3.50.

It didn't.... it cost the buyer the final auction price. I doubt £70 was the final price

We need seemly..... he had all that stuff logged didnt he?

 

[Opsuk]

It looks like £78 was the final auction price for the drop-catch. https://domainlore.uk/#tabs-4