[privacy] How many Bitcoin chips are out there?

[LoyceV]

ChipMixer started in 2017. It provides privacy by dividing Bitcoin into indistinguishable chips of the same size, starting with 0.001 BTC, and doubling the size up to 4.096 BTC (and since 2018 up to 8.196 BTC).
Chips existed long before ChipMixer, and anyone can create them on their own. For a while now, I've been curious how many chips are out there. The more there are, the better privacy they offer. So I collected the data

I took Blockchair's Bitcoin data, and used the following assumptions:

• Any address with a deposit of 0.001, 0.002, 0.004, 0.008, 0.016, 0.032, 0.064, 0.128, 0.256, 0.512, 1.024, 2.048, 4.092 or 8.196 BTC can be a chip.
• Any address that received more than one deposit is not a chip.
• Any address that sent more than one transaction is not a chip.
• Chips are counted if they're still funded at the end of a day.
• Chips are not counted from the day they're emptied.

The first chip
The first chip was created on April 29, 2011, with a size of 0.016 BTC.

Full data
As far as I know there are no false positives or negatives in my data. If anyone's interested in the full data (27 GB of "which chip exists on which day"), let me know.

Graphs
To keep the number of graphs manageable, I'll limit myself to Legacy addresses (starting with 1: solid lines) and native Segwit addresses (starting with bc1q: dotted lines).
Note: check the scale for each graph, I zoom in on the vertical axis to better show the larger chips.


















(0.001/4.096 and 0.002/8.196 have the same color, but it should be obvious based on the numbers which line is which)

No frequent updates
Currently, my data ends on August 20, 2022.
Processing the data takes a while, and updating the graphs is a lot of work. So don't expect too frequent updates, maybe in a few years, or maybe when Taproot becomes more popular.

No spam
Self-moderated against spam. Discussion is of course welcome.

[1714524959]

1714524959

[1714524959]

1714524959

[1714524959]

1714524959

[PawGo]

As far as I know there are no false positives or negatives in my data. If anyone's interested in the full data (27 GB of "which chip exists on which day"), let me know.

Do not you think the amount 0.001 or 0.002 could be so popular that connecting it with chipmixer transfer is a bit risky? Or course we may say it is a candidate because that amount was used, but I think we may have many many uncorrelated transactions.

How about source addresses, any duplicates?

[LoyceV]

due to instability on bigger chip i'd rather read table for short summary.

I can do that:

(sorry, I used Dutch date format)

I also made a 30 days running average, that should smoothen out the graphs:



(again: Dutch date format)

Full data
As far as I know there are no false positives or negatives in my data.

Now i wonder if there's rare occurrence someone accidentally mimic chip behavior. For example, user use one-time address practice and withdraw exactly 0.001BTC * 2^N

I'm pretty sure that's what happened thousands of times, before ChipMixer even existed. But what I meant with "no false positives" is that I hope I didn't miss anything according to my own definition of a chip.

Do not you think the amount 0.001 or 0.002 could be so popular that connecting it with chipmixer transfer is a bit risky? Or course we may say it is a candidate because that amount was used, but I think we may have many many uncorrelated transactions.

What do you mean by "risky"? I didn't mean to claim how many of those addresses are related to ChipMixer, but that's the beauty of chips: anyone can create them.

How about source addresses, any duplicates?

I didn't check source addresses, but I've seen many chips get funded at once in one transaction.

[buwaytress]

Cool, actually thought about it myself some time ago, but couldn't have imagined it was THIS many chips (sure, probably some false positives in there, but even halving the amounts exceeds what I'd have personally thought were the # of chips).

What's interesting is about 1/3 of chips are in segwit until you get to 0.016, then it becomes about half... while at the highest amount, almost 2/3 of chips are segwit.

[Little Mouse]

The more there are, the better privacy they offer. So I collected the data

Technically yes but I don’t think this has a lot of things to do. How many people would put the Chips from ChipMixer idle in the address? We don't have any data but I think most of them are spent from the Chips address. And if they are spent, this is unlikely that these chips will be exactly 0.001 BTC or 0.002 BTC and so on.
That's my perspective though. Who knows what people do with their mixed coin? Maybe they keep them idle. I would spend from the source address at least.

[LoyceV]

How many people would put the Chips from ChipMixer idle in the address? We don't have any data but I think most of them are spent from the Chips address.

Actually, we do have the data It's just that it's a lot of work to get to it, but it's totally possible to check how many chips were funded and emptied (per day).

[LoyceV]

It's rather surprising 0.256BTC chip suddenly grow >20x in 2022. It makes 0.512BTC chip looks far less popular than 0.256BTC and 1.024BTC chip on 2022.

The difference isn't that much, if you add Legacy and Segwit numbers:

July 1, 2021: 709+20=729 chips of 0.256 BTC
July 1, 2022: 389+2818=3207 chips of 0.256 BTC (4.4 times larger)

July 1, 2021: 646+5=651 chips of 0.512 BTC
July 1, 2022: 289+830=1119 chips of 0.512 BTC (1.7 times larger)

July 1, 2021: 381+11=392 chips of 1.024 BTC
July 1, 2022: 243+991=1234 chips of 1.024 BTC (3.1 times larger)

[DaveF]

I can say that at least 100 of the 0.001 are from me. They are the Dave and Julie DIY Bitcoin Penny Coins.

Yes I know a penny should be 0.01 but give them away to people so deal with it :-) or send me a bunch of BTC and the next time I am making things to give away I'll fund them with some of your money...

No real set time when they were made but were always done in batches of 5 or 10 usually just before events or gatherings and no idea how many still exist.

-Dave

[100bitcoin]

I took Blockchair's Bitcoin data, and used the following assumptions:

• Any address with a deposit of 0.001, 0.002, 0.004, 0.008, 0.016, 0.032, 0.064, 0.128, 0.256, 0.512, 1.024, 2.048, 4.092 or 8.196 BTC can be a chip.
• Any address that received more than one deposit is not a chip.
• Any address that sent more than one transaction is not a chip.
• Chips are counted if they're still funded at the end of a day.
• Chips are not counted from the day they're emptied.

Did you create a tool to find this with different variations? For example, I would like to know, how many chips with 1 BTC value are out there. Can something like the following URL be formulated for this calculation?

https://loyce.club/chipnumber.php?chipsize=1

[LoyceV]

Did you create a tool to find this with different variations?

More or less

For example, I would like to know, how many chips with 1 BTC value are out there.

Do you want only current data, or also historic data?

Can something like the following URL be formulated for this calculation?

Nope, I don't do PHP. I could run the script again with for 1000 mBTC though (but it's quite slow, it has to decompress all data). I've started it, if it doesn't fail I'll post the results tomorrow.

[DaveF]

I understand the 1 2 3 8 16 32 64 numbering.

But it makes you wonder why CM did it. I am sure there are other BTC0.128 chips out there, but I would imaging that there are several orders of magnitude more addresses with BTC0.1

Just for anonymity using more common amounts would be logical. With that in mind @LoyceV how difficult would it be to see how many BTC0.1 BTC0.2 BTC0.5 BTC1.0 and BTC2.0 addresses that meet the criteria.
i.e. I don't care about an address with BTC1 if it got 4 deposits go get there, so the same as how CM would have done it vs their amounts.

I could probably download your data and have one of the programmers I deal with whip something up, but if you have something close to it already it might not be worth it to have them do it.

Thanks,
Dave

[LoyceV]

But it makes you wonder why CM did it.

To <split> and <merge>

I am sure there are other BTC0.128 chips out there, but I would imaging that there are several orders of magnitude more addresses with BTC0.1

Maybe. By being big enough, CM created a market for 0.128 chips that wasn't there before.

Just for anonymity using more common amounts would be logical. With that in mind @LoyceV how difficult would it be to see how many BTC0.1 BTC0.2 BTC0.5 BTC1.0 and BTC2.0 addresses that meet the criteria.

I've added it to the script, and restarted. If it doesn't fail (I didn't do any thorough checking), I should have results by tomorrow.

[LoyceV]

Update: no results yet. I'm not sure why, but something ran out of memory. I'm trying again

[RocketSingh]

But it makes you wonder why CM did it.

To <split> and <merge>

Care to explain the advantage of using BTC0.128 over BTC0.1 to split and merge?

0.128/2 => 0.064/2 => 0.032/2 => 0.016/2 => 0.008/2 => 0.004/2 => 0.002/2 => 0.001

0.1/2 => 0.05/2 => 0.025/2 => 0.0125/2 => 0.00625/2 => 0.003125 => 0.0015625/2 => 0.00078125

Both works. No?

[LoyceV]

Care to explain the advantage of using BTC0.128 over BTC0.1 to split and merge?

You've answered your own question: 0.001 is more convenient than 0.00078125. And it's much easier to do the calculation in your head (most computer users should know powers of 2 up to quite a high number by heart anyway).

If you'd start at 0.1 and go the other direction, you end up with "weird" amounts again: 0.2, 0.4, 0.8, 1.6, 3.2 and 6.4. I'm much more used to 8.192 (from computer memory) than 6.4.

>> this goes off-topic for this topic <<

Still crunching numbers, I forgot how slow this is.

[casinotester0001]

#

instead of doubling the chip value each time you could use denominations like 0.01, 0.02, 0.05.

Could you extend your idea? How would it work with split/merge/bet function?

Would work too. (Maybe easier for humans who use money because of similar values?)  

existing chips:
0.001, 0.002, 0.005, 0.010, 0.020, 0.050, 0.100, 0.200, 0.500, 1.000, 2.000, 5.000

value -> split into
0.001 -> 0.001
0.002 -> 0.002
0.003 -> 0.002 + 0.001
0.004 -> 0.002 + 0.002
0.005 -> 0.002 + 0.002 + 0.001
0.006 -> 0.005 + 0.001
0.007 -> 0.005 + 0.002
0.008 -> 0.005 + 0.002 + 0.001
0.009 -> 0.005 + 0.002 + 0.002
0.010 -> 0.005 + 0.005
0.011 -> 0.005 + 0.005 + 0.001
0.012 -> 0.005 + 0.005 + 0.002
0.013 -> 0.005 + 0.005 + 0.002 + 0.001
0.014 -> 0.005 + 0.005 + 0.002 + 0.002
0.015 -> 0.005 + 0.005 + 0.005
0.016 -> 0.005 + 0.005 + 0.005 + 0.001
...
0.980 -> 0.500 + 0.200 + 0.200 + 0.050 + 0.020 + 0.010
...
0.990 -> 0.500 + 0.200 + 0.200 + 0.050 + 0.020 + 0.020
...
1.000 -> 1.000
1.001 -> 1.000 + 0.001
...

[o_e_l_e_o]

It kind of misses the point though.

It is trivial to identify an output as coming from ChipMixer. Take a look at this address: https://mempool.space/address/bc1qu225m44ere7sy89x5z0qhrhp7ma0yttrpwnyuh. Now, you can follow the coins on that address through dozens of transactions. Each transaction creates 30 outputs of the same chip size, and one change output. This chain of transactions creates hundreds of outputs for ChipMixer chips in total. There is no other service out there which does anything like this. It is completely unique and trivial to identify. This would not change if the chip sizes were 0.001/0.002/0.005/0.010/etc. instead of 0.001/0.002/0.004/0.008/etc.

The whole point of ChipMixer is to break the link between coins you deposit and coins you withdraw. It is not to hide the fact that you have used a mixer, and indeed, often the exact opposite. Being able to show quite clearly that all my inputs came from a mixer means I can prove that any claims you make about my coins being "tainted" or any other such nonsense are provably false. This same logic applies to coinjoins.

[DaveF]

It kind of misses the point though.

It is trivial to identify an output as coming from ChipMixer. Take a look at this address: https://mempool.space/address/bc1qu225m44ere7sy89x5z0qhrhp7ma0yttrpwnyuh. Now, you can follow the coins on that address through dozens of transactions. Each transaction creates 30 outputs of the same chip size, and one change output. This chain of transactions creates hundreds of outputs for ChipMixer chips in total. There is no other service out there which does anything like this. It is completely unique and trivial to identify. This would not change if the chip sizes were 0.001/0.002/0.005/0.010/etc. instead of 0.001/0.002/0.004/0.008/etc.

The whole point of ChipMixer is to break the link between coins you deposit and coins you withdraw. It is not to hide the fact that you have used a mixer, and indeed, often the exact opposite. Being able to show quite clearly that all my inputs came from a mixer means I can prove that any claims you make about my coins being "tainted" or any other such nonsense are provably false. This same logic applies to coinjoins.

Yes and no. I have made many transactions that create a lot of .001 (filling collectables) and back in the day when BTC was a lot less I made a lot of BTC1.x  into a bunch of BTC0.1 coins. So although not as obvious a the 30 groups that CM makes there are probably (waiting on loyces program to finish running) a lot more of those then .128 which was why I bought up the point of probably orders of magnitude more.

More a thought experiment then anything else, but still going to interesting to see what the results are.

-Dave

[casinotester0001]

It is completely unique and trivial to identify. This would not change if the chip sizes were 0.001/0.002/0.005/0.010/etc. instead of 0.001/0.002/0.004/0.008/etc.

I meant the value, not the privacy/mixing ability. It would be easier to calculate, at least for me 

[o_e_l_e_o]

So although not as obvious a the 30 groups that CM makes there are probably (waiting on loyces program to finish running) a lot more of those then .128 which was why I bought up the point of probably orders of magnitude more.

I'm certain you will be right, and there will be many more 0.1 BTC outputs than there will be 0.128 BTC outputs. That doesn't change my point, though. It's not just the individual transaction which creates 30x 0.1/0.128 BTC outputs, but rather you can follow the change output forward or backward and see transaction after transaction after transaction creating 30 outputs of various values. Although a 0.1 BTC output may look less obvious on first glance, to anyone who looks back even a single transaction it will still be immediately obvious it is coming from ChipMixer (and of course it is the people who look back at the history of a coin who are the very people you are protecting against by mixing).

It would be easier to calculate, at least for me 

I don't think it makes much difference. You can still withdraw 0.010 BTC with a 0.008 and a 0.002 chip, if you want. I like the powers of 2.