Does BIP39 allow for the same word to be generated more than twice in a seed?

[1980sFuture]

Hello there. I have a slightly more technical question for somebody knowledgeable about BIP39 standards and the rules involved.

During 24 word seed generation, when a wallet randomly generates your seed, I know it's possible for a word to be repeated in the seed - roughly a 12% chance from what I understand. But is it possible that the same word gets used THREE times during a 24 word seed generation? Obviously this would be even more unlikely than twice, but I'm just wondering if it's possible and what the chances are - or does BIP39 have rules that prevent the same word from being generated 3 times?

Appreciate any insight and answers I can get on this. Thankyou.

[1714502745]

1714502745

[1714502745]

1714502745

[Orpichukwu]

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase. BIP39 seed phrases usually consist of 12, 18, or 24 words chosen randomly from a predefined list of 2000+ words. There are over 30 trillion possible combinations of these words, so there is no need to repeat any word more than twice.

I don't know if those figures are exact, but it's something close, as I can't recall where I read about it but am certain of the content.

I have created more than 10 wallets this year, and I have not come across a single word repeated even twice, which is barely possible—talk more than three times.

[decodx]

does BIP39 have rules that prevent the same word from being generated 3 times?

No, I don't think so. There's nothing to stop the same word occurring more than once. But having the same word three times or more in a 24-word seed phrase is highly unlikely.

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase.

I don't think that is correct. Do you have a source?

I have created more than 10 wallets this year, and I have not come across a single word repeated even twice, which is barely possible—talk more than three times.

That doesn't prove it's impossible, just improbable.

[1980sFuture]

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase.

So you're saying there's actual code baked into BIP39 that will prevent a 3rd repetition if a word is already used twice? If this is indeed correct, then why do wallets accept phrases with multiple repetition? Wouldn't they be invalid if a seed contained 3 or more repetitions of a word + a valid checksum were used?

Andreas says in this video that you could use the same word 23 times + valid checksum and it would still be valid. If that's a mathematically valid wallet, then what's to stop a wallet from generating a seed with 3 of the same word? I'm not trying to get into the chances so much as if it's possible.

https://www.youtube.com/watch?v=kGClNWZcYV8

[Findingnemo]

does BIP39 have rules that prevent the same word from being generated 3 times?

There is no phenomenon that stops words from repeating more than once which can be twice, thrice, or even more while generating seeds based on BIP39.

OP your question has been answered here

The probability of 11.7% is being calculated using the birthday paradox formula.

"The paradox of birthdays is a mathematical problem put forward by Von Mises, who looks for the value N in the problem: In a group of N people there is 50% chance that at least 2 people in the group share the same birthday (day + month). The answer is N = 23, which is quite counterintuitive, hence the paradox."

So for the calculations: -Repetition (Twice): 2 (number of people sharing the same birthday) -Number of words: 24 (number of people in a group) -Number of all BIP39 words: 2048 (number of days in a year)

The result is 0.119 (11.9%). But for THREE repetitions it's getting really low: 0.00043 (0,043%).

Andreas says in this video that you could use the same word 23 times + valid checksum and it would still be valid.

Yes, it will still be a valid wallet but if you are doing such a thing then the time taken to brute force your wallet will be 0.001 sec.

[hosseinimr93]

Whenever you generate a BIP39 seed phrase, you actually generate a random number and your seed phrase represents that random number.

For example, if you generate a 24 word BIP39 seed phrase, the random number can be any number between 0 and 2²⁵⁶-1. This means that any of the words can be repeated multiple times.

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase.

This is completely wrong.

[1980sFuture]

Whenever you generate a BIP39 seed phrase, you actually generate a random number and your seed phrase represents that random number.

If you generate a 24 word BIP39 seed phrase, the random number can be any number between 0 and 2²⁵⁶-1.

I totally understand that. I know that the words are just a more manageable way of expressing a huge number. I was just wondering if - although unlikely - it was possible for a wallet to generate a 24 word seed that contains 3 of the same word.

This means that any of the words can be repeated multiple times.

So multiple as in 3 or more times is technically possible (although exceedingly rare), correct?

does BIP39 have rules that prevent the same word from being generated 3 times?

OP your question has been answered here

Thanks
Yeah I saw this. But I Wasn't sure if the answer was just strictly talking about the chances of it happening in terms of mathematics - or if it was possible within the boundaries of BIP39 wallet generation.

Yes, it will still be a valid wallet but if you are doing such a thing then the time taken to brute force your wallet will be 0.001 sec.

Right obviously that would be insane.

[hosseinimr93]

I was just wondering if - although unlikely - it was possible for a wallet to generate a 24 word seed that contains 3 of the same word.

Yes, that's possible.
(In theory) It's even possible that you generate a 24 word seed phrase and a single word is repeated 24 times.

The probability of a single word is repeated n times decreases exponentially with increase in n and that's why it's unlikely that you generate a seed phrase in which a single word is repeated 3 times, but as I said that's possible.


This post has been edited. Thanks LoyceV for the correction.

[1980sFuture]

I was just wondering if - although unlikely - it was possible for a wallet to generate a 24 word seed that contains 3 of the same word.

Yes, that's possible.
It's even possible that you generate a 24 word seed phrase and a single word is repeated 24 times.

The probability of a single word is repeated n times decreases exponentially with increase in n and that's why it's unlikely that you generate a seed phrase in which a single word is repeated 3 times, but as I said that's possible.

Thanks I appreciate the insight. 

[Orpichukwu]

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase.

This is completely wrong.

It's good this thread was created; at least what I was carrying around as knowledge is actually wrong.

This leaves me with this question. There is a higher possibility of words being repeated in a seed phrase, or 24 words. Let's assume up to 3–4 random words are repeated to make up those 24 words. won't it make it a bit easier for such a seed phrase to easily be guessed?

I know that even if one gets all the words correctly, the possibility of arranging them accordingly is also really small, but still, repeated words make me think otherwise. Forgive my curiosity by just asking to be sure.

[Zaguru12]

Yeah I saw this. But I Wasn't sure if the answer was just strictly talking about the chances of it happening in terms of mathematics - or if it was possible within the boundaries of BIP39 wallet generation.

There is chance although the probability isn’t high because logically picking a 12 or 24 series of words from a possible 2048 will widen the gap of it happening and that’s why we don’t see them occurring regularly but that doesn’t mean it isn’t. The thing stopping a group of words in right numbers from been a seed phrase either included a repeated words is if the checksum (which is contain in the last word) is not correct.

BIP39 seed phrases usually consist of 12, 18, or 24 words chosen randomly from a predefined list of 2000+ words.

It’s a total of 2048 from common words excepts the word Satoshi

[hosseinimr93]

This leaves me with this question. There is a higher possibility of words being repeated in a seed phrase, or 24 words. Let's assume up to 3–4 random words are repeated to make up those 24 words. won't it make it a bit easier for such a seed phrase to easily be guessed?

Take note the the seed phrase is not generated by selecting words. For a generating a seed phrase, a random entropy is generated and that's encoded to a series of words.

If the seed phrase has been generated in the correct way, there's nothing to worry about.
The chance of a word repeating 3 or 4 times is very very low and even if that happens, there's nothing to worry about in the case the seed phrase has been generated completely randomly.

[o_e_l_e_o]

There is absolutely nothing stopping the same word for appearing multiple times in a seed phrase. I've answered this before:

For a 24 word seed, there is a 12.65% of the same word appearing twice or more, and a 0.0479% of the same word appearing three times or more. So one in every 2,000 seed phrases roughly would have the same word three times. So not that unlikely by any means.

Here are some more numbers:

The chance of a 12 word seed having the same word at least twice : 3.18% (1 in 31)
The chance of a 12 word seed having the same word at least three times: 0.0052% (1 in 19,128)
The chance of a 12 word seed having the same word at least four times: 0.0000057% (1 in 17,407,725)

The chance of a 24 word seed having the same word at least twice: 12.65% (1 in 8)
The chance of a 24 word seed having the same word at least three times: 0.048% (1 in 2,089)
The chance of a 24 word seed having the same word at least four times: 0.00012% (1 in 814,729)

This leaves me with this question. There is a higher possibility of words being repeated in a seed phrase, or 24 words. Let's assume up to 3–4 random words are repeated to make up those 24 words. won't it make it a bit easier for such a seed phrase to easily be guessed?

No. There isn't a "higher possibility" of words being repeated - there is the exact same possibility as any other word. If my first word is "zoo", then my second word has the exact same possibility to be "zoo" as it does to be any other specific word. Indeed, if we introduced code to prevent a word from appearing more than once, then you would actually be decreasing the final entropy of your seed phrase, since by the time you get to the 23rd word in your seed phrase, you no longer have 2,048 possibilities for that word but only 2,026, since you could not repeat any of the 22 words you had already used.

[decodx]

The BIP39 standard doesn’t allow for any word to be repeated more than two times in a seed phrase.

This is completely wrong.

It's good this thread was created; at least what I was carrying around as knowledge is actually wrong.

In that case, you need to edit your first reply so that you don't pass your wrong knowledge on to someone else.

This leaves me with this question. There is a higher possibility of words being repeated in a seed phrase, or 24 words. Let's assume up to 3–4 random words are repeated to make up those 24 words. won't it make it a bit easier for such a seed phrase to easily be guessed?

Absolutely not. It's virtually impossible for anyone to guess the presence of a repeated word in a seed phrase. Therefore, a seed phrase with one or two repeating words is just as robust as a seed phrase with entirely unique words. Actually, I believe the opposite is true. Preventing duplicate words would, in fact, restrict the entropy of the seed phrase, making it weaker.

[apogio]

Just as an example, the following is a valid seed phrase:

Code:

office smoke grid creek office smoke grid creek office smoke grid credit

It's produced by the following entropy

Code:

10011001100110011001100110011001100110011001100110011001100110011001100110011001100110011001100110011001100110011001100110011001

As you can see the pattern is repetitive. Of course I produced it manually but every entropy generator must have the exact same probability for each word.

As oeleo said above, each word has the exact same possibility of appearing next, which is 1 out of 2048

By the way, the following is also valid:

Code:

zoo zoo zoo zoo zoo zoo zoo zoo zoo zoo zoo wrong

[DaveF]

Code:

zoo zoo zoo zoo zoo zoo zoo zoo zoo zoo zoo wrong

Also sounds like a kid in the back seat and the adult in the front.

...
The chance of a 12 word seed having the same word at least twice : 3.18% (1 in 31)
The chance of a 12 word seed having the same word at least three times: 0.0052% (1 in 19,128)
The chance of a 12 word seed having the same word at least four times: 0.0000057% (1 in 17,407,725)

The chance of a 24 word seed having the same word at least twice: 12.65% (1 in
The chance of a 24 word seed having the same word at least three times: 0.048% (1 in 2,089)
The chance of a 24 word seed having the same word at least four times: 0.00012% (1 in 814,729)
...

Makes you wonder how many times people have hit that 1 in X thing and not even noticed.
I know I have had at least 2 24 words with the same word twice that I was actually using for a while.

But, I have also generated so many for testing that I did not secure since thy were just for tests that it could have had 3 or more of the same but I never even looked.

-Dave

[apogio]

Also sounds like a kid in the back seat and the adult in the front.

Haha and this wallet has been used actually. There are transactions.

[philipma1957]

does BIP39 have rules that prevent the same word from being generated 3 times?

There is no phenomenon that stops words from repeating more than once which can be twice, thrice, or even more while generating seeds based on BIP39.

OP your question has been answered here

The probability of 11.7% is being calculated using the birthday paradox formula.

"The paradox of birthdays is a mathematical problem put forward by Von Mises, who looks for the value N in the problem: In a group of N people there is 50% chance that at least 2 people in the group share the same birthday (day + month). The answer is N = 23, which is quite counterintuitive, hence the paradox."

So for the calculations: -Repetition (Twice): 2 (number of people sharing the same birthday) -Number of words: 24 (number of people in a group) -Number of all BIP39 words: 2048 (number of days in a year)

The result is 0.119 (11.9%). But for THREE repetitions it's getting really low: 0.00043 (0,043%).

Andreas says in this video that you could use the same word 23 times + valid checksum and it would still be valid.

Yes, it will still be a valid wallet but if you are doing such a thing then the time taken to brute force your wallet will be 0.001 sec.

why would brute force do it so quickly?

if it is true brute force

and you have say 1500 word repeated 23 times.

it would start with

words. 1111111111…11
words 111111111…12


be a long fuck time before it gets to
1500,1500,1500,1500….

Or am I not understanding brute force.

A three digit combo brute force is

000
001
002… 999

so if you look at 000 first it takes time to get to 768

same idea if you run all the combo brute force.

or am I wrong

[o_e_l_e_o]

why would brute force do it so quickly?

if it is true brute force

and you have say 1500 word repeated 23 times.

There are hundreds of bots out there monitoring millions of easily hacked addresses waiting to sweep deposits the instant they show up. We've seen it countless times with brain wallets. Brain wallets generated from easy to guess phrases, such as common words, phrases, quotes, lyrics, etc., have any funds sent to the derived addresses stolen in under a second. These same bots are almost certainly also watching the addresses from any and all publicly revealed seed phrases, and any weak seed phrases, such as the same word repeated 23 times.

This is of course completely separate to a seed phrase including the same word twice or even three times, and these seed phrases are no easier to brute force than any other randomly generated seed phrases.

[LoyceV]

There is no phenomenon that stops words from repeating more than once which can be twice, thrice, or even more while generating seeds based on BIP39.

It wouldn't make any sense to do so. By limiting the number of duplicate words, you're limiting the number of possible seed phrases. If anything, that makes it less secure.
As long as your seed phrase is created randomly, it doesn't matter.

It's even possible that you generate a 24 word seed phrase and a single word is repeated 24 times.

Nope, that's not (realistically) possible. It would be more likely to generate someone else's seed phrase, than randomly ending up with 24 times the same word. It's just not going to happen.

Even if you would brute-force the creation of billions of random seed phrases, and select the one one with most duplicate words, it still doesn't matter. Nobody else is going to randomly create the same seed phrase.

This comes to mind:

(source, although I'm pretty sure this isn't the original location)