BitShares Forum

Main => General Discussion => Topic started by: hadrian on October 06, 2014, 07:16:06 pm

Title: Mistake in AGS distribution (Resolved)
Post by: hadrian on October 06, 2014, 07:16:06 pm
Look here: http://www1.agsexplorer.com/balances/14fXr2UTa17CadcZ911N7TBFqMLpqsfNae (http://www1.agsexplorer.com/balances/14fXr2UTa17CadcZ911N7TBFqMLpqsfNae)

Compare AGS balance with DNS balance. There is an obvious discrepancy in the numbers which doesn't appear to exist for other addresses which have AGS.

The mistake is persisting for each new DAC's genesis block it seems.

(The genesis block for LTS also had a discrepancy for this address - I first asked about these discrepancies in June, but nobody seems to have an answer.)

I've started a topic like this elsewhere, (e.g. https://bitsharestalk.org/index.php?topic=9510.0 (https://bitsharestalk.org/index.php?topic=9510.0)) but hope for more visibility here, because I'm so perplexed...

It's niggling me, so I may keep mentioning it until someone tells me to shut up, or some kind of answer is obtained!

Any ideas - what's going on?
In the previous link there was mention of a reward for checking AGS allocations.
I think we need a definitive record of distribution, if there isn't one which is sufficient/correct.
Title: Re: Mistake in AGS distribution
Post by: bitmeat on October 06, 2014, 08:11:47 pm
Ags explorer only tracks donations for AGS doesn't include the PTS snapshots from after AGS donation period ended
Title: Re: Mistake in AGS distribution
Post by: hadrian on October 06, 2014, 08:17:02 pm
Ags explorer only tracks donations for AGS doesn't include the PTS snapshots from after AGS donation period ended

Did you misread, "Compare AGS balance with DNS balance" for "Compare AGS balance with PTS balance"?

[edit : The DNS balance should be exactly 1000 x AGS balance. That isn't the case for this address (but seems to hold for other addresses).]

PTS are of no relevance to the problem here.
Title: Re: Mistake in AGS distribution
Post by: bitmeat on October 06, 2014, 08:18:01 pm
DNS rewards PTS as well as donations to AGS also done with PTS. No I didn't misread.
Title: Re: Mistake in AGS distribution
Post by: bitmeat on October 06, 2014, 08:19:30 pm
Same private key could get a balance in DNS from the amount you see on AGS explorer as well as PTS balance. Which is guaranteed to be a different number if there was also a PTS holding. (Should however be a higher balance not a lower than what AGS explorer is showing)
Title: Re: Mistake in AGS distribution
Post by: hadrian on October 06, 2014, 08:32:29 pm
DNS rewards PTS as well as donations to AGS also done with PTS. No I didn't misread.

Same private key could get a balance in DNS from the amount you see on AGS explorer as well as PTS balance. Which is guaranteed to be a different number if there was also a PTS holding. (Should however be a higher balance not a lower than what AGS explorer is showing)

I see what you mean, but this donation address is a BTC address, so PTS can be taken out of the equation here.

The problem / puzzle remains...

Thanks bitmeat though! I'd love to make sense of this!
Title: Re: Mistake in AGS distribution
Post by: bitmeat on October 06, 2014, 08:36:32 pm
Sorry I was on my phone and couldn't open the links in question. Now that I know it's a Btc question. I remember that some of the math in the AGS explorer was wrong. Especially at the first day which included prior donations. Could also be difference in the cut off time for the UTC timestamp
Title: Re: Mistake in AGS distribution
Post by: pc on October 06, 2014, 09:10:54 pm
Look here: http://www1.agsexplorer.com/balances/14fXr2UTa17CadcZ911N7TBFqMLpqsfNae (http://www1.agsexplorer.com/balances/14fXr2UTa17CadcZ911N7TBFqMLpqsfNae)

Compare AGS balance with DNS balance. There is an obvious discrepancy in the numbers which doesn't appear to exist for other addresses which have AGS.

The mistake is persisting for each new DAC's genesis block it seems.

(The genesis block for LTS also had a discrepancy for this address - I first asked about these discrepancies in June, but nobody seems to have an answer.)

I've started a topic like this elsewhere, (e.g. https://bitsharestalk.org/index.php?topic=9510.0 (https://bitsharestalk.org/index.php?topic=9510.0)) but hope for more visibility here, because I'm so perplexed...

It's niggling me, so I may keep mentioning it until someone tells me to shut up, or some kind of answer is obtained!

Any ideas - what's going on?
In the previous link there was mention of a reward for checking AGS allocations.
I think we need a definitive record of distribution, if there isn't one which is sufficient/correct.

I compared the donation history from agsexplorer with my own numbers ( http://ptsags.quisquis.de/ - AFAIK this was used to create (part of) the genesis block). The difference stems from 2014-03-03.

For that day and address, agsexplorer lists a donation of 0.17038387 BTC from a total of 24.10057009 BTC yielding 35.34851445 AGS, while my version gives only 34.77140927 AGS.
I manually went through the AGS donations on that day at blockchain.info and calculated a total of 24.50057009 BTC. Now, 5000 * 0.17038387 / 24.50057009 = 34.77140927 . Therefore I believe my version (and therefore all genesis blocks so far) to be correct and agsexplorer wrong.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: hadrian on October 07, 2014, 06:53:22 pm
Thank you bitmeat and pc for your time and effort  +5% +5% +5% +5% +5%

I will now stop the questioning, and consider the matter resolved. :-*
Title: Re: Mistake in AGS distribution (Resolved)
Post by: toast on October 07, 2014, 11:23:33 pm
Could you check all the ags days against your version? Want to fiz everything at once. I believe there are two generators out there already so check against those too. I'll give pc a tip and then fix everything at once.

Sent from my SCH-I535 using Tapatalk

Title: Re: Mistake in AGS distribution (Resolved)
Post by: boombastic on October 08, 2014, 04:44:38 am
As @pc claimed, I can confirm at least 2014-03-03, the total donation is different from what it should be.  And this is the cause for confusion for @hadrian.  I am gonna investigate why the difference exists and hopefully fix it ASAP to prevent future confusion for investors.  Sorry about the wrongful data.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: toast on October 08, 2014, 04:51:34 am
Could you check all the ags days against your version? Want to fiz everything at once. I believe there are two generators out there already so check against those too. I'll give pc a tip and then fix everything at once.

Sent from my SCH-I535 using Tapatalk

Hello please this.. task force?

Sent from my SCH-I535 using Tapatalk

Title: Re: Mistake in AGS distribution (Resolved)
Post by: pc on October 08, 2014, 01:18:32 pm
Using the master download API from agsexplorer.com I have found 166 differences in PTS balances + 202 differences in BTC balances (ignoring differences below 10 satoshis due to rounding errors):

http://bitcoin.quisquis.de/dl/btc.diff
http://bitcoin.quisquis.de/dl/pts.diff

Unfortunately, the agsexplorer API does not grant access to AGS balances on a specific day AFAICS. So for me it is impossible to automate scanning for differences. For someone with access to the agsexplorer database it should be easy to find the dates where these addresses donated.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: boombastic on October 08, 2014, 01:39:17 pm
Thanks @pc, I've come to the same conclusion.  I find some blocks are randomly skipped during the parsing period.  I am re-parsing the blockchain now.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: toast on October 08, 2014, 07:44:19 pm
Using the master download API from agsexplorer.com I have found 166 differences in PTS balances + 202 differences in BTC balances (ignoring differences below 10 satoshis due to rounding errors):

http://bitcoin.quisquis.de/dl/btc.diff
http://bitcoin.quisquis.de/dl/pts.diff

Unfortunately, the agsexplorer API does not grant access to AGS balances on a specific day AFAICS. So for me it is impossible to automate scanning for differences. For someone with access to the agsexplorer database it should be easy to find the dates where these addresses donated.

Those diffs say both sides are form the same source.

I need a diff of "keyid genesis block" vs "the real thing"
Title: Re: Mistake in AGS distribution (Resolved)
Post by: pc on October 08, 2014, 09:11:55 pm
Those diffs say both sides are form the same source.

Oops. C&P problem. The right hand side is from agsexplorer.com.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: toast on October 09, 2014, 02:12:22 am
Ok great.

pc, please give a DNS or BTSX account name for a tip.

Will wait for boombastic to fix and confirm the required changes.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: boombastic on October 09, 2014, 05:14:51 am
PTS re-parsing has finished.  I compared with @pc's quisquis datasource and now they have a match.  Except some minor unmatch (less than 6 satoshi due to rounding issue), other pts balances are match.

@pc can you re-run your comparison script and confirm this?

BTC parsing is still in progress, should be finished today.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: pc on October 09, 2014, 08:31:02 am
PTS re-parsing has finished.  I compared with @pc's quisquis datasource and now they have a match.  Except some minor unmatch (less than 6 satoshi due to rounding issue), other pts balances are match.

@pc can you re-run your comparison script and confirm this?

Confirmed.
Title: Re: Mistake in AGS distribution (Resolved)
Post by: pc on October 09, 2014, 08:33:09 am
pc, please give a DNS or BTSX account name for a tip.

btsx:cyrano
keyid:cyrano

I can send a receipt if that's of any use to you.

...and I wouldn't mind your vote for my KeyID delegate: delegate1-cyrano

Thanks!
Title: Re: Mistake in AGS distribution (Resolved)
Post by: boombastic on October 09, 2014, 03:44:42 pm
BTC re-parsing is finished now.  Checked against quisquis data, all are matched.

Thanks @pc for this.
@hadrian your AGS balance is accurate now.