Nxt Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Latest Nxt Client 1.11.7 - NEW RELEASE: Ardor 2.0.3e TestNet IS LAUNCHED!

Pages: [1] 2  All

Author Topic: The story of the IGNIS Crowd Sale (first round)  (Read 2093 times)

Riker

  • Core Dev
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1693
    • View Profile
  • Karma: +429/-42

They say necessity is the mother of all invention. By the end of day last Saturday, the developers at Jelurida were facing the grim reality. A single user had bought every exchange offer they published before anyone else even had a chance to participate. Hundreds or perhaps even thousands of participants sitting in front of their computer and waiting for their chance to buy, did not even see it enabled on the user interface before it was all over.

While the Nxt blockchain itself has performed remarkably well under the immense load, the social impact of having a crowd sale which turned into two over the counter transfers was turning into a PR disaster. Not because this hasn’t been a concern as the developers went to great lengths to try to prevent this situation by splitting the 60M IGNIS round into 12 batches of 5M tokens each and by randomizing the exact block in which the sale would start to prevent power users from using phased transactions to front run others.
Still, someone has found a loophole. In the beginning, the thought was that this user (nicknamed MAAC by everyone, due to his account address being NXT-MAAC-AZFN-MK6A-CE4N7) was using a bot, which became a common perception. However soon after, MAAC himself has confessed that he was not using a bot, does not even program, and that his day job is at the movie theater. Instead, MAAC, being an NXT holder and supporter since 2013, spent weeks learning the blockchain and attempted various techniques. He figured that by listening to new transactions, he can intercept the Jelurida exchange offer just after it was published while it was still unconfirmed and immediately submit his transactions. All other users were waiting for the user interface to refresh but this would only occur when a new block was forged. By then, MAAC’s transactions had already matched the Jelurida exchange offer and the sale was over.

To make matters worse, Jelurida had submitted a phased exchange offer then approved it using another approval transaction. MAAC was using the exact same technique, thus gaining priority even on other users who tried to match the exchange while being unconfirmed using regular transactions. For the developers, the choice on what to do was not immediately apparent. On one hand, the crowd sale progressed as expected and 2M NXT were bought every 12 hours. In addition, nothing that MAAC did had violated the terms and conditions, so no legal exposure existed. Still, the feeling was that something must be done and must be done quickly. Attempts to explain the situation were faced with angry comments ranging from “The worst ICO ever” to dirty words.

After studying the technique used by MAAC, the Jelurida devs realized that using a phased exchange, while more secure, is contributing to the unfair advantage gained by MAAC. So, one quick fix (which wasn’t implemented initially) was to start using a regular exchange transaction and allow anyone who is monitoring the unconfirmed pool to compete. But still, this wasn’t very helpful to most users who do not understand the exact inner workings of the blockchain.

While users like MAAC, who study the advanced features of the platform, provide immense contribution to Nxt, the realization was that the only way to prevent MAAC, and now also several others, from getting a head start, was to beat them in their own game by allowing anyone to front run the unconfirmed exchange transaction and thus equalize the playing field and make the process more similar to a random lottery.
Nothing like this has been attempted before, not in Nxt and perhaps not in any other blockchain protocol. The idea was to let users schedule transactions in advance, which their node will submit to the network immediately after seeing the Jelurida exchange offer. This had to be implemented in a way that will still reuse the existing user interface and not require users to re-learn the buying process or having to revert to command line tools or as such.

While most of the building blocks for this feature had already existed, the actual implementation, under tight schedule and extreme pressure from potential buyers, would prove daunting. Nevertheless, slightly more than 48 hours after the problem has surfaced, and less than 24 hours after coding began, a solution was implemented and released as official version of Nxt 1.11.7
Users had to frantically upgrade their software and those who did not rely on full blockchain nodes had to set up ones. Some documentation and FAQ pages were hastily prepared and some minor bugs had to be reproduced and documented. However, for the 6th batch by Monday evening, the scheduled transaction mechanism was successfully deployed. Still, only 15 users would actually buy the IGNIS marker token (named JLRDA). The demand outpaced supply by at least 20x. Still, the process itself became random enough in order not to obviously favor any given account.
Six more batches were sold over the next few days. Some heuristics were developed in order to improve the buying chances like trying to target the buy transactions at the next block forger, but overall, the process has become fairer.

Moreover, the immense knowledge gained during this exercise has made Nxt a robust solution for crowd funding platform. The built-in features of Nxt provides tight control over the ability to spend the funds raised and the new transaction scheduler makes a stampede on the blockchain a much more egalitarian process which no longer favors geeks and whales.
Would Nxt now become a popular choice for future crowd funding campaigns? Time will tell. What is sure, is that the Nxt legacy will proceed with a vengeance with 4 more crowd funding cycles on the way and while looking forward for the launch of Ardor, the ultimate solution to the blockchain bloat problem.
NXT Core Dev
Account: NXT-HBFW-X8TE-WXPW-DZFAG
Public Key: D8311651 Key fingerprint: 0560 443B 035C EE08 0EC0  D2DD 275E 94A7 D831 1651

Brangdon

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1268
  • Quality is addictive.
    • View Profile
  • Karma: +213/-25

Well done. Kudos for doing that.
NXT-RTYD-LJXQ-EPNJ-H7AQ5. Sponsoring 1 public node at brangdon.duckdns.org.

Neomadra

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 434
    • View Profile
  • Karma: +37/-6

Well written! I'm excited for the next round. :)
« Last Edit: August 12, 2017, 03:05:02 pm by Neomadra »

cayenne

  • Full Member
  • ***
  • Offline Offline
  • Posts: 183
    • View Profile
  • Karma: +9/-1

First of all, hire MAAC.

Second, it is still a serious problem that some supporters of NXT get a favorable 2.5 to 1 ratio, while stragglers get 0.5 to 1. There is too much demand and not enough supply.

Damelon

  • Administrator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2305
    • View Profile
    • Nxt Inside
  • Karma: +792/-53

Regardless of what may have gone wrong, it is very good to see Jelurida admitting they overlooked something, and making a post about it.

That's good to see and I want to thank you for that. :)

Distribution is of paramount importance for any healthy market, and that was being messed up. Fair game to MAAC, but not good for the continuing project.
It's good to see you addressed it by levelling the playing field.

Might I suggest that you also turn posts like these into Blog Posts on the Jelurida site, so people can share them wider on social media?
The team will be judged by the way you handle messy situations like this, and it would be good for this to be shared.

D.
Member of the Nxt Foundation | Donations: NXT-D6K7-MLY6-98FM-FLL5T
Join Nxt Slack! https://nxtchat.herokuapp.com/
Founder of Blockchain Workspace | Personal Site & Blog

TheWireMaster

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 334
    • View Profile
    • NXT Folks
  • Karma: +21/-0

Well done guys! I didn't get a single Ignis so far but I have to say that the core devs tackled the situation in a great way!
NXT-5WW2-XQ63-CFGM-G7YAJ

Brangdon

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1268
  • Quality is addictive.
    • View Profile
  • Karma: +213/-25

Second, it is still a serious problem that some supporters of NXT get a favorable 2.5 to 1 ratio, while stragglers get 0.5 to 1. There is too much demand and not enough supply.
I'm not sure that's a problem, as such. The difference in prices seems to reflect uncertainty about whether the ICO would be a success, and also a desire to give better prices to old supporters rather than newcomers. Both understandable.

The deeper problem is the first-come-first-served system. That works for on-going markets, but it doesn't suit a one-off ICO, and spreading the ICO over multiple offerings hasn't helped enough and may even have made things worse. In each offering after the first three, there were around 8 people who won, and 600+ who got nothing. Those people are now pissed off. The effect of the ICO so far has been to alienate people. This is still a PR disaster.
NXT-RTYD-LJXQ-EPNJ-H7AQ5. Sponsoring 1 public node at brangdon.duckdns.org.

websioux

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 306
  • Great changes grow bottom up
    • View Profile
    • Scriba.io the Blockchain Scribe
  • Karma: +67/-1

An ICO module may well be a killer nowadays..
Secret Miner <= communicate with style | NotBot <= timestamp digital docs

forkedchain

  • Administrator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1131
  • bite me
    • View Profile
  • Karma: +68/-10

thanks for this writeup.  and yes I have been frantically closely examining the blocks with PublishExchangeOffer, the results,  doing lots of my own testing, trying to figure something out.  Ive learned QUITE a bit on how things work with the consensus mechanism lately, especially in order of transactions.
NXT tips: 2319251 or NXT-8SWM-2224-YKWW-22222

Riker

  • Core Dev
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1693
    • View Profile
  • Karma: +429/-42

I think that in retrospect splitting the initial 60M batch to 12 batches was a correct decision.
Had we sold the whole 60M at once, MAAC would have bought most of it if not all of it.
NXT Core Dev
Account: NXT-HBFW-X8TE-WXPW-DZFAG
Public Key: D8311651 Key fingerprint: 0560 443B 035C EE08 0EC0  D2DD 275E 94A7 D831 1651

abctc

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1313
    • View Profile
  • Karma: +147/-13
Welcome to the Nxt generation of crypto!   Magis quam Moneta (More than a Coin)
"Do not worry, it is an attack" (c) Jean-Luc

cayenne

  • Full Member
  • ***
  • Offline Offline
  • Posts: 183
    • View Profile
  • Karma: +9/-1

I would like Jelurida to adjust the ratios for the remaining rounds, even if that means fewer rounds or more Ignis given out. Why not increase the ICO to 1.5 or 2 billion Ignis?

Virus

  • Newbie
  • *
  • Offline Offline
  • Posts: 7
    • View Profile
  • Karma: +1/-0

Came up empty handed for round one, but I enjoyed getting more familiar with the API and how the platform functions. Met some cool people, too.

Looking forward to the next few rounds; thanks for recognizing there was an issue and for getting 1.11.7 out as soon as you did.

neofelis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 405
    • View Profile
  • Karma: +68/-10

Thanks to Team Jelurida for an elegant solution to a critical problem. Again, NXT proves to be the most useful crypto in existence.

apenzl

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2466
    • View Profile
    • Nxter.org
  • Karma: +241/-10

Very nice roundup, Riker!

Nxter.org gave these live reports:

https://www.nxter.org/ignis-ico-report-1/
https://www.nxter.org/ignis-ico-report-2/
https://www.nxter.org/nxt-news-august-2017-ii/#IGNIS_ICO
https://www.nxter.org/ignis-ico-report-3/

from MAAC's attack and the implemented fair lottery solution - for which we applaud Jelurida - till the end of Round 1.

[EDIT]: Looking very much forward to Round 2.  :)

neofelis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 405
    • View Profile
  • Karma: +68/-10

Now we just need someone to write a positive article about this on cointelegraph.com. 

nrgskill

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 29
    • View Profile
  • Karma: +0/-2

Now we just need someone to write a positive article about this on cointelegraph.com.

I don't think that giving too much exposure will help, consider that despite the incident response from developer team has been amazing (my respect and appreciation to them), the overall outcome is still negative, with a whale that drastically reduced a fair distribution during the first days of the ICO.
« Last Edit: August 13, 2017, 06:07:29 am by nrgskill »

neofelis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 405
    • View Profile
  • Karma: +68/-10

The whale got about 15M IGNIS out of a total of 500M available.  I don't think that drastically reduced the fairness of the distribution.  If you were one of the frustrated many during that first 48 hours of the crowdsale, though, I understand why it might seem not fair.  As ICOs and IPOs go, though, this is by far the fairest I've seen.

I remember when Apple went public in 1980, I wanted in but whales and special friends of brokers took everything. This is life. It's not always fair and there are no guarantees of fairness.

Damelon

  • Administrator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2305
    • View Profile
    • Nxt Inside
  • Karma: +792/-53

The whale got about 15M IGNIS out of a total of 500M available.  I don't think that drastically reduced the fairness of the distribution.  If you were one of the frustrated many during that first 48 hours of the crowdsale, though, I understand why it might seem not fair.

Perception is everything.
The amount of resources that would have to be poured into containing the negatives would bleed you dry with little to no returns.

This is one of those cases where facts will mean little and any attempt to point them out will mostly be seen as sugarcoating something that went quite wrong.

It's a given in Customer Service that 1 happy customer will net you 5x positive word-of-mouth recommendations, while 1 unhappy customer will net you 10-15 negative word-of-mouth reports.

The net loss in reputation from this mistake (and most of the handling of it communication-wise) is large and cannot be undone.

While I am happy that Jelurida is now at least communicating with this one post, the only real way to repair anything will be via product, unless they invest in a professional PR and marketing team (and learn that these are two completely different things).

ICO is not over yet, and a lot may happen. I hope the lessons of this first round have been learned.
Member of the Nxt Foundation | Donations: NXT-D6K7-MLY6-98FM-FLL5T
Join Nxt Slack! https://nxtchat.herokuapp.com/
Founder of Blockchain Workspace | Personal Site & Blog

QBTC

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 575
    • View Profile
  • Karma: +146/-26

Thank you for this write up, Riker!
NXT-5TGU-S6Y4-X2B3-3QWZQ
Pages: [1] 2  All