Introduction
The purpose of this post is to see what we can learn from the new report from themarkymark called Krampus+ Leaderboard. I would like to preface this post with the fact that any single metric is almost useless when looked at in isolation in particular when it comes to applying it to Witnesses.
I am a fairly new Hive Witness myself, so I have some technical understanding of how Hive Witness Nodes work, but I don't pretend to know all the technical details or the history of what has transpired on Hive. This is just a fresh look at Hive Witnesses through the lens of someone who has been a technology consultant, principal software engineer and solutions architect for the last twenty five years.
Witness Weekly HP Rewards:
What is KE and What is KE+
I have not invented either of these two measures so I will quote @themarkymark:
Krampus Coefficient is a formula @azircon came up with to gauge how much a user sells off compared to how much they earned. The formula is simple: add up Author and Curation rewards, and divide by HP. Users who power down and sell a lot will have much higher KE number.
Krampus+ or KE+ is an modification of the formula to include Witness and DHF rewards.
The Krampus+ Leaderboard report includes both KE and KE+ as well as total author, curation, witness and DHF rewards that an account received. And you can read more about this report here:
Witness Schedule:
Special Considerations When Looking at KE+ and Witnesses
When it comes to looking at Hive Witnesses not all Witnesses are the same, first of all the top 20 witnesses are called consensus witnesses and every 63 seconds these 20 witnesses plus one backup witness will be selected to validate and sign blocks. Each witness has 3 seconds to validate the block and if they cannot do it for some reason then the block is passed on to the next scheduled witness. For each validated block a witness gets HP reward.
Hardware Costs
Based on this we have to make sure that our top 20 voted for witnesses have the best hardware and software setups and are geographically distributed and do not use the same hosting service. So these twenty have to be really special in order for Hive blockchain to function seamlessly. To be special in Tech world it means you run powerful up to date redundant hardware and your software is up to date. All of that costs money and arguably top 20 spend the most on their infrastructure. While we cannot look at the hardware or hosing service where these witnesses are being hosted (though some do provide witness reports and updates) we can definitely look at the software our top 20 are running:
I must say that to me as someone involved in servers and software it is really surprising to see that some of our top witnesses are running witness software seven to eleven versions behind the latest version of Hive Witness software! One is even running the initial release version which is v1.27.0 that you don't see very often in the tech world I am in...
Projects that benefit Hive
Another special consideration is that some witnesses use their witness earnings to fund various projects that benefit Hive. For example I know that threespeak has a video service they built and good-karma built my favorite Hive blogging front end Ecency. Mahdiyari is working on projects that benefit Hive Development same with blocktrades, themarkymark and others. So when we look at their KE+ we would have to take that into account.
Public API Servers
We also have special API servers that are critical to Hive infrastructure:
These have to be more powerful and require more storage space than regular Hive Witness Nodes. We have very few of these so those who are running these types of Nodes are contributing extra to the Hive blockchain and incur extra costs and put in extra efforts so they should be compensated extra.
KE+ Charts For Top 21
Now let's look at the top 21
What we see here is a really wide variation of KE scores! They range between KE+ of 0.443 for @blocktrades and up to good-karma of 35.393. While blocktrades score is great I had to remove him from the charts as his giant HP holdings make graphs really hard to read :)
As I mentioned above the score for good-karma is not necessarily bad as we see that it included 66K in DHF funding and we know good-karma as the face behind our top blogging front end Ecency that I am using right now to write this blog. So at least in my mind a bad KE+ score in this case is most likely meaningless.
I also mentioned threespeak and they have a video service which is also important and OCD has another high score and I know they are involved in a lot of important onboarding initiatives so those high scores do not raise any red flags for me.
The others with KE+ scores of 10+ are also involved in tools and projects that I have seen or heard about, but I think we as technical people need to do a better job touting our horn on the regular basis about what we are currently doing.
Another interesting chart I came up with is without blocktrades because he is a huge net positive that makes the chart basically un-readable. Here we can see that there are only two witnesses in top twenty who's HP is more than what they earned in witness rewards. This here is what this chart above is based on:
I have highlighted in red high KE+ scores. I don't know about many things that are going on with our top witnesses but it makes we wonder why some hold more HP than they have earned as a witness than others. Aside from funding projects from witness rewards there could also be transfers to other accounts where the HP is held, I didn't dive that deep...
Theoretically top twenty witnesses would run similar configurations and would have similar expenses to cover so it is interesting to see such a wide variations of KE+ and HP to Witness Rewards...
KE+ Charts For Witness Rank 22-60
Now let's take a look what is going on with witnesses that are ranked a bit lower that all server as backup witnesses. Though the higher ranked you are the more frequently you validate blocks. Think of it as cars racing around the track and the total HP voting for a witness determines not only the rank, but how quickly you are racing around the track getting to your opportunity to server as a backup witness and validate your block.
Here we also see a huge variation in KE+ scores which range from a great KE+ score for @ctrpch of only 0.168 to 94.818 for someguy123
Below are the KE+ scores for witnesses ranked 22 to 60:
Here we also start to see some things that indicate lack of proper maintenance such as broken price feeds that are highlighted in red below that range from 34 days to 1815 days!
We also have some really old witness software running such as v1.27.0 and one witness in 60th place is currently off and has been off for 167 days!
And some of these witnesses did get some decent rewards in the past:
Here we see that majority of witnesses hold much less HP than they have earned as a Witness... With a major outlier anyx that earned over one million hive but is holding under 400k... and someguy123 who is our KE+ highest scorer in this witness category...
We do have quite a few witnesses like solominer, neoxian, oflyhigh, @dalz, brianoflondon, engrave, techcoderx, vsc.network, hivecuba, sagarkothari88, stayoutoftheherz and reazuliqbal who hold more HP than what they have earned as a witness!
The big positive outliers that really stand out are @solominer and neoxian who hold almost 800K more HP than what they have earned as witnesses! Stayoutoftherz is also a bit net positive to Hive as he holds over 400K more HP than what he earned as a witness.
The biggest negative outliers are in the top 21-30 witness category with anyx and someguy and jesta the most in the negative.
Overall it seems that this group in general has a better KE+ than our top 20!
KE+ Charts For Witness Rank 61-101
My witness is actually in this category at number 88 and get an opportunity to validate a Hive Block every 5 hours.
Here we have witnesses that don't earn very much hive at all and we have not only inactive witnesses but witnesses who probably no longer on HIVE.
We have three witnesses with giant KE+ scores here. I will remove those really bad scores because they render the charts un-readable... So the giant KE+ score holders are: cervantes 108K, steempress 49k and thecryptodrive 82K.
Without those three witnesses we have the following KE+ score chart:
The big outlier here is Klye... and roomservice, but roomservice is no longer active as a witness so that leaves Klye dwarfing everyone else:
To see a better picture I am going to remove Klye and roomservice this gives us a better picture of what is going on in this group:
To get another view on this we can take a look at Witness Rewards earned vs HP Held:
The giant blue bars tell a story here of witnesses that must have been top twenty witnesses at some point but no longer are there and slid down below top sixty witnesses.
We have a lot of inactive witnesses in this group so when we remove those we get a better view of what is going on with those witnesses that are still active:
Interesting trend that I see here is many of the witnesses in this group that earned most rewards do not hold much HP anymore... and those with smaller overall witness earnings hold more HP... Could be that these are newer witnesses...
Conclusion
Looking at the Krampus+ (KE+) data across all witness ranks, one thing becomes clear - no single metric can define a witness’s true contribution to Hive. While KE+ gives us a glimpse into how much each witness retains versus sells, it doesn’t capture the bigger picture — the costs of running powerful infrastructure, the time spent maintaining nodes, or the value of funding ecosystem projects.
Many top witnesses with high KE+ scores, like good-karma, threespeak, and OCD, are actively building services that directly benefit Hive. Their higher KE+ values likely reflect project investment, not disengagement. On the other hand, the presence of outdated software among some top witnesses is a real concern - it reminds us that up-to-date infrastructure and technical diligence are as critical as any metric we can calculate.
Interestingly, backup witnesses (ranks 21–60) often show better KE+ scores and healthier HP-to-earnings ratios. This suggests that a number of mid-tier witnesses are quietly reinvesting more of what they earn back into Hive. The data also reveals positive outliers like solominer and neoxian, whose HP holdings far exceed their witness earnings - clear signs of long-term commitment.
At the lower ranks, we see a mix of inactive or legacy witnesses, many of whom were once top performers but are no longer active. This highlights why voter awareness and regular re-evaluation of witness votes are so important - to ensure Hive’s consensus layer stays secure, modern, and community-driven.
In the end, KE+ is a useful indicator - but context matters. Witness impact should be evaluated through a broader lens: technical reliability, HP commitment, project contributions, and engagement.
I also realized that there is a lot to learn about Hive and have more questions than answers...
I am also a Hive Witness and would really appreciate your vote for me as a witness: https://vote.hive.uno/@seattlea
I completely agree.
Thank you for this useful post full of interesting food for thought.
I need to review the witnesses I voted for.
@tipu curate 2
!discovery 30
You are welcome. There is a lot more to dig in the https://hiveanalytics.usehive.com/ke-leaderboard report from themarkymark...
KE is a good indicator but if you are unsure about someone or if you should support him or not it's worth a close check, in case of wtinesses withdraws could be to pay bills and hardware for the node... Some people also have an account for just doing the witness and then move the rewards to main account
Yes, I mentioned these things in the part where I said that I didn't dig deeper 😀
I have so much extra server space at work that I would really love to spin up a VM or two and run a witness, but I don't think my work would appreciate the bandwidth that it would take. There is a good chance I would be called out about the traffic and I wouldn't really be able to justify it.
I don't think you should risk it at work... The earnings below top 20 are nowhere near justifying even the minimal risk of doing something like that. The bandwidth is actually quite minimal, what it takes is a lot of fast storage and ideally good CPU. You can run a regular witness at home and you will not notice any load if you have a fast and reliable connection.
Oh really, that might be an option. I have an M2 Mac Mini that basically sits un-used in my basement. I was using it as a PLEX server for a while, but I haven't kept it on lately. I have three or four servers at work that are all decked out running Hyper-V. Recently moved away from VMWare after the Broadcom crap show. I have our infrastructure running on them and still have a ton of RAM, cores, and storage space available on all of the hardware. Like you said though, probably not worth it. Besides the questions it might raise, it is likely a bit unethical.
Don't know much about Macs, can they run Linux? I am running my witness on Ubuntu Linux 22.04 with 64GB Ram and 2 TB SSD space. But you can easily get away with 32 Gig of RAM and 1 TB SSD maybe even less if you run the latest software with the new rolling block log feature. Here is what you need: https://ecency.com/witness/@themarkymark/setup-a-hive-witness-using-the-new-rolling-block-log-feature-jte
Mac OS X that they rolled out a couple of decades ago is built mostly on Unix, so I don't think it would take too much to make it work. I'm not sure that Mac Mini has the specs I need though. I will have to look into it sometime.
tl:dr - jumped to the conclusion but still what is the key takeaway now for normal users regarding witness and their KE/KE+ numbers? What do the witness elite crowd think about the stats? Is there anything we shall do / change in terms of our votes?
For me personally if I see a KE+ over 10 then I better know what this witness does for Hive beyond just running a validator node. Didn't see any of the witness elite comment on KE+ yet.
ok got it - had a proper read across entire post - makes sense. As others already said and you indicated some have different accounts where staking might happen or use funds for other projects such as OCD or Roeland, some are not active anymore, some I am surprise to see such as Blurt witness @thecryptodrive :-)
Outstanding post mate. I am not ashamed to say that I learned a little bit about the underlying mechanics of how the witness schedule operates from your post and the linked resources.
No shame in learning, I must admit that your Author Report Card inspired me to do some coding against Hive JS API, will be releasing the first draft of the tool shortly on my Azure site.
I like learning and will build it out overtime in a slightly different direction from your tool and hopefully the two tools will be complimentary to each other.
It's an interesting breakdown. It has certainly been a gravy train for some but also we have to consider it has been over 9 years as well.
I do like that backups get a shot at producing blocks. That is one thing I don't like about the Avalon chain or the EOSIO implementations. No blocks are given to backups.
That is true. Here is what various Hive witnesses earned in September:
Certainly a profitable endeavor for that top 20. It's always interesting to see where the price would rest based on how much it costs to maintain the network compared to the POW chains.
As we all know the price going to $1 changes the dynamic a lot from those running nodes to the activity in the whole ecosystem. That makes it more viable for top 20 node operators to be fulltime without additional proposal support. In my opinion there needs to be massive action and deployment of the HIVE Funds or the price isn't going to catch much sun even in this bull market with BTC and BSC hitting all time highs.
The writing is on the wall though. The funds won't be utilized and everyone will clinch up and the price probably won't do great especially when the bear market comes.
A similar thing happened to Telos. Amazing tech but the price had gotten destroyed.
https://www.coingecko.com/en/coins/telos
First things first; who are the ones that the running outdated software and what can we do about it?
Is there a reason?
Can we tag them and ask?
I would say that those on this list that display in Orange are running outdated Hive Witness software:
https://hive.arcange.eu/witnesses/
So everyone running 1.27.4 and below, but 1.27.0 really looks bad to me as that is the first release of the current iteration of Witness software that works with the Hive blockchain since Hard Fork 27. It works and maybe that is OK, but coming from traditional tech it is unthinkable to run software that was not updated in three years and is 11 versions behind a current version...
These updates are usually full node only and have no impact on consensus. Thus no point in upgrading as it incurs more risk and potential downtime.
So witness software doesn't get any security patches? Like zero since last hard fork? (v1.27.0)
First off, witness nodes should not be exposed to the Internet.
There are only a few things like (RC) that are not consensus, which would require a hard fork.
Any attack vector on a witness node would be an attack vector on the chain and would require a hard/soft fork to address. I like my forks hard and consensual.
There have been a few changes I believe to address RC that some of us have upgraded to.
Most Hived updates are for full nodes and not for consensus (witness) nodes. So upgrading isn't always important.
Got it. Thanks for the info.
This post was shared and voted inside the discord by the curators team of discovery-it
Join our Community and follow our Curation Trail
Discovery-it is also a Witness, vote for us here
Delegate to us for passive income. Check our 80% fee-back Program
Thanks for sharing! - @azircon
