I wanted to compare personal curation versus community curation. When we are personally curating other user content, bias is bound to be involved. Whether you’re being altruistic at distributing wealth to random strangers or nurturing your favorite users, the weight of responsibility isn’t as great as using stake from a community account.


<center>![adamada curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/48GgTKrAHKHY7quxxBGpuFXXywrEBYLU5pGLgRyCUm7WqR4W447VBeqcHaQSqrby77.jpg)</center>


If you’re curating on your own account, you usually answer to no one. It’s you stake and you’re using it how you want it. But when it’s a community account where stake was received from delegations, there’s some responsibility in maintaining the trust given to you.

Beyond having access to liquid rewards and supporting a community program, there’s not much incentive to go on. The linear curation rewards curve has made it feasible for everyone to get almost the same returns as everyone else when it comes to curation. 

Community curation accounts are setup to support a wider audience (unless specified to only have it voting on a select few). In practice, you would get a lot of community members functioning as curators to get the community stake to vote on the post for fairer distribution of rewards. 

### Notes:
>I used the antiabuse bot created by @jazzhero for the post. It’s a community bot designed for curation and antiabuse functions powered by HiveSQL. 

>There are no absolute numbers here. If I say reaching this threshold means X, it’s just my opinion. 

>The bot displays the % of times the user has been voted on by the account. It doesn’t mean all the rewards being given by the account to the recipient goes to the recipient because beneficiaries to posts aren’t factored in.


<center>![curie curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tRxgGLVir4TgZjSFoasmQLx8EiEHw48E5V3sdcwgYMxs1MUQUrr387ghtt2sHTNFogk.jpg)</center>


I wanted to try answer when can one tell if a user’s voting pattern has entered to vote trading territory or just having personal favorites. I tested some users I suspected and reaching 8-10% is borderline while going beyond 10% yielded a high likelihood that they are just vote trading. 

Now vote trading has become even more popular since bidbots are out of fashion. The trick here is to keep it lowkey and consistent with daily posting, have a group of users join in the circle and fulfill a post quota like your income depended on it. It's not as rampant on the surface becomes votes are best controlled enough not to take the post to trending but just enough to make some substantial kickback for their time. 

The operation can be as sophisticated as just renting some HP from dlease. Nevermind the 10% fees imposed by the site for the delegator and the one leasing because the costs can be covered by a little self vote here and there to perpetuate the cycle. 

<center>![ocdb curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tRzbAPUSE7qTmjkj5MRgbnFfW6xARmPJKRqKurke8zjkFCHqmKjxa372ejvFVRTZ9MS.jpg)</center>

This is just arbitrary but I coupled these observations with taking the time to examine the accounts and which trade their votes back and forth. I didn’t include those searches here because picking fights would just distract the readers from the point I’m trying to hand out. 


<center>![curangel curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tbPQhihuTohFrwFjLCroVYT7hYVJKhvc4sfB2EnGVRipbCsz1CisFYzNhVgHBd4SM39.jpg)</center>


I think stepping beyond the 5% for the total votes you’ve given for the past 30 days is already a soft indicator of picking some favorites. This doesn’t apply if you’re set to follow some voting trails. 8/10 from the people on my voting list are people I would manually curate and set to autovote. The other 2/10 are probably from the people I chose to follow votes on. 


<center>![discovery-it.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tbPKxrNk7FgYzjrGVeq1ewugh9qZ8swRw1JSoFBG2GwCA4z9i1CvKQgiBAviMQKqfBR.jpg)</center>


For community curation accounts, the threshold is much lowered given how much votes these accounts can usually give a month and distribute substantial amounts even at lower voting weights. My bias again is just keeping the curations less than 2.5%. Below 1% is still green, 1.01-1.49% is yellow, and 1.5%-2.5% is already borderline biased unsightly for a community account to pick favorites on. 


<center>![ecency curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tRzbA48CKuFh6TQXNhbbaXkAyuUYVH1ErVycxiq7YR4byXoqS5UCidpiToTJXvaswNA.jpg)</center>


It’s not a clear cut case. The large the community account and the audience it intends to distribute votes on, the lower the threshold. This is inversely proportional to community accounts with lower stake and lower number of community members to support. 


<center>![hive.curation curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/Eo8LZPgyxRJQrxH5tDP2o1WDpgNfZ5RkfDc8szTPQxoGHs9KcjMTHh4EYZ9HevyXCK6.jpg)</center>


1% on a community account top 10 voting list could mean more than a 100$ a week. But it could also mean less than 5$ a week depending on the stake, and how often the recipient actually posts to get voted on. 


<center>![tipu curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tRzT3pFrTSBz2AGrzFmLBtoKYaicmPwbowzE77JvG8mmLFYr1HFP3gAfAY8avzoPtLF.jpg)</center>


> Important: Even if the recipient of the votes on top 10 have higher percentages, those accounts can just be distributing the rewards using the beneficiaries function. So while the bot will just display only one name, rewards are still diluted to multiple people on the posts. 


![hiveph curation.jpg](https://files.peakd.com/file/peakd-hive/adamada/23tRzfbGqiqS4gNkudb16r5GRgtPak49C6pwFP6BLkPNRfJzXKSXWjV413VYdE8ZXqMky.jpg)

The thing to mind here is that personal curation bias is more forgiving because having favorites is just being normal. Community curation accounts are a different story entirely because stake is delegated in good faith for profit and making better use of it for others (unless this wasn’t the goal for the community curation account). 

If you made it this far reading, thank you for your time.