A Data Science Approach to Steem: Upvotes, Curation Rewards and Delegations

This year in March (on the 27., to be precise), I started delegating to @ocdb (50.530 SP), @steem-ua (25.265 SP), @steemstem (25.265 SP) and @trufflepig (25.265 SP). Delegating 126.79 SP in total isn't much, I know, but since I only have 253.11 SP, it is quite something for me. I did it to support projects, which I find cool and to make some passive income. While I was at it, I also used @steemvoter to follow the upvote trail of @utopian.trail (with 50% voting power) and followed the guild of @nextcolony (with 25% voting power). Now, almost half a year later, I want to see, whether doing so paid off. Since I was a Python developer and still use it quite often, of course this was my weapon of choice (especially thanks to the beem library from @holger80). You can find the full iPython Notebook here, in this post I will share and discuss my results.

Upvotes per Month and Curation Rewards

2017-11:   39 upvotes, 0.000 STEEM curation reward
2017-12:  328 upvotes, 0.086 STEEM curation reward
2018-01:  330 upvotes, 0.084 STEEM curation reward
2018-02:  207 upvotes, 0.160 STEEM curation reward
2018-03:  368 upvotes, 0.245 STEEM curation reward
2018-04:  267 upvotes, 0.220 STEEM curation reward
2018-05:  283 upvotes, 0.262 STEEM curation reward
2018-06:  235 upvotes, 0.186 STEEM curation reward
2018-07:  191 upvotes, 0.074 STEEM curation reward
2018-08:  200 upvotes, 0.161 STEEM curation reward
2018-09:  221 upvotes, 0.287 STEEM curation reward
2018-10:  208 upvotes, 0.359 STEEM curation reward
2018-11:  212 upvotes, 0.218 STEEM curation reward
2018-12:  354 upvotes, 0.447 STEEM curation reward
2019-01:  375 upvotes, 0.472 STEEM curation reward
2019-02:  210 upvotes, 0.218 STEEM curation reward
2019-03:  353 upvotes, 0.276 STEEM curation reward
2019-04: 1007 upvotes, 0.176 STEEM curation reward
2019-05: 1350 upvotes, 0.071 STEEM curation reward
2019-06:  919 upvotes, 0.020 STEEM curation reward
2019-07: 1339 upvotes, 0.028 STEEM curation reward
2019-08:  691 upvotes, 0.075 STEEM curation reward
2019-09:  205 upvotes, 0.277 STEEM curation reward
Total Curation Rewards: 4.394 STEEM

As you can clearly see in this plot, my number of upvotes increased dramatically after I followed the trails, but at the same time my curation rewards dramatically decreased. As you can also see in this plot, @utopian.trail stopped upvoting in May 2019 (all of their operations are currently paused, as they stated in July) and there are (apparently) some technical problems with the @nextcolony guild, since the last upvote according to @steemvoter was on 20.8.2019.

The reason I have more curation rewards now, in September, are because of Hard Fork 22, but also because of my voting power not being drained by following the trails (as it was from April to Augst 2019). The reason it peaked in December 2018 and January 2019 are most likely because back then I had a lot of time for Steem (you can also see that in the increased amount of upvotes, I gave), for the peak in October 2018 as well as for the low in July 2018, I have no explanation.

A total curation reward of 4.394 STEEM in 22 months on Steem is also rather disillusioning to be honest, but that's all you get, when you are a plankton.

Curation Rewards for the Trails, I'm following

Curation Rewards from following Utopian:     0.150
Curation Rewards from following Next Colony: 0.003

And speaking of disillusionments: as you might have guessed, following those two trails doesn't really have a lot of financial benefits. @utopian.trail got me almost as much curation rewards, as I got from doing manual curation, after I got some Steem Power. That happened in February 2018, but as you saw, I wasn't curating that much back then. @nextcolony just got me 0.003 STEEM, so practically nothing. This is also related to the fact that I only followed that guild with 25%.

All in all, following a curation trail wasn't profitable before Hard Fork 21, especially due to your constant drained voting power. Nowadays it might be, but I leave that for further investigation.

Rewards from Transfers (aka Delegations)

ocdb          : 2.999
trufflepig    : 0.111
steemstem.pay : 0.177
reverseacid   : 0.038
roundbeargames: 2.590
magicdice     : 1.921
dtube.rewards : 0.033

trufflepig    : 0.048
ocdb          : 0.202
roundbeargames: 0.443
steem-ua      : 0.348

If you go through the iPython Notebook, you will see that to calculate my rewards from @steem-ua as upvotes and to calculate them, I had to do some tricks, basically assuming that $1 = 1 SBD, so you have to take this value with a grain of salt, the actual value is lower. I also put a threshold, in order to not include "spam" memos.

The most rewards for delegating I got from @ocdb, which should be no surprise, especially since I delegated double the amount of SP than I delegated to the others. But even if you account for this and divide the rewards by 2, it is actually quite nice for the amount I delegated and for the duration of my delegation: almost 1.5 STEEM and 0.1 SBD.

As much as I like supporting @trufflepig and @steemstem: in financial terms it doesn't really pay off, especially after Hard Fork 21. The opportunity costs (the rewards I would get from curating manually) are most likely too high: e.g. in the case of the 0.177 STEEM I got from @steemstem, this makes 0.177 : 6 = 0.0295 STEEM per month (on average), which I got for delegating 25 SP. Calculating the opportunity costs here is rather difficult, so I'm just using an estimation: in December 2017, when I had around 25 SP, I made 0.086 STEEM from curating manually, given that Steem is stake-based (the more SP you have, the higher your curation rewards are) and I have more SP now, today this number would be higher.

As you also can see, I also got some transfers from things other than delegating from

• @magicdice for when I tried the Dapp
• @reverseacid for when he promoted his articles via sending memos (the sum is higher than the threshold I set, so it does not count as spam :D )
• @dtube.rewards because getting rewards for upvoting @dtube videos was a thing once

And finally a very big shout out to @roundbeargames: Thank you very much man! Due to your generosity you are actually my number supporter: adding SBD at current rate to STEEM would make 4.681 STEEM from you to 3.949 STEEM from @ocdb. Again, thank you very very much for that!

Rank of Steem on CoinMarketCap over Time

Which brings us to my final point, something that doesn't really have much to do with my account, but rather with Steem in general, and since I was already coding I just included it: how did the rank of STEEM on CoinMarketCap developed since it launched? Sure, March 2016 was a completely different time for crypto and a lot changed since, but I was always interested to see, how STEEM did, compared to all the other crypto assets out there. If you want to see the actual numbers, check out the iPython Notebook. The data is scraped from https://coinmarketcap.com/historical/

As you can see, things don't look good. Again, 2016, when STEEM was even rank 3 for a couple of weeks, was a different time. From that on it was a slow, but steady decline, the reason for that for sure being the fact that so many other crypto assets launched. Unfortunately STEEM couldn't capitalise on its use-case as well as its first (or early) mover advantage. The bear market of 2018 was quite though for STEEM and in particular for Steemit INC, but according to this chart, 2019 hit STEEM even harder and the trend is clearly rather down than up.

One thing that I have to point out is, that STEEM's rank in the historical snapshots is much lower than in the live data:

Date	Rank	Market Cap
1.9.2019	96	$57.219.574
8.9.2019	94	$57.432.315
15.9.2019	98	$54.814.748
21.9.2019 23:53 (CEST)	80	$57.643.955
22.9.2019 10:00 (CEST)	82	$56.592.162
22.9.2019 18:23 (CEST)	81	$56.421.596
22.9.2019	97	$57.205.776

The reason behind this is, that CoinMarketCap lists some scammy looking coins, who have a high market cap, only after rank 200, in order to see them, you have to go to page 3, which of course nobody does. Just look at the market cap of current #200, BitKan: $13.244.456 and compare it with the current #201, INO COIN: $362.270.510, the difference is huge. Those coins have like Karatgold Coin, MineBee or EDUCare and seem rather on the scammy side. In the historical snapshot however, those coins are included and therefore Steem's there is currently around 16 places lower than when looking at CoinMarketCaps website.

Conclusion

So, what is the conclusion I draw from all of this for myself? Following upvote trails doesn't pay really off, when you have as little SP as I do and it only drains your voting power and therefore your curation rewards, so I will stop doing that. Also, as much as I would like to support @trufflepig and @steemstem, in my current situation it doesn't make a lot of sense (financially speaking) and it would be smarter for me to use the SP for myself and therefore stop the delegation. Regarding @steem-ua: that's a bit tricky, the upvotes I got in return for delegating were quite okay (but again, take the bar of @steem-ua with a bit of salt, it is most likely lower), but since I don't post that much, I think it would be again smarter for me to stop delegating and use the SP for curating manually.

After half a year, I will stop following curation trails and only delegate to @ocdb. Probably in six months from now, I will take a look at this again and post about it.