Statistics from 100 SolForge drafts

I've been playing SolForge for about a year, and draft is my favorite format. A few months ago I started a project to enter one draft per day, and record my results. I wanted to have quantifiable information about how good a drafter I was, and, although I knew I didn't have the 75+% win ratio needed to "go infinite" (i.e., to consistently win enough to fund my next draft), I wanted to see if the supplemental "income" I got from daily rewards and from forging my excess cards would be enough to make up the difference.

I started the project with about 20 tickets, if you include the purchasing power I had stored up in silver and forgable cards. I also had a full enough collection that most heroic cards I drafted would be immediately forgable, but most legendary cards would not. Although it is common for SolForge players to maintain multiple game accounts, I only have one, and I'm mostly on the free-to-play plan, although I did purchase enough gold to "seed" my drafts with a gold ticket.

It took me 140 days to record my first 100 drafts. There were a few days when I didn't log in at all, and about as many when I drafted twice.

After my 31st draft I ran out of silver completely, and could not draft again. I spent the next three weeks rebuilding my buffer. The remainder of the 40 days when I didn't draft were because I couldn't fit it into my schedule—it takes me about an hour to do a draft from start to finish—but I still earned most of the daily rewards on those days. At the 100-draft mark, I had about 35 tickets—the most I've ever had at once—and not much value in silver or forgable cards, outside of a few legendaries that I'm holding onto for trading purposes.

I've only run out of tickets the one time, so far, since beginning this project. So, while I couldn't afford to draft every single day, I've had a fairly good run. Major changes to the structure and payouts of SolForge tournaments are being made (and may already be in place by the time you read this). I don't have an opinion on them yet, but I'll be keeping my past experience in mind as I try them out.

Statistics on my results to date are illustrated below. It should be noted that this dataset is not without limitations, the most obvious one being that the whole thing is seen from the perspective of just one player. But, with that in mind, I hope people will find them illuminating, or at least entertaining.

Want to talk about this article? There's a Reddit thread for it.

Finishes

I actually started 101 drafts, but one was cancelled early when the servers went down for the Secrets of Solis update.

Win ratio

chart graphic

61.8%, to be more precise. One of the reasons I started keeping records was because I wasn't sure whether my win ratio was over 50% or not. I'm pleased to see that it is.

Finishes

chart graphic

More 3-1s than 2-2s. About as many 4-0s as 1-3s and 0-4s combined. Not too bad, I think.

Win ratio by day of week

chart graphic

I guess Garfield was right about Mondays. I kept track of this because I expected competition to be either harder or easier on weekends—wasn't really sure which—but this chart doesn't really seem to support that hypothesis.

Cumulative win ratio at intervals

chart graphic

Note that this isn't my win ratio over every set of ten games, but rather my win ratio after 10 total games, after 20 total games, 30, etc.

I was hoping this chart would show that I'm improving. I guess it... kind of does?

When going 3-1, which round was lost?

chart graphic

I generated this chart because I suspected it would look something like this, but I can't really explain why it does. Am I psychologically more likely to be confident about winning after I've been "placed in the losers' bracket"?

Factions

% of decks containing each faction

chart graphic

This is pretty much in line with how good I think each faction is in draft. I actually enjoy playing Tempys a lot, but it has not had strong cards recently.

% of decks with each faction pair

chart graphic

The Nekrium/Uterra slice of this pie was smaller before Secrets of Solis came out (bringing Seal of Deepwood, Dysian Broodqueen, and Aetherphage). But it was still the largest slice even then. I think that combination has always been the strongest, since it combines the most effective removal with the fattest creatures.

I like playing Nekrium/Tempys, but I don't trust it. Mostly I feel like it doesn't offer solid enough creatures. However, as we'll see in a moment, I should probably try it more often.

Win ratio with each faction

chart graphic

Until fairly recently, the factions I'd had the most success with were in the same order as those I picked most often, but now Nekrium has overtaken Uterra for the top spot. The difference here was always muted compared to the difference in my selections. I interpret that to mean that faction choice matters, but player skill matters more.

Win ratio with each faction pair

chart graphic

Nobody who's been following the game since the Secrets of Solis release will be surprised by Nekrium/Uterra's strong showing. I'm surprised, but only a little, by Tempys/Uterra's weak showing. I like playing that combination, but I never seem to do as well as I think I should with it.

Nekrium/Tempys is in the top spot despite my reluctance to pick it (only 6% of all decks). That could be a fluke due to the small number of data points, which will be regularized as time goes on. Or it could mean that I've been undervaluing N/T. I've heard it mentioned that removal is scarce in draft games; maybe running the two factions with the best removal isn't a terrible idea.

Win ratio with each faction, by format

chart graphic

This chart excludes Weekend Warrior drafts, since they may use non-standard card pools.

Sorry; I know this chart and the next one are a little hard to read.

The idea here is to examine whether the release (in set 3) and subsequent nerfing (in 3.1) of strong Uterra and N/U cards had as strong an effect as everyone thinks. The time period indicated by the white bars is when those cards were at their strongest. Uterra certainly did well then. It's interesting that Alloyin also was at its best then. I did have a very good run of drafting A/U decks for a while during the 3.0 period. Aetherguard and Aetherforge Oracle, both commons, have strong synergy with Seal of Deepwood. Nekrium, on the other hand, has actually been at its best after the nerf.

Win ratio with each faction pair, by format

chart graphic

This chart excludes Weekend Warrior drafts, since they may use non-standard card pools.

I haven't drafted N/T at all (except in one Weekend Warrior tournament) since the 3.1 release; that explains the missing bar.

Again we see that Alloyin/Uterra is strongest during the 3.0 period. Alloyin/Tempys looks just as strong during the same time, but I checked the spreadsheet, and it's because I only drafted A/T once during that time (and went 3-1).

Rarity

% of decks with at least one legendary

chart graphic

The total number of legendaries that I was offered over these 101 drafts was actually 25. Once I managed to draft two legendaries; they were Kas, Arcweaver (which I first-picked but never played) and Heart Tree (which was rarely played and never helpful). And I passed legendaries twice. One, ironically, was another Kas, Arcweaver (first-picked a Weirwood Patriarch instead and went 3-1), and the other was Shallow Grave (picked a Contagion Lord for a deck that already had 6 Abominations, but went 1-3).

So the rate at which one sees legendaries is roughly one in every four or five drafts. I feel like that's about right. Often enough that you have a realistic shot at opening one, but not so often that you feel like you always have to contend with them.

The legendary I've drafted most often is Palladium Simulacrum, with three appearances. It's good. Not earth-shattering, but worth taking.

Win ratio with vs. without legendaries

chart graphic

This isn't what you'd call a significant difference, but apparently I actually do slightly worse when I have a legendary.

A likely explanation is that I'm overvaluing the legendary. I think it must be good, and so I play it even when there's a common that might be more useful in the situation.

I think the lesson is, you should take a legendary if you want it, but don't feel bad about passing it. But, what this chart doesn't capture is that some legendaries are a lot better than others in draft. Everflame Phoenix is still an auto-pick, but that Kas, Arcweaver is more questionable.

Win ratio by number of heroics

chart graphic

Again, seems like there's not a whole lot of difference here, especially if you ignore the low-n sample at 0 heroics. Don't take heroics just because they're heroics.

Card Type

Win ratio by number of spells

chart graphic

Note that the sample size is very small (like, 1) at the extreme ends of this chart, but I think what we can learn from the middle is that you're slightly better off if you have fewer spells. This bears out the general rule that creatures are better than spells in draft.

I originally tried to limit my number of spells to 10, but as my record-keeping progressed, I realized that 7 was an achievable goal.

Average spells per draft by finish

chart graphic

Here's another way of looking at the data. Still makes a mild suggestion that it's better to have fewer spells.

About

If you want to record your own data, you can download a blank copy of my ODS spreadsheet here. However, it's going to need some edits to reflect the changes to SolForge's tournament structure. I plan to post an updated spreadsheet after I've had time to adjust it. If you want to use it to record constructed tournament results, you could probably do that with some minimal adjustment. You might want to add a column for "deck archetype", for example.

My filled-out spreadsheet is also available, in case you want to examine my data yourself, or if you just want to read short blurbs about my glorious victories and shameful defeats. I think my favorite story is the time I went 4-0 against one opponent with a Zimus and another with a Scorchmane Dragon.

Finally, the Ruby script that I used to turn the spreadsheet data into graphical charts is available. If you want to use it yourself, you'll need (a) Ruby, (b) ImageMagick, (c) RMagick, and (d) the gruff gem. How to install all of these things on your particular OS is, I'm afraid, beyond the scope of this article. Once you've done it, save the first sheet of your filled-out spreadsheet as a CSV file called "drafts.csv", put it and the script in the same directory, and run the script. Then send me bug reports about any horrible errors it throws. If I'm feeling especially industrious, I might fix some of them.