Welcome back #splinterlands fans! Star💫 here (@star-shroud) from the @SHIELD-OF-GLORY guild 🛡️ with the next edition of our guild-specific BRAWLS OF STEEL report ⚔️ In this edition we'll cover the results from our last two brawls. More specifically though, we'll discuss how we tried out a new brawl assignment system based on fray-specific winrate percentages. Then we'll contrast the results obtained using our current system with those obtained using the new fray-specific winrate-based system, showing that the new system produced better results. We hope this edition can be informative and helpful to other guilds who are determining an optimal fray assignment system.

Cover art in honor of @vugtis and his maxed GF epona

OUR CURRENT SYSTEM

Up until brawl cycle 98, SoG has used a brawl assignment system similar in style to the system used by many guilds today. At a high level, the system rewards top brawl performers by letting them keep their spot for the next cycle, and punishes bottom performers by taking away their spot (otherwise referred to as "benching") for the next brawl. In the case of a benching, another member in the guild takes the benched player's spot and tries to hold onto it by performing well in the next brawl.

Under our current system, we typically allowed to the TOP 5 performers to hold their spot, whereas the BOTTOM 5 were benched. Over the course of time, exceptions were made to the rules based on specific situations that felt unfair, or which just didn't make sense strategically.

• As an example of the former (i.e., felt unfair), at some point we instituted an exception to the BOTTOM 5 rule so that, if your winrate was more than 65%, you cannot be benched---even if you end up in the BOTTOM 5. This situation arose when everyone in the guild performed well in the same brawl, and even good performers were being benched simply by default.

• As an example of the latter (i.e., didn't make strategic sense), at another point we realized that in order to compete in our gold foil frays, we needed to keep the gold foil fray assignments relatively stable. This is simply because--even in a guild like ours, where some players have been playing for a long time---we don't have a deep enough bench of players with fully stacked gold foil decks to be benching from those frays. Therefore, another exception under the current system is that gold foil brawlers cannot be benched, regardless of their performance. Luckily for us, the performance of our gold foil brawlers has rarely if ever been an issue 💪

PROBLEMS WITH OUR CURRENT SYSTEM

In principle, the current system is setup to reward the best performing brawlers and punish the worst performers. The underlying principle behind the system is therefore sound. How have the results been though? Well, the results haven't actually been that reliable. In other words, sometimes we dominate, but other times we significantly underperform.

One problem we were concerned with was how the remaining open fray assignments were being filled. Obviously the TOP 5 and BOTTOM 5 were easy. They either kept their spot, or warmed the bench, respectively. What about everybody else though? We are a 30 member guild with only 16 fray assignments (for now, until we move up tiers). Under the current system, we simply allowed the remaining spots to be filled on a first-come, first-serve basis. A sort of free-for-all, depending solely on who was quick enough to sign up. As you might imagine, this is where our consistency problems seem to have been coming from.

For example, imagine the situation of hypothetical "Player X." Player X plays mostly modern, and has a chaos legion heavy deck that is light in the older editions. The frays open when it's 2:00am in Player X's country. When they wake up and go to check for open frays, the only ones left are Frays 7 and 8 (silver/regular foil/wild). Player X is all about brawls, wants to participate, and wants to help the guild. All good, admirable attributes in a guild-member. So they join Fray 7 and hope for the best. How does it go? Long story short, usually it does not go well.

It is because of situations like this that we decided to take a step back and analyze our fray-specific winrates.

FRAY-SPECIFIC WINRATES

So we took a close hard look at the fray specific winrates for each member of our guild from the last 35 brawls. Interestingly, some very clear patterns arose right away. Members fell into three general categories referred to here as SPECIALISTS, UTILITY PLAYERS, and JEKYLLS.

SPECIALISTS

The first category are fray-specific specialists. These members tend to dominate the same one or two fray spots week-in and week-out, thereby keeping them for each subsequent brawl. In some cases, this is due to deck constraints (e.g., @zagorlord, @azircon playing the gold foil frays). In other cases, it is simply due to the consistent delivery of dominating performances, week-in and week-out (e.g., @vugtis). With few exceptions, the winrates for our specialists were the highest in the guild, i.e., >75% for their primary frays.

UTILITY PLAYERS

The next category are referred to as "utility players." These members tended to play in a wide variety of frays---generally, in at least four to seven different frays. However, what really distinguishes the utility players as group was their tendency to win in whichever fray they played (e.g., @xsuilx). In this case, although xsuilx has settled into Fray 13 recently, you can see that he can probably play just about anywhere we need him to, and still be successful. These members are like "natural brawlers" and are extremely valuable to have in a guild. Among other things, they can contribute flexibility to a guild's lineup by filling in wherever they are most needed, AND consistently performing well regardless.

JEKYLLS

Finally, we have what are referred to here as our "jekylls" 😂 Yes, as in dr. jekyll and mr. hyde. What we're referring to are guild members that---like the utility players---play in a variety of frays, but---unlike the utility players---do not perform consistently well everywhere. Instead, jekylls perform very well in some frays, while significantly underperforming in others.

Simply as an example, take @petroni7. As you can see above, there are some frays that petroni simply slays (Frays 5, 8 and 9) 💪 whereas other frays appear best avoided. We believe that identifying your guild's jekylls is a critical exercise if you are looking to improve overall brawl results. As we will show below, once jekylls have been identified, they can be transformed from intermittent liabilities into consistently valuable assets. For example, by specifically assigning petroni to a fray he excelled in previously (Fray 5), we saw yet another dominating performance from him in the last brawl. This isn't rocket science 🚀 but it is where this type of analysis adds real value and pays out (staked SPS?) dividends in the long-run 💰

COMPARING THE DIFFERENT SYSTEMS

To test out the new system, we identified the top performers in winrate for each fray, and then specifically assigned them to those frays. Kind of like setting a batting order in baseball, or the starting lineup in a football match. This new approach contrasts with our current system in which most frays were taken in a first-come, first-serve free for all, without strategic thought as to which members were best for which fray.

For this experiment, two of the identified 'top-performers' in their frays (@bluestreak240 and @franzol) couldn't brawl and so were replaced, but otherwise the assignments went forth without a hitch. So ok, after all that build-up, and all that boring talk, what actually happened? Did the new system work?

Yes, in fact, it seems like it did. Not only did we turn in a blow-out first place win under the new system, but we increased the percentage of players with a winrate of >75% from 25% to 63% in just one cycle 👀 Moreover, the replacements for @bluestreak240 and @franzol performed particularly poorly. If we exclude those results, the percentage of players with a winrate of >75% was actually 71%!

Btw that winrate (>75%) is an important threshold for us. It is where we believe top tier brawlers should be performing.

Of course, the data here in this experiment is very limited---we are only comparing two brawls. It's very possible that other factors contributed to the difference in outcomes. Further testing and refinement of the system is no doubt needed. However, it is hard to argue that this wasn't a substantial turnaround. One we needed too. Getting us out of a slump and an unusually bad, four-brawl losing streak.

CONCLUSION

That's it for now folks. Please keep an eye out for our next special END OF SEASON edition (co-authored with @moros👑 leader of CARDS OF FURY ) for the only place to find out which guilds and which brawlers take the top spots this season! As always, thanks for reading and please post feedback, suggestions or whatever in the comments. We love chatting back with the community❤️💪