farm points

A tournament called WINLINE MPKBK CIS LAN Season 7 has wrapped up, and most of the teams in the field picked up a decent chunk of points. Aside from three teams that did not play a full five matches — teams that fall short of five matches do not show points or rankings — and one relatively “unlucky” side that lost points, almost everyone else gained. At a glance, the point gains were much larger than the losses: 820 points seemed to appear “out of nowhere,” while only 93 points were lost. So how are these points actually calculated? And why did so many suddenly show up?

Let’s start with the event itself: 16 teams competed for a $40,000 prize pool, with a double-elimination BO3 format, a single-elimination BO3 playoff bracket, and a BO5 final. Most importantly, it was played on LAN—it was a LAN event. The participating teams ranged from just outside the top 40 to beyond the top 100, and there were also some teams that had not played a full five matches, so they were unranked.

Valve’s points system (VRS) is roughly made up of five parts: BCOL (the “bounty” collected from beating opponents), BOFF (a tournament performance factor related to bounty value), NETW (the strength of the opponent network), LANW (LAN event wins), and H2H (head-to-head strength). Based on those factors, the points you get for beating a strong team are different from those for beating a weaker team; online and offline matches award points differently; and results from different events, or event tiers, are valued differently as well.

So why did this event award so many points? Once you factor in how these mechanisms are calculated, it was a LAN event. For teams that have not been playing LAN events regularly over the long term, this tournament was especially lucrative and a clear opportunity they could not afford to miss. LAN wins come with both a cap on the number counted and a cap on the points available.

Even though its prize pool was small, that does not mean it was worth few points; the event’s weighting only affects how much the results are worth in VRS.

Rankings affect how points change. For example, when a weaker team beats a stronger one, it can pick up points from that result. Take this match, for example:

Nculear Tiger, ranked 44th before the match, were beaten by CYBERSHOKE, who were ranked 81st pre-match. The winner picked up 59 points, while the loser dropped only 19, creating a “net” increase of 40 points. Those 40 points were not simply taken off Nculear Tiger and handed to CYBERSHOKE. Instead, the model recalculates how many points this team is “worth” based on “the lineup’s recent results, opponent quality, prize money, LAN wins, and other data.” “How much you lose for one defeat” and “how much you gain for one win” are not mirror images.

So why isn’t it just a matter of adding points to one side and subtracting the same amount from the other? VRS factors in the value of the opponent itself: the higher-ranked the team you beat is, and the stronger its network is, the more points you gain. Even though network conditions aren’t a factor here because this was played offline, it was still a LAN event, so the result also benefited from the LAN bonus.

Time also affects VRS: points have a recency factor. Playing a LAN event now does not just add one more recent result; it also refreshes the recent VRS data breakdown, while older points gradually decay. LAN event wins also get refreshed rather than being preserved forever. Winning a new LAN event match adds a highly valuable LANW (LAN event win) contribution, which is why you can see a team pick up a lot of points from this match even if the opponent they beat was a lesser-known underrated hidden-gem team.

If a team’s old LANW value has decayed, its new LANW is especially valuable, and its recent results have entered the VRS, then it can very easily gain a lot of points from a single match.

This event wasn’t worth fewer points because the prize pool was only 40,000. It was because it was a 40,000-prize-pool LAN, featured a field full of tier-two team, and had multiple BO3s, which made it a very good event for tier-two teams to pick up points.

The VRS points pool is not conserved either: teams can gain a lot while losing only a little. That is not Valve handing out points from thin air. It is because the event produced new result data, and VRS only calculates from a team’s best results. Right now, for bounty collected (points), opponent network, and LAN wins, VRS takes the ten best results from the past six months for those factors.