Measuring Up the Competition: A Review of the ITC Scoring System

The COVID pandemic is far from over but as hot spots have become increasingly regionalized, countries and communities are returning to a semblance of normality, including the competitive Warhammer 40k community. One thing that has been lost in the last year and a half of reduced competitive tournament play is the functioning of the Independent Tournament Circuit (ITC) scoring algorithm and the impact of outsized events. For those who played into the final season prior to the COVID pandemic, you may recall much discussion around how the Las Vegas Open or London GT impacted the rankings and the subsequent attempt to address how these events scaled. However, with the onset of the COVID pandemic the changes were never fully on display as to their impact on the course of a season. Now that a series of larger events have taken place we can start to assess that implementation. However, before we dig into the performance of the scoring algorithm, let’s briefly discuss what the ITC score embodies. While there are several sub-circuits (Lord Marshal Conference, UK Circuit, etc) that will utilize similar scoring methods, for the purposes of this article we’ll solely focus on the ITC due its status as the only all-geo-encompassing system of record.

Credit: Robert “TheChirurgeon” Jones

The Function of the ITC Score

Generally, the ITC score is a standard of measurement which seeks to assess an individual’s performance within any given event and then, in aggregate, across the course of a tournament season. By comparing an individual’s score against another individual’s score, you can get a relative assessment of their level of performance. As such, we generally look at the Top-10 or Top-3 within a given category as the ‘leaders’ within the ITC.

An individual event score is derived generally from three things:

  • Final tournament placing
  • Consecutive wins prior to the first loss
  • Player attendance bonus

The scoring system is meant to reward players for their performance regardless of the tournament structure or format. ITC Points are assigned by a player’s final placing, their path to victory (winning 5 games and losing 1 is generally harder than losing 1 game and then winning 5 because of the win pathing) and then their having fought through a greater, more diverse field. The aggregate of these event scores will constitute a player’s score on the season. Within any given season’s rankings, the maximum ‘score cap’ tends to shift upward as the season goes on based on:

  • Number of recorded scores
  • Types of events recorded
  • Size of the player base at recorded events

As is commonly the case early in a season, the top-players under the ITC will jockey around significantly as events are recorded on a blank slate – in a sense it is a race to establish the initial score cap. Under the current system, up to six events may count towards your score (though you may play any number you wish), of which only four may be RTT-sized events (3+ rounds, 8-27 players), five may be GT-sized events (5+ rounds, 28-57 players), and then any number of Major (5+ rounds, 58-255 players), or Super-Major (7+ rounds, 205+ players) events may be recorded. The cap is established by the maximum of those events in aggregate. As a result, rather than measuring performance, early season results will instead tend to reflect on those who have simply played the most events. It won’t be until mid-season where most competitive players have capped their initial six scores with a mix of GT+ and RTT sized events that the ‘true’ rankings will start to flesh themselves out. The score cap at this point is established by the mix of GT+ sized events vs RTT sized events – as the max score one can achieve at an RTT is lower than the score derived by a GT. This is certainly intuitive, as GTs are larger, have more rounds, and will generally have a more diverse set of competitors which means that a champion will very likely have been more thoroughly tested over a greater number of games than at an RTT.

However, by late-season, what has traditionally driven the final rankings and score cap has been the impact of the Las Vegas Open (LVO). A key multiplier of the ITC scoring system is the player attendance bonus awarded at an event – the larger the event the greater its value in ITC points, as measured by total starting players. Which begs the question…

Credit: Robert “TheChirurgeon” Jones

Is Assigning a Score Value to Total Players Relevant?

The intent behind a bonus for larger-sized events was two-fold:

  • Reward the difficulty of performance in a more diverse and larger field
  • Incentivize growth of events

The former is more directly applicable to player performance, in theory, as winning or performing well in larger events, at all levels of the event, is both more difficult due to the wider range of potential threats as well as the longer win path required in performing well. In many ways it’s an abstract measure because it represents the potential difficulty, rather than a realized difficulty and is applied evenly to all players. The latter is more of a circuit incentive for the overall health of the circuit – so in this manner a player bonus is a clever, non-conflicting dual-value component.

However, in practice the player bonus doesn’t truly capture the former, particularly at the extremes. This is most easily illustrated by the outcome of the recent London GT – a nearly 600-player, 5-round event with a four-player playoff tacked on. Despite being listed as such, it would be a bit disingenuous to call the London GT event a 7-round Super-Major as >99% of players will only have played 5-rounds and only ~25% of the remaining undefeated at the end of 5 rounds were given a shot at the championship. Regardless, when looking at the final ITC rankings you see the following;

  • 1st Place (7-0), champion, 275.3 ITC Points
  • 30th Place (4-1), top 5%, 236.66 ITC Points
  • 60th Place (4-1), top 10%, 219.89 ITC Points
  • 90th Place (4-1), top 15%, 202.43 ITC Points
  • 120th Place (3-2), top 20%, 186.95 ITC Points
  • 150th Place (3-2), top 25%, 173.88 ITC Points

Let’s compare this to a large event earlier this season, the Dallas Open which was an 8-round event with 172 players (I chose not to compare to Orlando due to pod structure of that event making the comparison challenging)

  • 1st Place (8-0), champion, 255.99 ITC Points
  • 9th Place (6-2), top 5%, 216.43 ITC Points
  • 18th Place (6-2), top 10%, 199.6 ITC Points
  • 27th Place (5-3), top 15%, 180.24 ITC Points
  • 36th Place (5-3), top 20%, 166.1 ITC Points
  • 44th Place (5-3), top 25%, 156.55 ITC Points

On average the latter event is worth ~20 ITC points less at each position. Again, the intent here is that the ~600-person event is more difficult and thus rewards more points, but I think it’s fair to ask whether that’s true.

At the very top, the champion at Dallas GT played one more game on his way to an undefeated championship than the champion at London GT. Additionally, the champion of the Dallas GT played the full field in that the conclusion of the event saw a single undefeated player. This same idea plays out at each level of the placings. At the LGT, with 19 undefeated players at its 5-round conclusion, a significant number of players will have yet to face their stiffest competition – this is not the case at Dallas. There is simply no dodging when your field culminates in a single champion.

It’s hard to justify that the field of 600 in LGT is more difficult on average in 5 rounds than the field in Dallas at 8 rounds, and even harder to justify a significant point premium at each ranked position simply because of the size of the player base. Of course, this could be further extrapolated to the 6-round winner of a 64-player event – do we think that such an undefeated champion, having gone 6-0, who scores ~219 ITC points is to be rewarded an equivalent score in performance to a 60th place 4-1 finisher at LGT? Do we think their win path and results are the same? I certainly don’t.

This is of course the same issue that is yet to be resolved with the upcoming Las Vegas Open early next year. However, where LVO is different is that it has historically maintained a shadow round of play-in games for undefeated players to make the final cut – essentially an uncounted elimination game.

So Does the ITC Scoring System Work?

In a word, no. At least not fully. But this isn’t exactly an unknown either. In 2020 FLG sought to address the issue of LVO being so dominant within the scoring system. Due to its unrivaled attendance, it made attending many GT-sized events completely irrelevant in terms of the final standings and their impact on them. However, in its first year after implementing a revamped system (and making the algorithm secret in the process) the pandemic struck and, in many ways, invalidated the need or even the ability to see how well it would address the issue. Now as 2021 closes out with the competitive scene seemingly in full swing and large events being scheduled or executed, some of the old issues are beginning to crop up again.

If the intention of the scoring algorithm is intended to measure the relative performance of players within the circuit, it fails miserably. What it actually measures is a player’s access to the circuit’s largest events first, and then for a very few, their performance second. By access I mean:

  • Was a player able to attend six events?
  • Was a player able to attend large events?
  • Was a player able to attend the largest events?

I think you can confidently argue that a player, to be at the top of the game, should need to both attend some minimum number of events as well as attend marquee events to be considered. However, that last bullet is a cause for concern. The way the player bonus is applied is such that at the extremes it rapidly falls apart – this was identified originally with LVO, but I think we can say that even with the mitigation derived at the time it simply doesn’t do enough as we see with LGT. But even further, it begs the question of whether applying a scoring bonus of total participants is even a relevant measure for scoring purposes if the reality behind it doesn’t match the intent?

Credit: Robert “TheChirurgeon” Jones

If Not Player Attendance, Then What?

There is no question that players should be rewarded for performance at more diverse and difficult events. However, the way we do that today is agnostic of the tournament structure itself which as we see with LGT, creates problem. Simply put, the ITC score should reward performance at an event based not on its gross size, but on its difficulty to a natural conclusion – a single undefeated champion.

As a measure, I think this is relatively simple in idea – the path to an undefeated champion is simply divisible by a factor of two. As a result, you can simply measure the player bonus based on the eligible rounds played. This works because the game has a binary outcome (draws within the event rankings fall to secondary or tertiary tiebreakers and are ultimately decided). The undefeated pathway follows the below thresholds of players and rounds to arrive at a single, undefeated champion:

  • 3 rounds = 8 player bonus
  • 4 rounds = 16 players bonus
  • 5 rounds = 32 players bonus
  • 6 rounds = 64 players bonus
  • 7 rounds = 128 player bonus
  • 8+ rounds = 256 player bonus

Of course, no event in the world will guarantee that they arrive at exactly these thresholds. Events which register a starting player base both above and below these thresholds will need to be accounted for at one or the other thresholds. As a result, there are two problems these thresholds create for scoring:

  • What if the total players exceed the nearest cap?
  • What if the total players fall below the nearest cap?

In both cases, I believe the solution is to simply round to the nearest cap for several reasons.

First, we should note that the player bonus is capped at 8 rounds/256 players. This is a deliberate cap. The additional rounds at doubling player sizes is subject to diminishing returns in terms of inherent challenge. As the rounds and event grows longer/larger, the event becomes more challenging to win, but that challenge diminishes with each new round at some point that the additional challenge being inconsequential or immeasurable. For our purposes, we have placed this cap at 8 games. It is certainly more challenging to win 9 straight games than it is to win 8 straight games, but it’s also easier to dodge tough matchups earlier in the event as the field becomes more diluted. For this reason, I think it’s appropriate to cap the max player bonus as a measure of a player’s performance vs leaving it openly scaling.

Second, events which exceed the cap for players but will round down will inevitably see at least one undefeated player and may have a second or even a third (in the event of draws) but are unlikely to have more than two. In these cases, it is fair to say that the event could only ever achieve a single undefeated and should only be considered as such. Events which fall below the cap but round up may see no undefeated players at its conclusion. However, in practice this has proven to be rare. In reality, every single event that historically has fallen below the above listed caps has run this risk of having a one-loss winner, but generally this simply does not happen. In the rare occasion it does, I think its fair to generalize the player cap upwards for one simple reason – all other things being held equal, it boosts the value of small and mid-size events and further incentivizes growth and attendance. In essence, it’s a boost to the mid-size tournament meta.

What’s important to keep in mind, is that under the proposed system we are still incentivizing large events, both to grow ongoing events and for players to attend these large events. What has changed is that we have realigned how the system measures a player’s performance – rather than simply scaling without regard to marginal return, we tie it instead to a meaningful measure – the difficulty to crown a champion inherent to the event’s field.

Credit: Robert “TheChirurgeon” Jones

What About the Structure of the ITC Itself?

While we’re discussing the ITC, I think more broadly it’s fair to ask whether a single global score is even relevant for international comparison. After all, it’s common to compare two international meta’s only to conclude that they simply aren’t the same and are incomparable – so why do we apply a single score? Wouldn’t it make more sense to place prominence on individual regional leaders who are performing well amongst like competition? By instead regionalizing the scoring system we alleviate the core problem of access – international players are not penalized for inability to attend a marquee event like the LVO or the LGT. In many ways this is what the sub-circuits do – they localize the competition to strengthen the local scene as well as provide a better measure of player performance within that scene.

Such a shift in structure also enables a direct comparison that is simply not possible under the current system – a season-ending invitational. While the problem of access is still relevant, top regional players may find opportunities to attend a single playoff format event reserved only for the season’s top regional players to name a single champion. Such a system, based on the current structure, is simply not possible because the fundamental rankings are skewed for all the previously stated reasons – it will always heavily overweight the US or the UK due to the current outsized impact of LVO and LGT. But with a rebalanced view and method of ranking the regions, we may start to see a more fundamental comparison of the players, and the potential to provide that ‘top cut’ and playoff that see’s the metas collide.

The Last Word

I think it’s important to keep in mind that the player base is not without a voice in this system. As our community winds back up, as FLG, GW, and others begin to drive more and more large events across the regions, the problems inherent to the system will begin to become more widespread. While this won’t impact many, those who find themselves seeking to compete for top spots within regions, factions, overall, or simply striving to be their best will find themselves tied to attending events such as LVO or LGT or be penalized. Moreover, the smaller or mid-size events will simply continue to be irrelevant by season’s end. As players, and as tournament organizers, we should question whether this makes sense for the health of our community and voice our concerns and potential solutions appropriately – and whether you agree or disagree I’d encourage you to join the conversation in good faith so we can shape a better competitive system for all.

Have any questions or feedback? Drop us an email at contact@goonhammer.com.