If you attended the LGT this year, you'll remember Tom as the ‘score guy’, enter Tom:
Time to grab your attention. Would you like to know how you win tournaments? (Tip - don’t just ask the person who won, how they won, as they might not know). We’re gonna break it down. A bit like the Hobbit trilogy, but less disappointing and more eagles.
So the first question to ask is, do you feel lucky, punk? Well do ya? Actually while luck does play an important factor, the first question you should be asking is maths related. How many people are participating in the tournament? If the number of people is greater than 2x – where x is the number of rounds to be played – it is almost certain that more than two people will end the tournament with the maximum number of wins (most tournaments are five or six rounds) and therefore have never face each other. So regardless of the points system in use for a "win", winning all your matchups is crucial to taking home the trophy.
Given that most big tournaments are five or six rounds (25 = 32 and 26 = 64) then it's certain that the winner will be decided on how much they won by, usually VPs.
If the tournament awards more points for "big wins" e.g. 12-8 rather than 3/1/0 for example, then there is scope for someone to win the tournament outright. However in a 12-8 scoring format, it’s not as easy as it sounds. At the LGT2016 the top two generals were separated by just two victory points (VPs) after five rounds and had identical gaming points.
Therefore, to win large tournaments, you need to win big on VPs, and the only way to do this is....
Does tabling win tournaments? In short, yes. Comparing the rank of the top 10 finishers in the LGT, with the number of opponents they tabled, we can see that there is a direct correlation between the number of people tabled, and how high generals finished. See table below. Note players ranked 1st, 2nd and 3rd tabled 4 of 5 games.
As we can see in the below table, Eldar and Tau are absolutely running away with it as far as tabling opponents is concerned. However, sharp eyed among you, will notice that while Tau do most tabling in the first four rounds, in round five they ended up with fewer tablings than Space Marines, and significantly fewer than Eldar.
While round 1 and possibly 2 can often be based on ‘softer’ match ups, rounds 3-5 are against generals and armies that have a similar win/loss ratio to you, so naturally you would expect to see the tabling count to decrease after this point.
From our third table below, you see a definite peak that proves round 1 can be hell, and round 2 still sees a fair number of tablings, but after this, where you’d expect to see a significant drop, there isn’t much more of a decrease. This shows that as the tournament progresses and you face typically more appropriately skilled opponents the Eldar and Marine players have the edge over the Tau player. In short, playing Tau will make it easier for you to table your players in the early stages, but when the going gets tough in round 5: you’re better off with a Craftworld Codex under your arm.
The obvious differentiator between Tau and Space Marines or Eldar is the ease of access to psykers. Marine and Eldar generals have as both are particularly potent psychic armies. Additionally, anti-Tau weaponry such as Warp Spiders and Grav-anything are almost exclusive to these armies. This would suggest Tau have an Achilles heel, which only a few races can exploit.
What can you do if you’re a Tau player? A Culexus assassin and some Riptides? It just so happens that a maxed out Riptide wing with nine Badboys, an interceptor and target locks, leaves you just enough points for that sneaky assassin at 1850pts. But, we’ll leave the list building to you guys.
So that’s it for the first in our series, I hope you found it informative. Next up on the analysis front, we’ll be looking at win rates of factions vs other factions.
Final fun stats:
Of the 50 games played by the top 10 players, 33 ended by them tabling their opponent
Of the 102 tablings that occurred in the tournament the top 10 players accounted for a third of them.
Disclaimer: in some areas the data sample size is pretty small. This is part of a growing project, and it will take time to create a full picture