How to Review Your SMITE 2 Match History

Why KDA misleads more than it informs

KDA is the number players quote and the number that explains the least. It combines three different things — kills, deaths and assists — into a ratio that treats them as interchangeable, and they are not. An assist on a kill you contributed almost nothing to counts the same as one where you landed the setup that made it possible. A death while pushing a winning advantage counts the same as a death from walking into the enemy jungle at minute four.

The specific failure mode is that KDA rewards passivity. The easiest way to improve your KDA is to take fewer risks: stay back in fights, avoid contesting objectives, do not dive for a kill that would have ended the game. All of those raise the number and lower your actual contribution. This is why you will regularly see a player with an excellent KDA on a losing team and find, on review, that their caution was part of why the team lost.

The numbers that actually tell you something

Three measurements consistently correlate with playing well, and all three are more boring than KDA. That is generally the sign of a useful statistic.

Farm relative to game length

Gold and experience accumulated per minute is the closest thing to an objective measure of whether you spent your time productively. It is difficult to inflate, it rewards exactly the behaviour that wins games, and it exposes the most common invisible mistake in MOBA play — standing around doing nothing while a wave crashes into a tower. A player who dies twice but farms efficiently is usually contributing more than a player who dies never and farms badly.

Deaths, categorised by cause

Not the count but the reason. Sort every death in a session into one of a few categories: caught out of position with no vision, lost a fight you chose to take, died to a gank you had information about, died while contesting something worth contesting. The distribution tells you what to fix. A session of deaths from the first category is a warding and map awareness problem. A session from the second is a fight-selection problem. They require completely different corrections.

Objective participation

Whether you were present for the objectives your team took or lost. Objectives are the currency games are actually decided in, and a player who is farming a side lane during every objective fight is, in a measurable sense, not playing the same game as their team. This is the statistic that most often explains a losing streak that otherwise looks inexplicable.

Finding the decision that lost the game

Games are rarely lost at the moment they appear to be lost. The teamfight you lost at minute twenty-eight was usually decided by a positioning or objective decision several minutes earlier that made the fight unwinnable before it started. Reviewing the fight itself teaches you very little; reviewing what put you into it teaches you most of what there is to learn.

1. Identify the moment the game became unwinnable. Not the surrender vote — the point where the outcome stopped being in doubt.
2. Work backwards two to three minutes from there. What was the state of the map? Who had vision? What was the wave doing?
3. Find the decision, not the mechanic. 'I missed a skillshot' is almost never the real answer. 'We started an objective without vision while two enemies were unaccounted for' usually is.
4. Ask what information you had at the time. A bad decision made with good information is a judgement problem; a reasonable decision made with no information is a vision problem. They have different fixes.
5. Write down the one change. A single specific adjustment carried into the next session beats a list of five you will not remember.

A ten-minute post-session review

Consistency matters more than depth here. A short review done after every session will improve you faster than an exhaustive one done occasionally, because the point is to notice patterns across games rather than to fully understand any single game.

1. List the session's results and note which queue and role. Mixing these produces conclusions that describe nothing.
2. Count your deaths across the session and sort them into cause categories. Look for the category that dominates.
3. Pick the single worst game and find the decision point using the backwards method above.
4. Pick the best game and identify what was different. Success is as informative as failure and is reviewed far less often.
5. Write one sentence: the specific thing you will do differently next session. Not a goal — a behaviour.
6. At the start of the next session, read that sentence before you queue.

Frequently asked questions