What to Upgrade First in CSR2, Ranked by Measured Gain

Every CSR Racing 2 guide tells you to upgrade. Almost none of them tell you what to upgrade first, because that requires knowing what each path is actually worth in seconds, which means measuring it rather than asserting it.

We have now measured all four separately — driving, Elite Customs, stars and tune — against the same record sheet. This article puts them in order. It is the practical conclusion the rest of our analysis adds up to, and the ranking is not what most players assume.

The order, by measured seconds

Two things jump out. Elite Customs returns by far the most of anything you can buy — roughly eighteen times what stars return at the median. And driving, which costs nothing, returns more than twice what stars do.

So the default order for most players is: fix your driving, commit to one Elite Customs climb, tune properly, and treat stars as the last increment rather than the first purchase.

How much driving is actually sitting on the table

The gap between a car's dyno prediction and its best real recorded time is pure driving — launch and shift timing, nothing else. Across the 1,246 cars with both figures, here is how much is available:

Three quarters of the field has at least five hundredths available from driving alone, and more than half has a tenth. A quarter of all cars have two tenths or more waiting to be found by someone who shifts properly.

None of that costs anything. If you have never deliberately practised launches, this is the largest free gain available to you and it is not close.

But the order inverts as you climb

This is the finding that changes the advice. Driving is not worth the same amount at every tier:

Driving is worth more than three times as much in Tier 1 as it is in Tier 5. The return falls steadily and consistently the higher you climb.

That makes intuitive sense once you see it. Lower-tier cars are slower, races are shorter, and a mistimed shift costs a larger share of a smaller total. By Tier 5 the cars are so fast and so heavily built that the margin for a driver to add anything has narrowed to a tenth.

The practical consequence is a genuine reversal of strategy:

• Early game, in Tiers 1 to 3, driving is the dominant lever. A median 0.18 to 0.32 seconds is available for free, which is more than most purchases will give you at that stage. Practise before you spend.
• Late game, in Tier 5, driving still matters but only returns about a tenth at the median. This is where Elite Customs and stars start to be the only remaining ground, because the free gain has mostly been taken.

Most advice has this backwards, telling new players to chase parts and telling experienced players to get good. The data says the opposite on both counts.

Why Elite Customs beats everything you can buy

Across the 70 cars recorded at more than one Elite Customs level, the median gain over a full recorded span is 0.907 seconds. That is nearly eight times the median driving gain and roughly eighteen times what purple stars return.

It also produces the largest single transformations on the sheet. A Toyota Supra goes from 9.070 to 6.856 and climbs from 873rd in its tier to 151st. A Ford Mustang Shelby GT350 goes from 706th in its tier to fourth in the entire game.

The catch is that the gain is cumulative and the per-level return is tiny — about 0.021 seconds. Buying five levels gets you roughly a tenth, which is less than you would get for free by practising launches. Elite Customs only beats everything if you commit to a long climb on one car. Spread across four cars it is the worst use of parts in the game.

Where stars and tune actually belong

Purple stars return a median 0.050 seconds, measured across 226 cars that appear on the sheet in both gold and purple form. That is real and it is consistent — the purple copy was quicker in 218 of the 226 pairs — but it is half what driving returns and a fraction of an Elite Customs climb.

Tune is the smallest lever of the four and the one players spend the most time arguing about. At the very top of the game it has effectively converged: seven of the eight fastest cars run identical nitrous and final-drive settings. There is very little left to find there, though it costs nothing to copy the recorded tune for your car and start from a known-good point.

A decision you can actually follow

• You are in Tier 1 to 3: practise launches and shifts. A median 0.18 to 0.32 seconds is free and no purchase at that stage comes close.
• You are in Tier 4: still practise, since the median car here hands back 0.178 seconds, but start choosing cars carefully — this tier has a 6.6-second spread and a bad car cannot be driven around.
• You are in Tier 5 and mid-pack: pick one car and commit to a full Elite Customs climb. Nothing else buys 0.907 seconds.
• You are in Tier 5 and competitive: stars, at a median 0.050 seconds, and the last hundredths of tune. By this point the free gains are gone.
• At any stage: look the car up before you build it. The sheet tells you both the ceiling and how much of that ceiling is driving rather than parts.

Every figure here comes from measuring the same dataset four different ways, and each measurement is written up in full if you want to check the working rather than take the summary on trust.