Elite Customs in CSR2: What Each Level Actually Buys
69 CSR2 cars are recorded at more than one Elite Customs level. Measuring every span shows what EC really returns per level, and where it is enormous.
Elite Customs is the longest-running money question in CSR Racing 2. The parts are expensive, the levels go on forever, and the game tells you almost nothing about what you are buying. Players either commit to it completely or avoid it entirely, usually based on what someone told them once.
The record sheet has an unusual property that lets us answer this properly. 70 cars are recorded at more than one Elite Customs level, with a real time at each. Comparing those spans gives a direct measurement of what EC returns, on the same car, rather than a guess.
The short answer: about two hundredths per level
Across all 70 models the median gain per level is 0.0214 seconds. The mean is 0.0236. 69 of the 70 got faster as the level went up.
Two hundredths per level sounds trivial, and taken one level at a time it is. The reason it dominates the top of the record sheet is that nobody buys one level. The median total gain across a full recorded span is 0.907 seconds, and the biggest are far larger than that.
The levels that appear on the sheet are 0, 35, 36 and 50, so these are mostly measurements of a long climb rather than a single step.
The spans that transform a car
This is where it stops being a marginal upgrade and starts being the whole build:
| Car | Levels | Before | After | Gain | Tier rank |
|---|---|---|---|---|---|
| Toyota Supra | 0 to 35 | 9.070 | 6.856 | 2.214s | 873 to 151 |
| Bugatti EB110 SS | 0 to 50 | 9.823 | 7.786 | 2.037s | 888 to 691 |
| 2017 Nissan GT-R (R35) | 0 to 35 | 8.874 | 7.100 | 1.774s | 855 to 324 |
| Subaru 2020 BRZ tS | 0 to 35 | 9.513 | 7.895 | 1.618s | 882 to 727 |
| Porsche 911 Carrera RS 2.7 | 0 to 36 | 8.523 | 6.918 | 1.605s | 834 to 196 |
| Ford 2018 Mustang GT | 0 to 35 | 8.476 | 6.938 | 1.538s | 830 to 213 |
The Supra is the clearest case in the game. Untouched it runs 9.070 and sits 873rd in Tier 5, which is the bottom third of the field. At level 35 it runs 6.856 and sits 151st. That is the same car, moved 722 places, by one upgrade path.
The Ford Mustang Shelby GT350 goes further still. At level 0 it records 7.834 and sits 706th in its tier. At level 36 it records 6.528 and sits 4th in the entire game. The upgrade path is the only variable that changed.
| Toyota Supra | 2.214s | |
|---|---|---|
| Bugatti EB110 SS | 2.037s | |
| 2017 Nissan GT-R | 1.774s | |
| Subaru BRZ tS | 1.618s | |
| Porsche 911 Carrera RS | 1.605s | |
| Median of all 70 models | 0.907s |
Per level the gain is about 0.021s. The totals come from committing to a long climb.
Rank movement is the number to watch, not seconds
Compare the Supra and the Bugatti EB110 SS in the table above. The Bugatti gained 2.037 seconds, nearly as much as the Supra, and moved from 888th to 691st. The Supra gained 2.214 seconds and moved from 873rd to 151st.
Almost the same time gain, wildly different outcomes. The reason is that Tier 5 is densely packed in the 6.8 to 7.2 range and sparse below 7.8. A second gained in the sparse part of the field moves you past very few cars. The same second gained as you arrive in the dense part moves you past hundreds.
This has a practical consequence. If you are deciding where to spend, the question is not how many seconds a car will gain, it is where those seconds will land it. A car that ends up at 6.9 is in the fight. A car that ends up at 7.8 has spent the same parts to stay out of it.
Two results that do not fit
Being straight about these matters more than a tidy conclusion:
- BMW 3.0 CSL (E9) records exactly 6.759 at both level 0 and level 36. Identical to the thousandth. Either the car is already at its ceiling without Elite Customs, or one of the two entries is mislabelled.
It is also worth saying that none of these spans were cheap. Every one represents a long commitment of parts on a single car, and the sheet records the result rather than the cost.
A correction worth recording
An earlier version of this analysis reported 69 models and flagged two of them as suspect rows, including a Corvette that appeared to get 1.850 seconds slower as it was upgraded.
That was our error, not the data's. We had grouped Elite Customs variants by car name alone, and two models — the Porsche 959 SC and the 2020 Corvette Stingray — exist in both Tier 4 and Tier 5. Comparing a Tier 5 entry against a Tier 4 one produced a phantom regression, because the two race different distances.
Grouping by name and tier gives 70 clean spans, and the only genuine oddity left is the BMW. Every figure above has been recomputed on the corrected grouping.
What this means for your garage
A few things follow from the numbers that are not obvious from inside the game:
- Stopping part-way is mostly wasted. At roughly 0.02 seconds a level, stopping at level 5 buys you about a tenth. The gains that matter come from committing to a long climb on one car.
- Concentrate on one car rather than spreading levels across several. Four cars at level 10 is a worse garage than one car at level 40, by a wide margin.
- Pick a car that is already reasonably close to the dense part of its tier. The rank movement, not the seconds, is what you are buying.
- It dwarfs everything else we have measured. Purple stars return a median 0.050 seconds; a full EC span returns a median 0.907. They are not in the same category.
That last point is the one worth sitting with. Across this dataset the ranking of upgrades by value is consistent and not especially close: driving well beats Elite Customs for cost-effectiveness because it is free, Elite Customs beats everything you can buy, and stars are a final increment once the rest is done.
All 70 spans are in the record database. Search a car name and you will see each of its recorded Elite Customs levels listed separately, with the full tune for each.
