I don’t play video games as much as I did growing up, but one of the few I’ve played in the last several years is Mario Kart Tour. I mostly played casually, until I decided I wanted a challenge: to secure a spot in the top 10 worldwide, a group dominated by only the most skilled players and/or those spending thousands on the game.

Here’s how I finished 8th spending just $5 a month.

Crash course on Mario Kart Tour

Mario Kart Tour, the mobile app and only version not native to a Nintendo device, is a little different than the other games in the series. Perhaps the most glaring way is that it’s by default single-player. Most races are against bots.

While most Mario Kart games are about winning the race, Mario Kart Tour is about winning with style. There’s arcade scoring, and you’re trying to score as many points as you can. You earn points for doing “actions” during the race, such as driving through a coin, hitting an opponent with an item, or doing a mini-turbo.

In this game, Nintendo replaced the Star item with “Frenzies”. The idea of a frenzy is that you’re still invincible, but each frenzy is randomly assigned to an item. So you can get a Red Shell Frenzy, Banana Frenzy, Mushroom Frenzy, etc. And for the duration of that frenzy you get to use that item by repeatedly tapping the screen. Frenzies are the best way to accrue a lot of actions and get good scores.

Besides getting actions in races, you also get bonus points to your scores by racing with better drivers, karts, and gliders. That makes the game very pay-to-win. Players who spend hundreds or thousands of dollars each year into the game are doing to have higher-leveled items than those who don’t.

Because the arcade scoring is heavily driven by how strong your items are, the community generally makes apples-to-apples score comparisons based on how many actions someone completes on a track. I’ve written about separately how for a several-month stretch I held the all-time record for actions in a race.

But the action-count world records are unofficial. Nintendo’s official competition is called the All-Cup Ranking (ACR). Each “tour” is a two-week period in which 45 tracks are rotated in (there are 546 tracks in the game). To rank in ACR, you try to get as high a score as you can on all 45 tracks. There are three tiers: top 10, top 100, and top 1000.

So yes, ACR is pay-to-win, but it’s a fun challenge to rank as high as you can anyway. If your items are strong enough, you can place in the top 1000 easily. If your items are really strong, you might be able to place in the top 100 easily. For most players, top 100 is already pretty hard to crack though.

Top 10 is different, though. There are so few spots available that you need an account strong enough to compete, and the skill to match. Most of the players who frequently finish in the top 10 are among the top spenders (thousands spent over four years) and most-skilled players.

I was an exception.

The Estimate

The most influential person outside of the development team to the game is a player on the other side of the world who goes by the name DKR.exe.

We mentioned earlier that competitive players measure performance on actions instead of raw scores, since scores are so dependent on your items. However, Nintendo doesn’t provide any guidance on how many actions is good.

So for each of the 546 tracks in Mario Kart Tour, DKR.exe has determined how many actions a good player should expect in a race. He literally counts jumps, mini-turbos, coins, etc. manually. These go into a spreadsheet that players can reference. Plug in your driver, kart, and glider and it spits out an expected score.

Plug in that data for all 45 tracks in a tour and you have a total score estimate. Of course, DKR’s estimates assume someone capable of playing at a very high level. It’s very hard to hit 100% of your score across 45 tracks. The best players can get 103-105% with a lot of work. The very best players have gotten up to 110%. They’re on a whole other level. I’d say a player is good if they can get to 90%.

I’m pretty good, but not 110% good. In spring 2023, bored at home with a broken foot, I came pretty close to hitting 100% of my estimate in a tour while placing in the top 100 for the second time. So I figured I had it in me to hit 100%, but the second top 100 finish didn’t feel very accomplishing. You get no in-game rewards besides some pixels, so it’s purely a matter of pride. So I wasn’t sure I wanted to hit 100% for the sake of hitting 100%.

I wondered if I should set my sights higher. I started to save my in-game resources, without a specific goal at first, hoping one would emerge.

Where’s Waldo?

There are a few ways Nintendo monetizes the game. The most popular way is called the “Gold Pass”. For $5 a month you get a decent boost in resources to compete easier. It’s a good value, and really hard to compete at the higher levels without it.

In addition, players can purchase “Premium Challenges” for ~$40. I honestly don’t know the prices for these as I’ve never bought them. There are also paywalled items, usually $4 each. Almost all paywalled items later become generally available.

While many players have achieved top 100 buying only the Gold Pass, I actually only knew of one other player who reached the top 10 with just the Gold Pass — a Japanese player named Hikarun. In fact, they made it several times. They executed a strategy of building up their account very shrewedly, were very patient to find the right tours to go for top 10, and could hit 104+% of the DKR estimated score.

There were probably a few others, but generally everyone thought Hikarun was such an outlier talent-wise that not many others attempted to follow his path.

In a slightly different direction, a player named Waldo came out of nowhere to finish in the top 10 late summer 2023. I knew of him from Discord servers, a decent player who a long time frequently finished in the top 100.

Waldo wrote a lengthy Google Doc explaining his strategy for how he planned his top 10 run. I read it, multiple times.

He hoarded resources, maintained a lot of flexibility, and built his plan assuming moderate skill level and limited time to play. Those parts seemed sharp.

There were a lot of opinions and assumptions that I thought were less so. He repeatedly downplayed his level of spend but it was clear that it was still quite a bit more than most players, probably closer to the “whales” he held in contempt. He also saved up more resources than he had to, and insisted on ignoring a few conventions that most players typically observe when trying for high scores.

As ambivalent as I was about Waldo’s specific strategy, I came out of that Google Doc believing if he could do it, I could, too. I was going to try to place in the top 10.

Pre-Planning

I started doing research on historic top 10 scores. To my surprise, in the Halloween 2022 tour (my first top 100 finish), I could’ve finished 10th if I hit 101% of my estimate. I didn’t realize I was already that close.

I assumed that Halloween 2023 would be my best shot, since some of my strongest items were Halloween-themed and helpful on Halloween tracks like Luigi’s Mansion and Boo Lake.

Nintendo then threw a wrench into the plan by announcing that no new content would be made for Mario Kart Tour, though the game would still go on. Eventually, it became clear that this meant the tours from fall 2022 to the end of summer 2023 would be identical copies.

This had pros and cons for me. One of the advantages I expected to have was for a new track and new items to come out. New items tend to be the most valuable items of their debut tour, and if I maxed 2-4 new items, that would give me an edge over the spenders. That advantage was gone.

The other side of this was that instead of guessing which tracks would appear, I’d know exactly which ones to expect. I could plan much more efficiently, and get very strong (though not perfect) loadouts.

This meant I knew which tracks would appear in Halloween 2023, and I wasn’t thrilled at what I saw. I expected to abort that plan, but waited until a few days into the tour to officially postpone. The last slot in the top 10 went to someone who scored a little over 100% of what my estimate would have been. I probably could have made it, though it would have been close.

Waiting a few more months for a better tour meant I’d have more resources and therefore better items when I did go for it, but it also meant the top players would continue to improve their accounts at faster rates. I identified a handful of other candidates and settled on the Doctor tour. I could have scored higher in other tours, but I expected the others would have more contenders.

That last part turned out to be really important. Usually there are 9-10 contenders that emerge, and other players decide not to grind as hard knowing they don’t have a chance to finish 10th. Nobody spends an extra 20 hours finishing 12th when they could finish 25th instead. In some tours, like the Ninja 2024 tour, there were 11 or 12 contenders and the total score needed ended up much higher than initially projected. The worst I’ve seen was Bowser 2024, where every single player scored 105% of a maxed-out estimate. I wasn’t going to be able to compete with that, and fortunately I didn’t have to.

One tactic I knew I’d use regardless of which tour I selected was to not broadly share my plans, and to try to stay out of the top 10 leaderboard. The top 10 cutoffs can be a self-fulfilling prophecy — some players would score more if they feel they have to but relax if they think they are safe. I wanted to avoid influencing the projected cutoffs as best I could. (I later confirmed that the player who ran the projections didn’t see me coming.)

The leaderboard updates once a day and is about 24 hours behind, so you don’t have perfect visibility. I’d try to not reveal myself until it was too late for players to fully adjust.

One more advantage — I took a week off work. I was due for a vacation anyway, and I liked the idea of being paid to play Mario Kart.

The marathon

Getting 45 good scores in 14 days is already hard. But some scores are harder than others.

Like in other Mario Kart games, each driver has a special item. Mario has the Fire Flower, Yoshi has Yoshi’s Egg, etc. But some special items provide way more actions, and therefore higher scores. These are: Coin Box, Boomerang, Lucky 7, and Giant Banana. We talked about frenzies replacing the Star item — Coin Box, Boomerang, Lucky 7, and Giant Banana frenzies are the best scoring frenzies in the game.

Coin Box and Boomerang usually get the best scores of that group. Giant Banana is usually the quickest and easiest to get a good score with, and Boomerang the longest.

When you are trying to get your highest possible score, you pretty much need to use the driver with the best skill for that track, and have them at a high level. Unfortunately, 14 of my 45 tracks required using a Boomerang driver. This could lead to getting behind schedule if Boomerang was taking too long on any track.

How to distribute the Boomerang tracks was an interesting question. I could do them all first, all last, or sprinkle them in. Most top players say to do them last, because you’ll have more knowledge about how well you’ve done on other tracks and how you’re faring against the competition.

I chose to do them first. I had 19 Coin Box tracks and am a much better Coin Box player, so I wanted to give my highest energy levels to the more difficult Boomerang.

I expected to do two Boomerang tracks per day for the first week, plus maybe one other track a day if time allowed. Then I’d do four tracks a day for Coin Box, Lucky 7, and Giant Banana drivers and have a day or two of cushion at the end to redo any tracks as needed.

As expected, the Boomerang tracks were a slog. Most of them were taking 4+ hours each and I still had some so-so scores. Yes, absolutely I second-guessed whether this was worth continuing. But slowly and slowly I got them done. I was tracking for 99% of my estimate on Boomerang courses, which was good enough to not sink me but below where I’d hoped.

The other time-consuming effort in the first week was playing with drivers without a “plus skill”. Usually in this case, most players will try to get two Coin frenzies (instead of shooting out shells at other players, for example, your driver just collects coins, doesn’t help you win the race but scores well) in a race. But since you can get up to three frenzies in a race, the top players will sometimes try for a third Coin frenzy. There’s no skill in this, your fate is completely in the hands of RNG. And unlike the plus skills, you don’t really get much out of the other item rolls (a single Coin Box, Lucky 7, or Boomerang can be quite powerful) so scoring is less forgiving if you mess up turns, miss coins on the track, or don’t spam mini-turbos.

I hoped that with only 13 courses complete in the first five days, I could pick up the pace once I moved onto the easier tracks. That started to play out. On day 6, I completed five tracks. On day 7, I did three, but two of them were aiming for Triple Coin frenzies (although I had to settle for Double Coin and a Banana on one). On day 8, I did seven, largely thanks to getting all four of my Giant Banana tracks done in an hour combined. I got the right RNG so quickly I didn’t even know the right lines to take a few times. On day 9, I did five. Just like that, I had four days left and only 12 tracks remaining.

The 12 remaining tracks were a mix of the hard tracks I’d been avoiding (including all four battles), lower-scoring tracks where it was less important for me to 100% nail, and two easy high-scoring tracks I was saving for a surprise jump on the leaderboard last day.

I burned most of that time on battles. Many players, including myself, didn’t like when battles were added to the game. Only the Balloon Battle was brought over. You’re still trying to hit other players with items to pop their balloons (lives), but in Mario Kart Tour fashion you’re doing that while trying to maximize actions. You can also get a significant point bonus by winning in under 90 seconds (called the “Well Fought” bonus).

The best players figured out how to get 200+ actions in under 90 seconds, leading to ludicrous scores sometimes 15,000 points above their estimate.

I couldn’t match that. And I tried. Danny, the best player in the world, even hopped on a call with me for a few minutes to coach me on battles — it helped but not enough to get me to 200 + Well Fought status immediately. I finished above my estimate for all four battles, but my biggest gain vs. my estimate was 7,300 points.

I returned to my last few tracks, and left just one track for the last day. It was the normal variant of Peach Gardens, a pretty tough Boomerang track. I was worn down, burnt out, and had terrible luck. I spent several hours on it, only to get sniped at the finish line of my only good run of the day. I was about 7,000 points short of my estimate for that course, so I just had to hope that wouldn’t make the difference.

But I finished the marathon. Then it was time to wait a few days for the official final standings.

Results

The Peach Gardens mess depressed my final percentage against the estimate, but I was still plenty happy with 100.5%. I had 24 of the 45 tracks above the estimate.

I was relieved that I had 26,000 points of cushion over 10th place and 65,000 points over 11th. It turned out I could’ve completely ignored battles and taken the last few days off, but there’s no way to have known that in the moment. Here are the final standings. I was the only player who had not previously finished in the top 10.

So that’s how I placed 8th in the world, surprising the whales, on $5 a month.