To get a sense of how the simulation works, click through the demo below!

- The name highlighted in white is the player whose turn it is.
- You can see each player's hand. The card in the middle is the most recently played card that players must match with.
- The number next to the name represents how many games they have won.
- Yes, I did take the time to draw each card in pixel art.

Oh! I won! Must not be a very accurate simulation...

You may notice some funny behavior. Especially around Wild Cards. You may notice a player choosing to draw a card instead of using their Wild, or use a Draw 4 Wild when they had other good cards.

These are on purpose, and meant to test Uno strategy. I'll get more into that later.

In order to test different strategies, each player has a number of different parameters that determine their play style.

As it turns out, Uno is pretty much a game of luck. That shouldn't come as a surprise! Because it's so random, a pattern won't emerge after just playing a few games with your friends.

That's why this simulation is so handy! It can play hundreds of thousands of games in just a few seconds. With that large of a sample size, patterns can start to be seen.

There's no way to guarantee a victory, and even with all the odds stacked in your favor, the best I could ever simulate was a player winning about 65% of the time (See above! Remember the numbers next to the names represent how many times they've won).

These are the best odds you'll ever get with Uno. But I'll take what I can get! So how do you actually win at Uno? I'll walk through each parameter to find the best strategy on each.

The simplest test I could do here is try throw three players together: one evil, one random, and one nice.

Evil never wins. Except for in Uno. Hit your friends with those draw 2s whenever you get a chance, 'cause it turns out using the mean cards right off the bat is the best strategy!

Takeaway: Use mean cards like draw 2 and skip right away.

In Uno, you must place a card that matches the previous card's color or number. Matching order changes which cards a player will try to place down first. Consider this example: I need to match a card with a green 7. I can use either my green 2 or blue 7. But which is better?

I made 4 players, two match by color first, two match by number first. After 100,000 games, we can see who does better!

The edge is so small, I didn't even bother making a graph. But the numbers would suggest that matching by number is a bit better!

Takeaway: Prefer number matches over color matches.

The last strategy I tested was saving wild cards for when you really need them. It's always satisfying to get down to your last card only to reveal that it's a wild and nobody can do anything about it!

To simulate this, I originally made the computer players just refuse to use their last wild card, but turns out that really doesn't make a difference:

Their scores are about the same. Saving wilds always is a little worse even.

So instead, I modified the simulation to do this: If a player has no useable cards other than a wild card, they will draw some number of cards first. If one of the cards matches, that's great. If not, they will use the wild.

Here's a demo of the behavior. Chris has no red cards, but he does have a wild. Instead of using it, he draws a card first, but when that still doesn't match, he uses the wild.

So does this actually help? Yes, quite a bit actually!

But that's for saving your wild card for only 1 draw. What if that next card is the one you need? Does it help at all to draw a second or third card before using the wild card?

Here's a plot of win percent vs number of turns to save your wild.

It looks like only one card is all you need. Any more than that and the extra cards in your hand start to cause more harm.

Takeaway: If a wild is your only option, try drawing 1 card before using it.

Uno is pretty much random. But there are a few things you can do to improve your odds of winning.