Faking randomness

[martinburley]

Here's an interesting website where you can see how random you are, or can pretend to be. It's based on choosing 'heads' or 'tails' rather than RPS throws, but it's still worth trying, to get an insight into how difficult or easy you find it to approximate randomness:

[Edited to update the link: this is now the correct link - the test is II A - Four Tests For Randomness.]

As we all know, one of the main criticisms that foolish people make of RPS is that, "It's just random." This website might prove otherwise - if it's that hard to be random with two choices, how much harder with three!  Although the most practical way to counter their argument is, of course, to challenge them to a game of RPS... :-)

[Wayne]

This is so fasinating.

WOW, thx for finding this site, fun and informational!

[martinburley]

You're welcome, Wayne.

Sparked off by that website, I've put together an Excel spreadsheet that helps analyse Rock Paper Scissors sequences for randomness in the same way. I think my RPS Analyser is quite useful already (with more in store for the next build), so if you want to try it, message or email me.

Once you've got it, here's how to use it:

1. Type in your throws into the green cells in the 'Input Throws' worksheet. (You can enter up to 60).

2. The 'Match Analysis' worksheet now has some useful feedback for you:

A: Total throws (percent R, P and S).

This is the most obvious measure of randomness. The closer to 33% for each, the better your throw sequence approximates a random sequence.

B: Subsequent Throws: After Rock throws, After Paper throws, After Scissors throws.

This is a very useful part of the data. It shows what you're most likely to throw after each throw. Do you tend to repeat your Rocks? Do you usually switch from Scissors to Paper, or to Rock? This can be very revealing. Again, the closer to 33% for each, the more your sequence is like a random sequence.

C: Longest run.

This shows the longest run of one throw (e.g. 5 rocks in a row). For random sequences of length 60, the expected longest run is 3.4 throws. So your sequence is most randomlike if your longest run is a run of 3 or 4. (For shorter sequences the expected longest run is shorter.)

D: Number of runs.

This shows the total number of runs. For example, R,R,P,S,S,R,P has five runs (RR, P, SS, R, P). For 60 throws, the expected result is 40 runs. The closer to 40, the more random your sequence appears to be. Again, this number is lower for shorter sequences (in fact, it's 2/3 of the total number of throws).

 
By the way, please bear in mind that I'm not recommending you play RPS randomly (though if you want to, the RPS Analyser would help you). Personally, I use the RPS Analyser for three main purposes:

1. Testing my ability to create randomish sequences in order to check that I understand what randomness is (which helps me to spot non-random patterns in opponents' play).

2. Analysing my throws after matches to see what patterns I was using. I don't expect my throws to be random, but I want to check that they're biased in the ways I knew about. For example, if I found I'd been following rock with scissors and I hadn't been aware of that during the match, that would be a useful warning sign.

3. To show people who claim, "RPS is all random," that it's harder than they think to be random, by inviting them to enter a sequence of throws. Usually it's at least somewhat non-random, and they'd have to admit that a sharp-minded opponent could exploit that pattern.


For the next version, I plan to add in a column for your opponent's throws, to make the post-match analysis side of things even more powerful. In the meantime, hope you find this useful - happy throwing!

[Wayne]

Sounds very tight, but the link does not work

[martinburley]

...

[martinburley]

The updated version is now available!

In this version, you can use it exactly as described above, but you can also analyse matches!

Enter a match into the 'Input Throws' worksheet. Again, you can enter anything up to 60 throws.
Now, on the 'Match Analysis' worksheet, you get:

1. Stats on your throw sequence (as described above)
2. Stats on your opponent's throw sequence - find the patterns in their play! Perfect for getting an edge over regular opponents!
3. Stats on the match - for example, after drawing with rock, did you tend to throw rock again? Or switch to scissors?

(If you also want to see what your opponent got in 3., you can do this by entering their throws in your input column, and your throws in their column.)

The discoveries you make by using the RPS Analyser will enhance your self-awareness, and take your game to the next level!


I can't seem to upload it at the moment. If you want to try it, give me your email address (message or email me) and I'll send it directly to you.

[Wayne]

hey, it's really cool

however, it seem to only work if you imput both oponent and your throws. you can't analyze yourself only by enter one column.

very cool anyhow, how does one make such a cool program? i thought microsoft excel is only something for like a check list of stuff you need to buy, etc, never knew you could program it like that!

[martinburley]

Thanks Wayne, glad you like it! :-)

Are you sure about not being able to analyse one column? When I enter my throws only, I still get the stats on my sequence (the first section on the Match Analysis worksheet). Just not the stats on my opponent or the match (since there's no opponent and no match).

Excel is surprisingly nifty at times. On the other hand it can also be a real $%#@. I've been having to design and use spreadsheets that are 50mb+ in size for work lately, so believe me, this was a refreshing change. :-)

[Wayne]

Would you agree that if you could get you hands on the full complete analysis from your excel of your opponent before playing him, it would almost gurantee your win right?

~~~~~~~~~
edit:

I posted Martian's program on this site:

http://www.yousendit.com/transfer.php?action=download&ufid=E6C2EC2A66D744CA

now everyone should be able to download it

[martinburley]

Wayne, looks like your link is broken too.

No worries though, this tool is still available to anyone who requests it (message or email me) - we at the Strategic Forces Department are glad to provide this tool free of charge to enhance the strategic understanding of RPS players everywhere.

[martinburley]

The WRPSS Strategic Forces Department is pleased to announce that the RPS Analyser 2008 Edition is available now. Our Outreach Initiative funding allows us to offer this powerful piece of technology free of charge to all RPS players. Simply message or email me if you want a copy.

We've already had multiple requests for this award-winning program - professional-level statistical analysis, now available to all RPS players. This is the very same spreadsheet that martinburley developed and used to his advantage in the 2007 Master of the Bull Board Tournament. The spreadsheet applies unique statistical algorithms - known collectively as Burlistics - to reveal crucial insights into not only your opponents' patterns, but also your own.

Whether you're a novice or a Master, whether you're playing against amateurs or professionals, this spreadsheet can deepen your insights into your play and enhance your ability to discover and exploit your opponents' patterns to your advantage. Don't be left behind - order your copy of RPS Analyser 2008 and harness the power of Burlistics in your play today!

[Wayne]

huh? is that suppose to cost money? sounds like it :p

[Master Roshambollah]

Our Outreach Initiative funding allows us to offer this powerful piece of technology free of charge to all RPS players.

So that's were all the Outreach money went.

It's probably a good thing; we were planning on giving all the money back to rank-and-file WRPSS members in the form of a rebate check.