Fantasy Sports vs. Poker - Long Run, Variance, and Similarity to SNGs

[Snipe]

So this one is a bit long; it's actually an email chain between a friend and I regarding the similarity between Fantasy sports and Poker or specifically SNGs. I maintain that each fantasy sports season is really more closely related to an SNG and due to the 'real time' involved, your results will likely never reflect your true skill. He maintains otherwise.

I'd love to hear what you all have to say about the subject.

Me:
Your team sucks.

Friend:

YOUR team sucks. I've just been unlucky so far. My top pick has been injured, teixera has yet to hit a bomb, and my pitching has gotten shelled. i guarantee i finish higher than you.

Me:
Lol @ "No YOU suck – I'm just unlucky" – Hilarious. Seriously – you think I've been exceptionally lucky? I love that Mr. Long run believes his Fantasy success has been anything less than variance, and that when he's NOT winning, it's just luck.

You are in fact the ball licker.

Friend:
exactly. IN the long run, i should have an overall ranking of first. Variance however takes me down a notch every once and a while. Don't worry though, the fun has just begun, its a long season, and while you think you're dope for picking up sheffield, it was in fact a donk pick up. He's washed. 290 with 27 hrs at best.

Me:
Lol – 290 / 27 – TERRIBLE – AHHAHAHAHAHAHAHA.

And come on – you know damn well that you'll never even get close to the long run with fantasy sports due to the length of the seasons. You can keep thinking you're dope and all, but it's just variance.

Friend:
I know you're trying to make a joke and all, but due to the length of seasons? That dosen't even make sense. and i said at best he would do 290 27. And for a fantasy league thats only 8 deep or whatever we are, that in fact is the very bottom of the barrel. I think its very possible he goes 275 and 20. Or better yet, goes in and out on the dL, much like the past couple seasons. Look at his numbers from the past years and get a grip.

Me:
Wow – I wasn't trying to make a joke. You've actually begun to buy into your own hype. Scary.

The length of the seasons thing makes perfect sense, but you have to think of it. Unlike poker where each (cash) session can be considered just part of an ongoing session, fantasy sports are one-and-done events – more closely related to donkaments than cash games. Being that, due to the length of the seasons, you get only one season per annum, it stands to reason that you will never hit the long run.

If you play fantasy baseball for 50 years, you will get 50 seasons. Lets say you play 2 leagues a year in basketball and baseball for 50 years, that's 200 'sessions'. Would you claim that you were the best Sit n Go player in the world after 200 SNGs? I really hope not. Particluarly when you play in a super high variance game full of maniacs who don't know what they're doing half the time.

Pull your head out Shaw – you're nowhere near the guru you believe yourself to be simply because you have a couple victories in donk leagues over the last 5 or 6 years.

Don't make the mistake of buying into your own hype.

Friend:
your numbers are heavilly skewed. You could certainly argue that there are 152 sessions within each season, giving you 152 sessions per year. Either way, to count a whole years worth of fantasy play as 1 session is riddiculous. If you wanted, you could do the same exact thing for poker. tally up all your wins and losses, come up with a bb/100, and rate that for the year against other poker players. At the end of the year would that be 1 session? Obviously it would be a bunch of sessions both good and bad. If you get rooked for a grand on one day, it dosen't mean that you'll end up poorly at the end of the year. Same thing with fantasy. One day my whole team might be 0 for 40, but the next 20 for 40. One good session, one bad. So by your logic then, we cannot know that michael jordan is in fact a good basketball player. He only had about 15 sessions and might have just been luckboxing it. Gimmie a break.

Me:
Congratulations! You will be receiving the 'Worst Argument of the Year' award in the mail shortly. The package should include a plaque, a stamp, one cookie, a coke, and a smile.

NOT counting a year as a session is ridiculous. There is a predefined start and end, and results do not roll over. They're SNGs in which you get one hand per day. End of story. In light of this fact, your argument about tallying up wins and losses makes no sense as after 152 hands, the session ends and the results are marked.

Friend:
okay, so michael jordan has only played 15 sessions then?

Me:
Sorry, not applicable as his career stands alone based on individual rolling stats. He's not considered the greatest simply because he won his championships, but rather based on his rolling career accomplishments.

Friend:
so "rolling" is the term that your'e gonna use? Fine, so my results are "rolling" then as well.

Me:
I'm not trying to take anything away from you – you're a solidly mediocre fantasy player on a hot streak – and nobody can take that from you. It's just that you understand long term results versus short term variance, and 10 - 15 sessions is not even close to long term results. It's just a fact.

Friend:
I don't think you can correctly use this poker analogy of sessions here. You have to remember that a poker session consists of something that takes place over hours, possibly minutes. This makes it all the more visible why in poker a winning player can have lots of loosing sessions in a row, but if you tally up the wins and losses after a year, there will be more wins than losses. I understand what you mean when you say a years worth of fantasy counts as "one session" because when the season is done, your rank at that point is what you are measured by. However i disagree with using this as one session, because there are oodles and oodles of stats that you must accumulate to get to that point. It is NOT like a sit n go that has a lot to do with how you run in the short term where luck may dictate how you finish. Rather, it is more like a bunch of sit n goes strung together, and at the end of the season you take count. Wouldn't you agree that if a player in the NBA averaged 20 points and 10 rebounds that was an all star this season, would produce similar numbers next year? Next GAME he might go for a bad one, maybe even 2 or 3 in a row, much like the poker player in the short run. But over the course of the season, his numbers will average out to 20 and 10. If you can say that for the basketball player, then why can't you say that for the fantasy player, who simply is an extension of the basketball player? I guess you could argue that all of your players happened to get hurt that year, but the same remains true for the player in real life.

Having said all that, you might be thinking that i'm contradicting myself when i say that i was unlucky and blaming my poor season on bad luck and injuries. You may have misunderstood me, and what i meant was that THUS FAR in this very short season, my top guys have been hampered with injuries, causing me to struggle early on. However, i would expect that this would even out by the end of the season, as more players will surely get injured on other peoples teams, or other star players will go through slumps. I am in no way writing the season off as variance, only the short comings in the young season. Now, it is possible that at the end of the year i will do worse than usual, just as the player in the NBA. For whatever reasons (nagging injuries, family issues, personal), players can have up and down years, just like the fantasy player. You wouldn't expect AI to score less than 24 points next year, and i wouldn't expect to come in any less than 3rd. If AI gets injured though, it may happen, but as statistics show, this is very unlikely and you would have to chalk that up as a downswing. Sure there are a handfull of injury prone players, but if you average them out, they are the minority. Same goes with fantasy, it can happen where i finish last in the league, but it would have to be something that i would expect only once every 12 seasons or so

Me:
I continue to completely disagree with you. The sit n' go / tournament analogy is dead on in the world of fantasy sports as that is, EXACTLY what it is. Every day is not a session, but rather a hand. In a sng, you take your hands (days), play them as best you can, accumulate chips (points), and at the end, the person with the most chips wins. Slightly different in that the tourney ends after a given amount of time versus when one person has all the chips. That said, even if you apply that structure to a sng, it's still an sng – nothing materially changes about the dynamic (though obv strategy would be time dependent – moreso than currently).

The fact is, one bad hand or big suck out can ruin a SNG tournament in exactly the same way that losing a top pick can ruin a fantasy season. You don't have the opportunity to move down and rebuild like you would in cash as you are locked in for the limited duration of the fantasy season, again, like a SNG. In this way, variance is HUGE in fantasy sports. Do the better players have an edge? Absolutely – in much the same way as they do in a SNG and this is the basis of my argument. Fantasy seasons are high variance, limited duration tournaments. Due to the fact that they happen in 'real time' – much like playing live poker versus online – you will likely never get to the true 'long term' results and thusly, your extremely limited sample size is subject to enormous amounts of variance.

Friend:
well i continue to disagree with you. By your standards then, pro players will never see the long run. END OF STORY. I don't care what you say about "rolling stats," because if they are considered the best because of rolling stats, then so can the fantasy player.

btw, we can argue the session thing till we're blue in the face. You say each day is a hand, i say every at bat is a hand, and the at bats in a game add up to a session. To me, this seems more reasonable since players can be pretty safely measured by the season they had. If a guy averages 20 and 10, do you say, "well, he very well could have been a fish that got lucky in one sit n go"? I think not, and that he will continue to average similar numbers throughout his career. But in poker, this can very easily happen. Therefore, if judging a player by a year (or two or 3), can give you a good idea of what they will do for their CAREER, then i don't think you can say each season is simply one sit n go. If this were the case, you would see many more players having luckbox seasons which for the most part dosen't happen. I understand where you're going with the sit n go thing, i just don't think it transfers over well for this example. It is true that at the end of a fantasy season points are tallied up and a winner is chosen, just like in a sit n go. However, and this is a HUGE however, these are stats and numbers that are compiled over months over the course of the season, whereas the sit n go takes an hour. If you were to find a sit n go where you started with 50 million chips and the blinds started at 1 and 2, and slowly increased so that the sit n go took 6 months to conclude, i don't think you would see anybody luckboxing it anymore.

Sure, there's an element of variance to fantasy, but i don't think its as huge as you make it out to be, and your analogy to the sit n go just dosen'tn work for me because of the huge discrepency in volume.

Me:
I don’t believe you fully understand what I mean by ‘rolling stats’ so I’ll explain. I realize you are biased and want to believe you’re right on this one, and therefore will remain steadfast in your belief regardless of any evidence provided, however, I’ll explain anyway.

By Rolling Stats, I mean a group of stats which continue to be complied infinitely. This is analogous to a poker player’s cash game record. If you chart the results, there will be pronounced ups and downs, however the general trajectory of the graph will be “up and to the right”.

This is NOT what SNGs or MTTs, or Fantasy are. Those examples are limited duration tournaments in which a player has a certain amount of time / hands / at bats, etc – to do their best with. At the end of the period, the result is taken as a single event and the win is assigned.

The fact is that simply by grouping a linear set of results over a given time period and assigning a win based on that sample, you automatically increase the amount of time it will take for overall results to converge.

Now granted, a good SNG player (or fantasy player) will win in the long run. The difference in this case is that while you get 60 – 150 hands over the hour duration of an SNG, the fantasy season allows for one hand per day and as such, you will never reach the long run.

Your point about the real ‘Time’ involved is ridiculous by the way as it has absolutely no bearing on the outcome. It’s not like we’re seeing who can grow a plant the fastest, in which case, the amount of time the plant is exposed to sunlight plays an important role. Instead, all of the things we’re measuring here are defined simply by occurrences – at bats, hands, etc. Time means nothing and you should know that. Is a single hand any more important because one guy stalls for an hour before making the call? Obviously not.

[Doug Funnie II]

You both have decent arguements. The only thing that would truly settle this argument would be to compare the standard deviations of finish position in 9 handed leagues versus 9 handed DnGs, and find the # of DNGs where their standard deviations were equal.

I'm pulling these numbers out of nowhere.
If your standard deviation of finish in 1 sng is 2, and your SD of finish position in a fantasy league is 1, then clearly the league is worth more than a sng. How many sngs is it worth? My guess would be 4. (1/2 = SDofset/number in set = (sqrt(4))/4 = 1/2)

This would give you a benchmark so you could determine what "long run" really means for fantasy sports. My gut says that it's longer than your friend wants to give credit for, but much shorter than the long run for sngs or mtts.

Also your argument that "time means nothing" I think is poorly thought out. SNGS are a game of extreme incomplete information, as well as massive uncertainty. Regardless of how long it takes to come to a decision, you can only do so much with what data you're given. Fantasy sports however are a game of slight incomplete information with massive uncertainty. In theory the longer you have to make decisions, then better your decisions should be because of all the extra available information.

Lastly you seem to want to draw the parallel that 1 hand in a sng = 1 at bat, but to me it makes more sense that 1 hand in a sng should parallel a decision that you make (i.e, to play so and so). Perhaps every card of the flop, turn, and river would be a better parallel for an at bat.

[Buffalo66]

I was hoping someone would post about the similarities/differences of poker and fantasy sports.

There is a site I play on that offers DAILY fantasy contests for real money. It's run just like a SNG - pony up $10+1 (or whatever stakes you play) and the 10 goes in the pot. I have been playing heads up every day, mostly baseball and doing very well (70% win rate). And thanks to the UGIEA carve-out, it's all perfectly legal!

I think the daily contests offer the variance similar to a SNG, especially if you play a 10 or 20 player contest. Rainouts, injuries, freak plays, overtime/extra innings can cost you in the short term.

However, IMHO fantasy baseball is more skill based than poker. There is so much instant information at your fingertips, and you don't have to make your selections until moments before the 1st game starts. I'm certain I'm playing against opponents who aren't using the same statistical/news/weather data I'm using - I suspect they are playing "hot" players or going on hunches.

As for a season long fantasy game, If you were limited on roster changes, I think it would somewhat mirror a SNG. After all, you can't change your cards at the table. If a player gets injured, you can drop him for another.

[KingMedicine]

buffalo - on what site are these fantasy SNGs-format games?
sounds pretty interesting.

[Buffalo66]

The site I play on is called gamedaydraft (dot com). Paypal accepted. They have "freerolls" if you want to get your feet wet.

Basically pick your team by choosing one player from 6-8 choices in each group. It has live scoring, which is a nice feature. It certainly makes the games fun to watch.

However, I do think someone with decent web skills could do a better job with the concept.

When you think about how much MLB/ESPN, etc. charges for a full season team, and how little they pay out, it's mind boggling. No wonder they lobbied to make it legal (post UIGEA).

[KingMedicine]

interesting site. i think youre exactly correct that if some web developer got his [censored] together, this idea could take off and be really lucrative.
is there a refer a friend bonus that youd get if i referred you when i sign up here? if so, whats yr username (PM me if youd like).

[Buffalo66]

I'm Buffalo66 on the site.