Use access key #2 to skip to page content.

TMFCrocoStimpy (91.82)

CAPS Scoring Tweaks - Please break them

Recs

28

August 18, 2011 – Comments (46)

Having posted numerous times over the last year that we are analyzing how picks and scoring are working, I wanted to share with the CAPS community some of the general results and a few of the ideas as to how to nudge the scoring metric back to reflect the original intent.  The idea behind this and future blogs is to help answer some of the questions about why we haven't pursued some paths already, show some concise demonstrations of certain issues (such as accuracy) that players have expressed concern with, and propose some possible tweaks to the system that might eleviate those issues.  With the tweaks, what I really hope to find are people who are willing to think about how they would game the new system and seek to break it in certain ways.  As always, we don't want to introduce new issues that are just as troublesome by solving old ones.

In the beginning.......

Admittedly I wasn't working for TMF when CAPS was first envisioned, so I'm making up any specifics here, but the general idea was to create a stock picking community that was fun and engaging to participate in, and to aggregate the information into a stock rating system.  One of the key ideas here was to keep it engaging, which among other things required the construction of incentives for players to keep making picks.  This isn't the level of incentives that get people battling it out for the top fool quote spot, but rather the incentive for people who might or might not want to participate to jump in and get their feet wet with stock picks.  To do this require two basic components: 1) make the potential CAPS player ranking grow with the number of successful picks, and 2) make the scoring metrics sufficiently transparent that it was easy to figure out how your score is being calculated.

So what was the metric that was created?  Anyone truly worried about this topic has probably got this memorized.  The raw score from each pick is accumulated in an open ended fashion, so it can grow or shrink to unlimited size.  This is the aspect of the scoring which draws people in to keep making picks that they believe will beat the market, even if just by a little, because they can grow their lifetime score.  Because the cumulative raw score would encourage picking any ticker that a player believed had the slightest chance of beating the market, a closed metric was introduced to calculate an average value component as well.  This was the accuracy score, whose intention was to provide a guide to the likelihood of an individual pick that was made exceeding chance (hold your objections - we'll get to the problems and tweaks to accuracy below).  This bounded metric, if it were working as intended, would get players to balance the shear volume of picks made for points with their ability to consistently pick stocks that beat the market.  Thus someone might opt for picking fewer stocks that they have great confidence in to move up on the accuracy metric, whereas a different player might pick a large number of stocks that they have moderate conviction in to garner points.

Reality sets in......

In the early stages of testing the scoring metrics, it rapidly became obvious that some level of minimum positive performance had to be set to determine positive accuracy, since it was simple to capture miniscule price fluctuations that always ended up positive.  The threshold of +5% was set on an empirical basis, and has been the subject of much discussion and analysis since then.  The problem that this threshold exhibits is that with the level of volatility in the markets over the past several years, exceeding this threshold in very short periods of time has become a fairly straight forward endeavour, to which quite a few players have systematically set out to exploit.  I don't say exploit in an overly negative sense - it is an artifact of the scoring metrics which simply "is" - but it has potentially served to distract these players from examining and picking stocks with other types of strategies which were historically beneficial to the community.  The loss of strategic diversity is what concerns me the most, since that is the greatest edge in information that community intelligence embodies.

Changing Player Behaviour.....

Saying that I want to change player pick behaviour is a somewhat dangerous thing, because that smacks of hubris in me thinking I know better than the community about what it should do.  Taking the comments that have been coming for a few years, and the rather narrow strategy that is employed to boost accuracy so that it has distorted its intended information, There are a couple tweaks to the scoring system that I would like people to think about, and to figure out how these might be exploited themselves.

Quality of score:  a simplified version of examining the raw score return vs. the volatility of the pick.  High quality picks would have a high (score/volatility) measure, indicating that the return is more due to real price movement than to chance price fluctuations.  For picks that are relying more on high volatility than actual directional movements to cross the +5% for accuracy capture, the Quality metric would end up with a low value, so there would be some balance that has to be struct between the two.  This is still a fairly straight forward metric to understand, but should have an impact on player pick selection.

Accuracy Decay: Banking accuracy and holding on to it forever is the scoring bugaboo.  With the typical pick volatility of ~30% annually in the past year (volatility is this high due to the relative movements of the market and the individual stocks), you can play a simulation game of making hypothetical picks that don't have any expected return (meaning the same rate of return for the stock and the benchmark).  If you let that run and make a habit of closing every pick when it exceeds +5%, you find that you get an accuracy of 68%.  If you have an actual information advantage (meaning the stock really does have a higher rate of return than the benchmark), the accuracy rate jumps up significantly from there.  The reason that accuracy isn't working as it was intended is that a player can capture all the picks at different times in their random walks.  As a deux ex machina players are no longer playing in the same statistics arena with each other, so the comparison of their accuracy is not so simple.  In the extreme, a player who is solely focused on closing out +5% gains vs. a player who holds picks for very long periods of time cannot be compared to one another.

The level of complexity to "fix" the accuracy issue is pretty high, and requires a lot of assumptions about the statistics of stock price movements that may or may not be valid.  Rather than structure some overly deep compensation mechanism, I'm more interested in a simple heuristic that allows the closed pick accuracy to decay over time in a sensible fashion.  The open pick accuracy I have no issue with - these should always contribute fully to the accuracy score, because they are still being actively bounced around by the market.  But when a pick closes, you only want that pick to contribute to your future calculation of accuracy in proportion to how much information it contained when it was closed.  The simplest thing would be to set a fixed timeline, say 1 year, and have the contribution to accuracy decay linearly with time.  But that doesn't make a lot of sense, because both the timeline for decay is arbitrary, and the value of the accuracy information is assumed to be equal.  Kicking it up a notch, the three relevant factors seem to be: 1) magnitude of pick score, 2) volatility of pick, 3) duration for which the pick lasted.  Let's examine this in more detail.

The magnitude and the volatility go hand in hand, in that we are interested in how likely we got the pick score that we have given the volatility.  This works for both winning and losing scores.  The time that the pick lasted is relevant because it can be used to reflect the typical pick style of a player, in particular having players who largely make short term picks only hold on to that accuracy for a short period of time, so they would have to constantly keep making picks or start holding longer to maintain their accuracy rating.  I do not want to punish short term picks, but if that is primarily a player's style then I would expect them to need to maintain the activity to maintain their player standing.  The pick decay might look something like this:

decay factor = exp(-time*(decay function))

The decay function would go faster for shorter pick hold times, and go slower for pick scores that are more standard deviations of the pick volatility.  So, even if you have a very short duration pick, it might not decay if the stock score is so large (compared to the volatility) that it offsets the short time decay.

There are a lot of other details in both of these tweaks, and multiple other ideas being batted around, but I am hoping that some of my fellow CAPS gearheads will think about these ideas and see if they a) seem like they might make the scoring system better, and b) not introduce additional problems.

Fool On!

Xander

46 Comments – Post Your Own

#1) On August 18, 2011 at 3:09 PM, TSIF (99.96) wrote:

Most of the gearhead stuff is over my head, but I appreciate these efforts. I'll have to try to digest this. 

As for the accuracy on the +5, it might be nice to NOT count the accuracy against us on a ZERO to NEGATIVE 4.99.  It's really hard to close a pick now adays with all the volitility. I've tried using limits and the market drops disporportional to the equity I'm trying to close and I get a negative 0.07 and get stuck for life with the negative accuracy!  ;)

Just a pet peeve!

Again, I appreciate the concepts, gearheads unite!

TSIF

Report this comment
#2) On August 18, 2011 at 3:17 PM, TMFCrocoStimpy (91.82) wrote:

TSIF, thanks for the response.  I'm definitely looking at how to make the accuracy more symmetric without leaving it overly susceptible to a selection bias of people quickly closing out anything that is losing.  That is definitely part of the ongoing analysis to try and understand if decaying short term negative picks should operate similar to short term positive picks, but rest assured that negative accuracy will decay if positive accuracy does.

-Xander

Report this comment
#3) On August 18, 2011 at 3:34 PM, ETFsRule (99.93) wrote:

The biggest issue I have with the scoring system is that it's biased against newer players. If someone joins CAPS today, they are already trailing the top Fools by thousands of points, so it's going to be extremely difficult for them to ever catch up.

It might be nice to add another score, which is your outperformance vs the S&P 500 over the past year.

Report this comment
#4) On August 18, 2011 at 3:37 PM, DaveGruska (95.84) wrote:

I definitely appreciate the effort, but a few examples would be helpful, as I'm having a hard time understanding as well.

As a bit of an aside, I think seeing more meta-data (i.e. players who primarily pick large caps, dividends, player sector leaders, etc.) would help to keep people more engaged. I tried putting together a simple C# application using the Fool's CAPS API to try and do some of this myself, but it was choking on penny stocks (or at least stocks with a "." in them), and I never got a response - actually, the error reporting mechanism is broken itself (which I also reported seperately, with no response, either). I'm not sure if the CAPS API idea has been abandoned, but I would LOVE to try out some things.

I've got to say, though, that CAPS is easily the best stock market "game" I've ever used. Tracking a stock pick vs. an index is brilliant, the pitches my CAPS members are usually helpful, and there are a number of great CAPS bloggers.

Report this comment
#5) On August 18, 2011 at 3:38 PM, DaveGruska (95.84) wrote:

"It might be nice to add another score, which is your outperformance vs the S&P 500 over the past year."

+1 

Report this comment
#6) On August 18, 2011 at 3:42 PM, TheDumbMoney (53.33) wrote:

This is way over my head, too, at least without re-reading it a couple of more times. 

But what about, as an antidote to "bank-some-accuracy" "close'm at +5" issues, just incorporate points-per-pick into the final rank determination?  (To some degree.)  I like to know how accurate people are, and to some extent it is important, especially if one takes a trading vs. an investing viewpoint.  But if you created an incentive for people to run up significant points-per-pick, that would seem to change the dynamic.  And on the negative side, it would create an incentive to close certain near-permanent-loser picks that otherwise never get closed because of current incentives and risk/reward (at least as it seems to me).

I'm sure this raises all kinds of other concerns I just haven't thought of yet.

Report this comment
#7) On August 18, 2011 at 3:43 PM, TheDumbMoney (53.33) wrote:

By points per pick, I mean average pick score. 

I'm not an avid follower of the CAPS rules in general though.

Report this comment
#8) On August 18, 2011 at 3:49 PM, MaxTheTerrible (87.33) wrote:

Couple things I would like to see are (i) the ability to make multiple calls on the same stock (say up to 3), so that if the stock drops, for example, you could "double down"; (ii) some type of a weighting system to pick the size of the position, so that if you are sure about the stock you could make a larger allocation (I suppose, if multiple picks for the same stock would be allowed you could use that as well); (iii) totaly agree with #3 above; (iv) underperformance by x number of points really should be counted as -2x in your total score, since it takes a 20% gain to wipe out a 10% loss.

Cheers!

Report this comment
#9) On August 18, 2011 at 5:06 PM, SultanOfSwing (96.36) wrote:

This post makes 100% sense ... after you've had 2 margaritas!

You guys are in an unenviable position of damned if you do, damned if you don't.  My only request is that any change you propose, be explained with total clarity and "elegance" to a non-gearhead like me.  The CAPS Help pages clearly explain the rules, so good job there.  Any new rules that fall short of the Help pages will lead to ambiguity, which will lead to uncertainty, which will lead to doubt. And we've seen recently what investors do with doubt!

SoS

Report this comment
#10) On August 18, 2011 at 5:23 PM, SultanOfSwing (96.36) wrote:

some balance that has to be struct between the two (sic)

Aha! you must be a C programmer.  :)

So yeah, I did have to read this twice -- and without alcohol.

If I get what you're saying, you want to discourage the practice of making many, many quick picks with small positive scores (lower score average and thousands of closed, small positive picks).  Players using this strategy, like say TMFEldrehad or me, may have a tougher time of it.

Report this comment
#11) On August 18, 2011 at 5:44 PM, memoandstitch (< 20) wrote:

I stopped playing CAPS a year ago.  As you try to write more rules, CAPS stops simulating real investments and anything that doesn't simulate real investments isn't worth looking at.

To be specific, the volatility picks you mentioned would not be an issue if CAPS allows bet sizing and takes into account volume i.e. score of a pick is weighted by the size of the bet and the size of a bet is capped at the volume of the stock (translation: stocks with low volume cannot yield much score). 

Report this comment
#12) On August 18, 2011 at 6:36 PM, TMFCrocoStimpy (91.82) wrote:

I have definitely gone to the dark side of geek speak - this is one of the reasons that I am rarely let out of the programming closet under the stairs unless it is for a refill of doritoes and jolt cola.  Thankfully there are far more elegant writers here at TMF who would take a crack at explaining how any of these functions would work, but this goes to illustrate one of the original points: anything that tackles the real underlying technical issues can become very difficult to explain, which could result in becoming a turn-off to participating in CAPS because you don't have a good feel for how your score is calculated.  The potential loss of players due to this effect could be a higher liability than the distortion that occurs from the accuracy exploits.

I have no beef with short term calls, as long as they contain information that is better than chance.  In fact that is the idea whether the stock is held for a long period of time or a short period of time.  What is happening now is that players are able to take advantage of the high volatility in the market to capture +5% moves that give them an accuracy rating higher than what we believed would be the measure.  Because they can hold onto that accuracy forever, it becomes virtually impossible to reduce someones accuracy who is making very large numbers of calls.

Again, we want people to make a lot of calls, but to reflect how "accurate" you are requires looking at the open picks, plus some understanding of whether the past picks were truly accurate or not when you closed them.  And that is where it gets messy, because the only way to make some estimate about how "accurate" a pick was when you closed it is to get into some statistical modeling of the stock market volatility.  Now I love that kind of stuff, but then again I'm a statistical modeler - we're the guys who suck the life out of the room at a party of mathematicians - so before contemplating doing something that will make the scoring system far more opaque we want to know if anyone can see how to sneak around the ideas.

Streeter123 suggested an example, so lets see if this helps.  Without getting into the gory details of the calculation, let's just look at some results to compare.

Example 1: picking a 3x ETF

The volatility of the 3x ETF is huge, especially when combined with the SPY.  These pairs boarder on having 120% annualized volatility.  If you managed to get +5% in 2 weeks and then closed the pick, the rate that we would decay the accuracy contribution from this pick would be fast.  So if an open pick has an influence of 1 whole pick, then as the influence of our example decays, it would only have the influence of 1/2 a pick at 2 weeks after close, and 1/4 of a pick a month after close. 

Example 2: Pick IBM

Really low volatility relative to the market, only about 12% annualized.  You wouldn't expect that the difference between the SPY and IBM would be very large by chance, so if it outperformed the SPY by 5% in only 2 weeks it would be considered a very significant event, indicating that it had truly outperformed the market in that time period.  If it were closed at that point, because the size of the jump is so much larger than the expected volatility, the influence of this banked accuracy would decay really slowly, still contributing the equivalent of 9/10 of a whole pick at the end of a year.

The reason why the second example would be considered more "accurate" than the first is simply because random chance could have played a much larger role in the first example than the last.  The longer a player holds onto a pick the longer the pick will contribute to accuracy after it has closed.  However, if the pick was producing a positive score due to chance rather than actual positive performance, than the longer the pick is held the greater the chance that it will go negative again.

I will have to check what happened to the email links on the CAPS API page - it has definitely not been abandoned.  Meanwhile, anyone who wants to contact me can reach me at xtwombly@fool.com.

dumber, max, ETFs - I'll tackle each of those ideas (which we have dutifully tucked away both now and in the past) in some upcoming blogs.  Nothing is ruled out.

Fool On!

Xander

Report this comment
#13) On August 18, 2011 at 6:44 PM, totallyoblivious (30.75) wrote:

I've stopped trying to play CAPS as it appears to be largely a function of red thumbing leveraged ETFs and closing them frequently to inflate accuracy and score.

Want to make CAPS transparant, fun, and somewhat representative of accuracy?  Do this:

* Eliminate all scores from leveraged ETFs and similar things derivatives with a time decay component (VXX and the like)

* Determine a player's CAPS rating based on a combination of total score and the "average pick score" rating.  Consider adding a token 10% of the rating being from accuracy, but its value is far overratd in this system.  Similarly, total score is overrated without factoring in the average pick score as you can make an infinite number of picks without dilution in the current system, whereas with a real porfolio you have finite funds and have to either reduce cash levels or trim down other positions to make room for new ones.

* Apply a minimum number of picks to qualify for a CAPS rating, or apply a % penalty for individuals with less than X picks to prevent someone from picking a few momentum stocks that do great.  Also consider a decay function to penalize players who do not actively continue to make new picks.

Report this comment
#14) On August 18, 2011 at 6:48 PM, dwot (40.93) wrote:

Ugh...

On the accuracy, can you not just make it so one stock can only be counted once in the accuracy score?  

When you green thumbing, you can get hundreds of percents if the stock does more then a double.  Going to zero limits to 100% loss and that is the end.

If you close the red thumb at 5% each time, it takes going down less to reach that 5% each time and you can end up with considerably more then 100% chasing a stock down, which I did.  The accuracy programming got thrown way out of wack because each time you do that it counts the same stock.

Now, I don't see maximizing the percent your score gives by closing and reopening red thumbs as some kind of unfair advantage.  You have percent gains being calculated for green thumb items based on stock prices several years ago and dividend coming off base prices.  To me, these all fall into the same kind of issue.  But I do think the recounting of the same stock is problematic, which when I was actively playing, I exploited to the fullest as I am competitive based on how things are scored, not whether they make sense.

Report this comment
#15) On August 18, 2011 at 7:17 PM, truthisntstupid (92.68) wrote:

When I can look through a top player's early picks and figure out he's just not really much of a stockpicker, and it wasn't until he started redthumbing leveraged ETFs that his score and accuracy really started making rapid gains, it really irks me that this kind of player has more weight in rating a stock than many of us who do not use any leveraged ETFs.

Eliminate leveraged ETFs.  

By the way.  We can see who you are.  We can look through your early picks and pitches, and if you are ignorant, it shows.

Report this comment
#16) On August 18, 2011 at 7:18 PM, TMFBlacknGold (98.91) wrote:

You guys will have a tough time balancing any adjustments to the scoring metrics with retaining new players. If it is made too complicated it could deter newer investors from joining CAPS entirely.

Change wouldn't be a bad thing though. I absolutely agree with factoring in average pick score to a player's rating calculation. That would be simple to understand and, better yet, would encourage players not to close picks immediately after they cross +5%.

Accuracy decay is also a good idea, although it's tough to choose a timeframe for such a metric. One idea would be to have different ratings for different timeframes. Why not have a lifetime rating (all picks across all time) so as not to penalize players like Babo and UltraLong AND a past year rating. It would be cool to have a CAPS winner every year (like a league champ in sports).

One REALLY easy (temporary) way to fix CAPS would be to delete accounts that haven't been touched in forever. There are quite a few accounts that haven't made picks in years, but are clogging up the rating ladder.

Perhaps TMF holds a live chat to get player's opinions on how to make ammendments to the scoring system.

I'd be interested in your thoughts on my ideas =)

BlacknGold

Report this comment
#17) On August 18, 2011 at 7:40 PM, TMFCrocoStimpy (91.82) wrote:

totallyoblivious:

I will admit a personal distaste for the leveraged ETFs - it is just an admitted bias, largely because their excessive volatility has led them to be used widely in the accuracy banking.  We have had several discussions about the value of allowing these in the CAPS universe, but to date have erred on the side of not removing something that can be traded on the open market.  What I am hoping some of the above modifications to do would be do reduce the effectiveness of pumping accuracy with these instruments, and that players would turn their strategies to other types of picks.  With some form of accuracy decay a player who relied on many short term picks would only maintain their high accuracy rating if they kept playing - as soon as they slacked off, their past accuracy would begin to decay and within a few months their accuracy would be pretty close to that shown by their active picks.

Combining cumulative score as one metric and average score as another is actually something that we looked at, and still have some variants in the works.  The one thing that this still encourages, though, is the use of higher volatility stocks that can be closed on an upswing that is driven more by chance than the actual movement of the tickers.  I am hoping to include some incentives for players to look for lower volatility stocks as well that will both outperform and underperform the market, and I'm not sure if substituting the average score in place of the accuracy would do that.

With regards to a minimum number of picks, we already do that.  It takes 7+ picks before you are assigned a rating, and then your accuracy is adjusted slightly (see How is My Rating Calculated) based on the number of picks you have until you get up to 100.

dwot:

Ugh indeed.

I don't think the issue is about whether a single stock is used multiple times.  If I can make an accurate call on a single stock 10 different times that is as valuable as making 10 separate accurate calls on different stocks, provided that the 10 separate stocks have the same statistical characteristics as the single stock.  What I'm trying to eliminate here is more the element of random chance masquerading as accuracy, which conveniently allows me to treat every pick as a separate entity, regardless of the actual stock.

I've never had a problem with the points generated from chasing a short down or the excess accuracy.  What I would like to do is have put that accuracy into the proper context of the movement of the stock, and made sure that players were catching significant moves.  I'm actually going to use your account as one of my test cases to see the impact of some of these decay methods on a player who has left the game for an extended period of time.  My guess is that your current accuracy will be highly dominated by the active picks and only marginally effected by the closed picks, and you will have dropped out of the upper 1%, but I won't know until I've actually run the tests.  I'm going to contrast the impact on your account with players like TMFEldrehad and TMFBabo (and many others) to see what kind of shuffling in the top 100 takes place over time.

-Xander 

 

Report this comment
#18) On August 18, 2011 at 7:45 PM, truthisntstupid (92.68) wrote:

totallyoblivious

Why penalize people who do not actively continue to make new picks?

There are those who are more inclined towards trading.   There are also those who are more inclined towards making picks based more on a long-haul perspective, based on company-and-business-specific fundamentals todrive their return over years rather than weeks or months.

If this game favored traders even more than it does now, I would go back to Seeking Alpha and stay there.

Report this comment
#19) On August 18, 2011 at 9:15 PM, devoish (96.47) wrote:

Have you considered just giving us a real world penalty?

Take a 35% tax in points from picks held less than one year, 15% from longer holds. That way gaming the system for accuracy would be balanced against lost points in your overall score. You would probably get 1/3 of my points off the board that way. This way "short term picks would also have to reach 7.9 before they left you with the 5 points needed for accuracy, long term 5.9. And as TSIF points out in reality you have to get far enough past those point totals that volatility doesn't get you in the twenty minutes it takes to close a pick.

A player could still game accuracy but it would be more difficult and less rewarding.

And I also agree with TSIF that it would be nice if a pick is closed at negative .07points we don't lose accuracy.

And while we are talking about scores, is it possible to add a column with our original starting price on a pick before the deductions for dividends?

Best wishes and thanks.

Steven

Report this comment
#20) On August 18, 2011 at 9:43 PM, ETFsRule (99.93) wrote:

Good ideas in #19.

With so many different ideas coming up in this thread, I'm thinking that there should be more than one "game" on CAPS. Let the players choose which set of rules they like the best.

Report this comment
#21) On August 19, 2011 at 12:01 AM, SultanOfSwing (96.36) wrote:

#12) So with your examples, it's like saying that when you close a winning 3X ETF pick, your accuracy credit slips away like sand through your fingers because a levered ETF = volatile = bad?  Whereas, when you close a winning large cap pick, like IBM, your accuracy stays intact like a lump of clay because a stalwart stock = low volatility = good?  I thought accuracy was a binary thng: you either get it or you don't.  There's no point in imposing extra artificial penalties on levered ETFs.  The contango (volatility decay) already takes care of that.  If you go down that road you can argue that small cap biotechs are too volatile for the game.  Over the counter & pink sheets are too volatile.  Companies in bankruptcy are too volatile.  Does anyone else see the slippery slope here?

Continue to allow anything that's traded on the open market to be pickable and you'll keep attracting new players.  If you monkey too much with the rules in an attempt to achieve a desired result, you might just wind up with a game full of zombie players.

Report this comment
#22) On August 19, 2011 at 1:21 AM, TMFCrocoStimpy (91.82) wrote:

Sultan,

I'm definitely not arguing that volatile is bad, and non-volatile is good.  However, I am looking at accuracy as something other than binary.  You are definitely looking at accuracy the way the original (and current) design was viewed - that if you are positive then you are accurate, and if you are negative you are not.  The difficulty comes in based on the random moves that come from the market.  Let's step back to when we didn't have the +5% threshold, so any stock closed with any positive value added to the accuracy.  Under those conditions it was simple to close virtually every pick when it reached some point of being positive.  In the case of having stocks that had the same expected return of the SPY we can get more than 95% of the picks to be positive at some point within a 3 month period, so just through random chance we could produce a 95% accuracy measure.

The question I've tried to address, and the original proposals above may be completely off base, was to view the accuracy of a closed pick of being made up of the probability of it being an accurate pick and the probability of it having achieved the closing score by chance.  For a given score (the putative +5% gain) the probability of reaching that score by chance is much greater than for a high volatility stock than it is for a low volatility stock, so we would award a greater degree of potential accuracy to the lower volatility stock.  Keep in mind, however, that for a large number of 3x ETF picks compared to a large number of IBM like picks, you will get a much greater percentage of the 3x ETF picks crossing the 5% threshold than the IBM like stocks.

You are hitting on the one theme that keeps us from making any moves very quickly (some would argue we are paced glacially): if solving one issue introduces another, then we may not have bought ourselves any progress, and we run the risk of alienating players any time a fundamental aspect of the system is changed.

Your feedback is appreciate.  Fool On!

Xander

Report this comment
#23) On August 19, 2011 at 7:31 AM, SultanOfSwing (96.36) wrote:

Thank you Xander.  I understand better now where you're coming from and I appreciate TMF's efforts to make the game better.

Fool On indeed!

Report this comment
#24) On August 19, 2011 at 8:26 AM, geneticbiscuit (98.78) wrote:

One relatively easy fix for the ultra-ETFs would be to only allow green-thumbing since most of the controversy surrounding them seems to be people who red-thumb to bank accuracy (and points).  This would force players to make a real call rather than just taking advantage of the time decay, and since they (mostly) have an inverse counterpart both bullish and bearish sentiments are represented.

Report this comment
#25) On August 19, 2011 at 8:46 AM, DaveGruska (95.84) wrote:

TMFCrocoStimpy -  thanks for the examples - they really help clear things up. I get what you're trying to do, but I also think that may be making that a bit too complicated.

I like devoish's suggestion, but maybe just simplify it to favoring long-term pick accuracy. Any move in a stock you pick in the short term is going to be more luck than anything else.

"we're the guys who suck the life out of the room at a party of mathematicians"

:-)



 

Report this comment
#26) On August 19, 2011 at 8:54 AM, SultanOfSwing (96.36) wrote:

biscuit,

If I had a vote, I would be ok with your fix.  If the ultras have to have a "restrictor plate" on them so people think it's more fair then so be it.  I'd just green thumb the ultras and still collect my points.  I would be sad losing my underdog charm, though.

truth guy, oblivious, are you guys cool with that?

Xander, I think I got a glimpse of how the U.S. Congress operates, but that's a whole other blog.

SoS

Report this comment
#27) On August 19, 2011 at 9:18 AM, geneticbiscuit (98.78) wrote:

SoS,

I have no problem with correctly calling the direction of a particular sector via green thumbing ultras - they are, after all, legitimate trading vehicles.  I do think that many people would be more hesistant using them due to the decay and minumum holding time, and ending them with a positive score would take more thought than just waiting it out like red thumbing...

I think you'd probably be grandfathered into your underdog charm (although getting it via shorting ultras is sort of like getting the blogger charm by simply writing a "hello world" blog), so no worries there.

Report this comment
#28) On August 19, 2011 at 10:54 AM, leohaas (31.35) wrote:

Nice ideas.

I am against changing any of the rules. With every set of rules, players will try and use the rules to their advantage. Go ahead and call that "exploiting". I see nothing wrong with playing by the rules. Only full anarchy cannot be exploited. Ask David in Qatar!

If we were to change the rules, I would close my current account and start a new one, "exploiting" whatever the new rules are. And with me, more and more players would find ways to game the game.

If your goal is to be transparent, I'd suggest some simpler way of scoring. Do away with the comparison to a benchmark (that confuses the h3ll out of at least half the CAPS players), and completely get rid of accuracy (you and binve are probably the only ones who fully understand it as is). And definitely do not introduce something that includes an exponential function. Although many aspects of the real world are exponential rather than linear (definitely growth, something we all want of our real-life portfolios), only very few of us appreciate that.

Report this comment
#29) On August 19, 2011 at 11:13 AM, latinoeconomist (26.40) wrote:

1.  I really like the CAPS game, as is.

2. I didn't know before that positive picks were not counted for accuracy unless the gain was more than 5 %.  Good to know (I had been wondering why my "accuracy" score was lower than my actual accuracy).

3.  I do think a "depreciation" of picks would make sense, if nothing else to encourage maintaining active picks.  Rather than any exponential function, a simple 33.3 % annual depreciation would be simple and easy to understand.

4.  I do NOT agree with the suggestion on #28 above, of doing away with the comparison with a benchmark (the S&P 500 index).  Then it would be VERY EASY to gain 5 % simply on market fluctuations.  I don't think it confuses anyone, and it is how most mutual funds, ETFs and hedge funds are compared.

5.   In summary, I would not advise making any major changes, except perhaps for the depreciation of previous years' scores, and perhaps excluding leveraged ETFs (or dividing their gains or losses by the factor of leverage; for example, if it is 3x, divide by 3).

Report this comment
#30) On August 19, 2011 at 11:27 AM, edwjm (99.87) wrote:

I DO agree with the suggestions of leohass in #28, especially on eliminating the comparison with the S&P 500 and not introducing something that includes an exponential function.

 

I am bothered, however, but the fact that no more attention is being paid to inactive accounts.  I have no problem with players who chose to leave their picks unchanged for an extended period of time because they still agree with all of them.  But I know that there are many who have simply stopped playing at all, yet their accounts remain.  Players who make no changes should at least be required to check in periodically and affirm that they still agree with the picks they have made.  Those who leave their selections untouched and don't even bother to check on them for weeks or even months should have their accounts automatically closed.  The utility of the CAPS system is severely compromised by stale data.

Report this comment
#31) On August 19, 2011 at 12:40 PM, goldminingXpert (29.52) wrote:

I think you should change accuracy so that the value of a pick toward your accuracy rating is determined by how long a pick is left open. That takes care of a good deal of the exploiting short-term bursts of volatility problem.

Accuracy decay is, in my view, a poor idea as it punishes people unnecessarily for short-term performance fluctutations. We judge Warren Buffet by his body of work, rather than his last 6 months of calls.  

Report this comment
#32) On August 19, 2011 at 12:59 PM, goldminingXpert (29.52) wrote:

One other point. Regardless of what changes you make, I will keep reloading my red thumbs on stocks I think are going to zero (cough most Chinese RTOs). Reason: By constantly moving down my cost basis, I can score far more than 100 points by the time the thing finally heads off to the pink sheets.

Red thumb at 10, falls to 5, I make 50 points. Close pick. Reopen. Falls to 3. I make 40 points. Close. Reopen. Falls to 2. I make 33 points. Close. Reopen. Falls to 1. I make 50 points. Close Reopen. Etc.

I'm already up to 173 points already, and I can keep doing that down to 50 cents now before a stock becomes unrateable. This is only fair (not a game exploit), since green thumbs naturally compound similarly to the longside and by constantly reloading, I am getting the same sort of compounding on my red thumbs.

If you tweak accuracy so that a picks accuracy value is based on how long it is open, I still get 1 years worth of accurate picks if a Chinese fraud takes 1 year to die, regardless of whether I red thumb it one time or a dozen. 

Report this comment
#33) On August 19, 2011 at 1:25 PM, dwot (40.93) wrote:

Well, when you say what you are planning on doing, I sure don't like it.  Everything I have left open is crap and under this new idea, which I think is better, the best play would have been to close stuff to reduce the emphasis of the inaccurate picks because they would also decay.

Why I think what you have suggested is better is because of how closing bad picks would decay in the accuracy scoring and it would increase motivation to do so and that would eliminate the effect on the overall scoring of the stock from essentially dead picks, of which I'd describe most of mine that are still open. 

Of course I don't find implementing this change almost 4 years after a lot of my picks have just been sitting there all that fair.

I tend to think it would be interesting to have a duel scoring system, one which represents your score from the last year, and then your long term score. 

Report this comment
#34) On August 19, 2011 at 1:44 PM, TMFCrocoStimpy (91.82) wrote:

devoish:

I haven't actually dug very deeply into the issue of score, so haven't given much thought to any corrections that would be "real world" like in terms of trading times.  With respect to it changing the threshold value that needs to be hit to count as "accurate" like we do know, I'm not sure that it would have a sufficient impact to really change the fact that high volatility stocks could achieve the threshold by chance much more easily than low volatility stocks, so we still wouldn't be getting a very clear measure of accuracy as intended (meaning a reflection of how likely a pick made by a player is to truly beat the market).

If there were some modification to decay accuracy based on time the pick were open and magnitude of the score, then something that was close at -0.07% score would decay very quickly, and have essentially no impact on the accuracy calculation from the time it was closed.  I find this type of symmetric treatment (picks that are barely positive or negative really cannot be said to reflect anything about the true return of the stock relative to the benchmark) appealing.  But that is because I'm a geek.

Features questions I'll try and collect - I'm not a UI person, but I'm going to be gathering up the ideas for new features that might make the system more engaging.

leohaas:

"Exploit" definitely wasn't meant in the negative sense, simply a designation of making use of the rules that exist.  What it turns out is that there is a potential weakness in those rules that we might want to correct, but of course the correction needs to be better than the original weakness!  Given the general premise of decaying accuracy on closed picks based on the length of time that the pick was open and on the "significance" of the score (score/volatility), are there obvious ways that you could see to game such a rule set?  That is at the heart of my original question, because I'd like to see how any new rule set could be exploited.

The use of a benchmark is probably one of the strongest components in the intended design of the system, given that the goal is to identify the stocks which perform better or worse than the average stock in the market place.  Stock picking, as opposed to money management, is really the focus of the CAPS system.  My interpretation of that is that we are not as concerned with the unknown drivers of the total market place but rather want to focus on how the stocks will perform relative to one another, to ferret out the differences in comparable qualities even if we are in a raging bull market or watching the financial system go to h3ll in a handbasket.  To do so requires us to subtract out a common market benchmark such as the SPY, which is really a very convienent and consistent proxy available to us.

latinoeconomist:

Just to be clear, it is closed picks which only count for accuracy if they have a score of +5% or higher.  If a pick is open than any positive value counts.

We will definitely be testing a constant decay idea rather than the more complicated formulations I posted to see if there is any marked in performance.

edwjm:

It isn't visible, but inactive accounts do lose their impact on CAPS ratings over time.  We don't advertise the formula, but in our rating calculation of individual stocks we take into account the length of time a pick has been open to calculate its influence on the final rating.  The longer a pick has been open, the less influence it has.  If an account gets abandoned then even though it is still visible to the community, and still receives a CAPS ranking, the actual impact on the stocks will eventually go to zero.  We are also looking at the difference between accounts that seem truly abandoned, and accounts where the players still come to the CAPS site but haven't made any picks (the latter still seem engaged, they just don't have any new opinions).  We will examine using different influence decay models on picks from the two different types of accounts.

GMX:

We may be closer to the same page here than you think.  The idea is to partially determine the decay of the accuracy based on how long picks are held open.  For WB, his past accuracy would be based on picks that had been held for many years, and consequently the accuracy measures from those picks wouldn't decay away for many years and his accuracy would not be based soley on his last 6 months of performance.  However, if you had somebody who was a short term trader and then they switched to a long term buy and hold strategy, you wouldn't want your estimation of their accuracy to be given by the short term trading results.

Fool On!

Xander

Report this comment
#35) On August 19, 2011 at 1:53 PM, TMFCrocoStimpy (91.82) wrote:

dwot:

Let me emphasize that we are not necessarily planning on doing any of these tweaks - this is an exploration phase, and I want to see how people would think about the potential changes, break them, beat them, etc :)

You have brought up an excellent point about the incentive to close losing picks that I forgot to mention.  Because the length of time that a pick's accuracy decays would be tied to the magnitude of the raw score, there is definitely incentive to close out negative picks so that the negative accuracy will decay away faster.  This may be one of the areas of weakness - will players start closing picks prematurely to try and protect accuracy?  Hard to say, but this is something that I would definitely like our strategic players to weigh in on.

Implementation of any of these changes (if any should occur) in a "fair" manner is a whole nother nut to crack - to right the ship, so to speak, would require some type of retro-active application, but how to do that and also reasonably accommodate the fact that you would have likely taken very different actions previously is something that will take a different debate (and it will likely be longer and more contentious debate than Keynesian's vs. Austrian's).

Fool On!

Xander

Report this comment
#36) On August 19, 2011 at 2:04 PM, SkepticalOx (99.50) wrote:

I honestly think something has to be done about the leveraged ETF's. I mean, how feasible is it to short ultrashorts in real-life (I haven't tried but I've heard it's not so easy to borrow shares). It totally warps the game because you would be at a disadvantage if you don't use them. 

I know this keeps being mentioned over and over again, and you even addressed it, but it really takes the whole stock-picking and analysis out of CAPS and rewards just mechanincal downthumbing of ultrashorts. 

Report this comment
#37) On August 19, 2011 at 2:13 PM, TMFCrocoStimpy (91.82) wrote:

#32 GMX:

Now you're giving me some strategic moves that we can really test out the decay ideas with.  For starters, I've always been fine with following a stock down into the abyss while closing and re-opening.  The fundamental point scoring strategy in CAPS is to minimize your cost basis, and to do that with shorts means that you need to continually close and re-open the pick as it moves down in price.  Every time you close and re-open the pick the risk of larger point losses goes up if the call goes against you, so it seems that the reward/risk amply works here.

What would change is the way the closed picks would change the accuracy contribution.  If you were to hold onto just a single pick for the year, then you would have a contribution to accuracy that was only equal to one pick, but it would decay away slowly after you eventually closed.  On the other hand, if you continually closed and re-opened, you would have the equivalent of many picks contributing to accuracy, each starting at a different date, but they would all be decaying at a far more rapid rate.  During the period of time that you were following the pick into the ground you would likely have stronger contributions to accuracy from the multiple picks, but once you stopped all together the decay of the multiple picks would drop away faster than that of a single pick.  So, you would be making a decision about points, short term accuracy and long term accuracy when deciding on the strategy you want to take.  This in turn may be influenced by how you see yourself picking stocks in the future as well - if you believe that you can keep finding sufficient picks to make short term plays on then it would be optimal to get all the points you can and bolster your accuracy constantly with new picks.  If, however, you thought that you had just a window of opportunity you might either strike for many points or try to garner some longer term accuracy since it wouldn't be clear how you would maintain the accuracy in the future.

Would definitely like to hear more on how you might approach your strategy.

Xander

Report this comment
#38) On August 19, 2011 at 2:29 PM, Melaschasm (57.74) wrote:

If you make accuracy decay, but the score remains, it just means that players will have to constantly farm accuracy rating.

If you make accuracy and score decay, it removes the advantage of players who have been successful over a long period of time.

I think it would be easier to give less weight to the accuracy score when calculating a persons rank, than to mess around with decaying accuracy scores and the confusion it may cause to players.  

Report this comment
#39) On August 19, 2011 at 2:51 PM, TMFCrocoStimpy (91.82) wrote:

#38 Melaschasm

That would indeed be one of the worst pathological cases that could occur, exacerbating the problem rather than helping it.  If, rather than nudging the players who are banking score and accuracy in small bites with shorter term picks into considering some additional picks with other strategies, these changes simply increased the number of short term picks, we would not be better off.  This is where trying to predict a players motivation and actions comes to light, and it is a very speculative area.  There are some players, without a doubt, who would not change their behaviour at all and maintain a high turnover rate of short term picks.  This would produce a reasonable degree of accuracy as well as an ever increasing number of points.  However, it requires a lot of effort.  As soon as such a player slacks off the accuracy begins to drop, though the points are maintained, so their overall CAPS ranking begins to sink fairly quickly.  Thus someone who pursued a large number of points via short term picks would have a harder time staying up on the leaderboard if they stopped making short term picks, because of the sinking accuracy rank.

As I mentioned previously I haven't given any real thought to the decay of score, but my gut reaction is to leave that alone.  One possible addition would be to include a metric of the significance of the score relative to the volatility, which is part of the accuracy decay as well.  This gives an open ended accumulation of raw score to be maintained, but then gives a second metric of accumulation of score that we might attribute to something other than chance.  These dueling metrics might get players to try and consider the impact of the timeframe they are holding picks for more thoroughly.

Fool On!

Xander

Report this comment
#40) On August 19, 2011 at 2:51 PM, goldminingXpert (29.52) wrote:

Let me think about comment #37 to get an idea of how I would strategize.

Comment #38: Either reducing the weight of accuracy, or making it harder to bank accuracy. Maybe raise the requirement for an "accurate" pick to be +8 or +10 points, or take someone else's suggestion from earlier up in the thread and stop people from red thumbing Utlras.

You simply can't make enough points from random accuracy banking of market noise to have a top profile. Those of us who tend to be serial accuracy bankers have been exploiting easy-to-use instruments such as penny stocks and leveraged ETFs because we know their long-term direction, thus making our picks nearly risk-free and likely to generate significant alpha.

When I have attempted to accuracy bank on dying sectors (homebuilders, banks, etc. in 2008 and '09) my strategy absolutely blew up in my face, ending up with -500+ point picks for things like GGP. Accuracy banking on instruments with no inherent bias (penny stocks, levETFs, China RTOs) is a fairly stupid game strategy. Stop people from picking pennies and ultra ETFS and the accuracy banking problem would solve itself. I think this is a better, less complicated, and more user-friendly solution than some sort of accuracy decay system.

That said, if you do move to accuracy decay, I will figure out some sort of strategy to accomodate it. 

Report this comment
#41) On August 19, 2011 at 2:53 PM, goldminingXpert (29.52) wrote:

Accuracy banking on instruments with no inherent bias ( that is, just about anything other than penny stocks, levETFs, China RTOs) is a fairly stupid game strategy

Report this comment
#42) On August 19, 2011 at 3:10 PM, TMFCrocoStimpy (91.82) wrote:

#40

"You simply can't make enough points from random accuracy banking of market noise to have a top profile."

Absolutely agree.  I wouldn't contented that anyone with a top profile got there due simply to banking accuracy with noise, but I would suggest that the high volatility provided a much higher measure of accuracy when used to bank than if you had a low volatility ticker that was following the same return trajectory into the toilet.  Ultimately, with a large group of stocks that are going downhill, the ones with higher volatility will present you with more opportunities to close at 5% gains than those with lower volatility, so you will get a higher accuracy rating even though the actual undeperformance is identical for the two groups.  And you are correct about this type of strategy blowing up in peoples faces - for all the accounts that are quite successfull at harvesting accuracy, we have a treasure trove of rotting account corpses that kept red-thumbing the ultra ETFs in late 2009 and 2010 who where convinced another crash was just around the corner.

Xander

Report this comment
#43) On August 19, 2011 at 3:49 PM, binve (< 20) wrote:

Xander,

Here are my thoughts as requested.

CAPS is not a strict portfolio simulation game. Each of the 200 picks you can have open at any time is not equivalent to 1/200th of a portfolio. Since you can close a pick for a gain or loss, and then either re-open it or open a new position for the same 'cost' (1 out of 200 slots), a CAPS score does not equate to 'profit' per se. Adding to that idea is the CAPS concept of 'outperformance'. Because you don't gain based on on score, you gain based on outperformance against the SPX. I know I am not saying anything new that anybody is not aware of, I am not setting up the issue as I see it.

So the point of the rating system and player participation is to rate issues (stocks, ETFs and funds) that will either outperform or underperform the SPX. However, this is not how players think, when a player makes a pick they are thinking first: "Will this pick be 'profitable'?" and secondly "Will the pick be profitable in the timeframe that I have selected?". However, this does not necessarily jive with the CAPS scoring system, which in a very strict sense means when a player makes a pick: Will the pick beat the SPX (regardless of whether the pick is actually 'profitable'), and Will the pick beat the SPX in the timeframe selected.

Since these two viewpoints don't jive all the time, we have the 'bugaboos' as you describe above. Again, nothing that hasn't been said before.

This results in the biggest problem, which is the fact that investors don't 'bank accuracy' in the real world. Investors care about profits. An investor does not necessarily care that they mistrade an issue 4 times for -2% each, if the entry on the fifth attempt was a bottom that subsequently saw the issue rise >10% (or something). This means that *Timing matters*, and investors should not be penalized for excercising proper risk management in getting the correct entry price on a stock (either buying in a falling market or on pullbacks). People use stops in real life for good reason. If a 1% loss on a stop preverves my ability to pick up the issue on a further discount, then that is viable risk management strategy. That proper unilization of opportunity costs should not be improperly penalized by a system designed to encourage investors to rate stocks.

With that framework, here are my comments and suggestions:

1. Time decay on the accuracy of closed picks.

I like this idea quite a bit. This means that investors will not be improprely punished on employing good risk management in a volatile market, nor will it improperly benefit those who bank accuracy.

The role of accuracy affecting the score to begin with is to 'dis-incentivize' overtrading. But there already is a 7-day rule in place anyways, and many short-term swing traders hold for much shorter periods than that. So I don't think it is a concern. You might make the arugment that there is some benefit to 'picking the right direction' as well as getting a positive score out of a pick. But I disagree. That is not how investors measure their own personal performance. They are looking for profits. Profits results from making more correct bets in *magnitude*, not in *number*, than incorrect bets. So the proxy for accurate profit making is from a positive scores, not from 'accuracy'. Successful picks beget accuracy, not the other way around.

So with that in mind, I would advocate for a *steep* decay rate on all closed picks. Depending on how you want to classify these descriptions (subject to what TMF deems is important), I would recommend seting up your decay function such that these trends are what follow:

Short Term picks (closed in less than two weeks after opening) or low score picks (± 5 pts): Decay to ~0 within 1 week
Medium Term picks (closed in less than 6 months after opening) or medium score picks (± 20 pts): Decay to ~0 within 2 months
Long Term picks (closed in > 1 year after opening) or large score picks (± 50 pts): Decay to ~0 within 4 months

With some appropriate smoothing/interpolation between the different classes

2. ETFs and Leveraged ETFs

Let them be. I hear people complaining on this all the time. It is like with margin (which is what a leveraged ETF amounts to). If you want to gain the reward with holding a Leveraged ETF, then you have to take the risk that your bet will go against you. And if you don't manage risk properly (just look at my old picks, and you will see what I am talking about :) ) you will be penalized with the score that results.

But when you combine this behavior with point 1 above regarding the accuracy decay, then you will get more realistic behavior of investors utilizing these instruments in their portfolios: i.e. You will see their use in CAPS dinimish greatly as the accuracy banking aspect becomes less important based on the Player rating changes.

3. Picking the same Issue multiple times

Let it be. If you can succesfully trade the same stock multiple times, good for you! If you can successfuly bank a positive score with it many times, good for you. You deserve any points you get, because trading the same stock ten times is just as hard as trading 10 different stocks ten times. But the key point here is that trading for accuracy will not be rewarded over the long term. If you are trading the same stock to bank accuracy, then idea 1 will prevent your score from benefitting that.

IMO, those are the big 3. And since 2 and 3 are just ideas to not change anything, then it is only 1 proposal.

Accuracy is the biggest 'break' in the system, and a steep decay rate on closed picks is the best compromise between giving a small amount of credit to good directionality choices for the short term, but not unfair and unrealistic punishments/benefits to long term rankings.

Report this comment
#44) On August 19, 2011 at 4:00 PM, TMFCrocoStimpy (91.82) wrote:

binve, appreciate your feedback.  I'm going to digest your points and continue with the thought experiments.

-Xander

Report this comment
#45) On August 20, 2011 at 7:32 PM, SolarisKing (< 20) wrote:

Caps is a stockpicking game, not a papertrading game. For one thing, there are many many paper trading sites and they barely differ, though some do have forums, etc. For another thing, what caps is supposed to do is rate stocks, not stockpickers. Now i agree that rating stockpickers is inherent in our paradigm, but the distinctions needs to be made clear so that we all pull in the right direction.

On the other hand, i feel that the players want/need the game to seem like it parallels investing (i know i think of it that way).

As far as some of the main issues above: On the leveraged ETFs; It might be harder than it sounds, but it would probably be more fair if you could only pick what you could trade. It follows the logic that if it doen't have volume Caps won't allow picks (nor can we pick many other investments).

As far as double/triple down; I want to say yes, but not really. Close the pick and re-open it if you want to, and a good picker could do that fine if our current system didn't disincentivise it.

As far as the accuracy decay; not bad. .. but i think it would stink to be the all time best at Caps but get booted for not picking often enough, or having made a bad set of calls that took a few years to get through.
  Also, if it is too easy to get around bad picks fools will make even more silly picks. The accuracy decay would need to be more inconvenient for the players, and good for the rating system, than vis a vis.

** I've always thought that there should be a reward for keeping the pick open for as long as you say. It's cool if you thought the pick was good for 5 years and then changed your mind, but if you are picking a stock with a time frame, you should get rated by the accuracy of your advice. 


   So i more support the "This years Top Fool", and "All Time Top Fool" idea. Keep both ratings; (1) Current Accuracy and (2) All Time Accuracy. To be the ATTF you would need some strength in both metrics.
   This would allow new players to get Top Fool, and old players to keep their game (even if you wiped the slate and started a two layer system tomorrow).

I also think that skeletons need a skeleton charm. Badly. This could be used to tell the difference between someone who hadn't picked in months and someone who had not signed in for years.
   Perhaps after one year of no sign on an account gets a skeleton charm, and after a whole year with a skeleton charm (total of two years of no sign on) the account could be removed from the rating system (even if it was kept visible).

More charms. More charms. More charms. 

There should be a dividend player charm, and maybe sector specialist charms.  A time frame accuracy charm, a +1,000 charm, a trader charm, and a way to search with multiple charms.

Accuracy Decay:

It would be nice if the players understood the formula (as was opined above). 

Is Caps for picking swings, or rating stocks? Make the minimum pick 1 month (two weeks?)? THIS IS PROBABLY A GOOD IDEA, but will probably get little support. It would reflect stockpicking more. .. . .. . naahhh.

````````````

You should pay twenty of us to fly to where you are and let us beta test your ideas daily in real time. Make sure to pick some advanced and some novice players.

 

 

 

Report this comment
#46) On August 31, 2011 at 9:03 AM, MKArch (99.70) wrote:

Xander,

 I have a solution for you. You need to make a small tweak to the existing system to nick score points in proportion to real life brokerage and tax costs. I don't think anyone would complain about that and it would help address the concern that winning strategies in the current system are not practical in real life.

Then you need a second alternate system based on score per pick. It needs to require a minimum number of picks per year to eliminate the players who created an alias in March 2009 and are sitting on 8-10 high score picks. Maybe you incorporate some sort of weighting based on the number of picks and the probability of an average pick score over the number in order to encourage constant participation. The explanation for this would be that you are more interested in finding players that can consistently find high score picks.

You could allow players to see both scoring systems or to choose one but be able to toggle to the other when ever they choose.  IMO this would mostly preserve the existing system with a minor tweak that I don't think many would object to while giving the players who care more about the quality of individual picks a system that they feel more comfortable with.

 Mike

Report this comment

Blog Archive

2011
August (3)
2010
February (2)

Featured Broker Partners


Advertisement