So, is there too much random in the match engine? To be honest, it’s impossible to answer on as it’s very subjective. There is no right or wrong here, it really depends on what you prefer (which may not be what the guy next to you prefer). But lets make an attempt to break things down a bit at least.
First we have to separate “real random” from what one could call “feeling of random”. Each week more than 1 million matches are played in Hattrick. The vast majority of them are “normal”, but these matches never get recognized. Quite naturally, I mean why would they? Still, you have to remember them as well.
But of course, every week there will be some few matches which are “unnormal” – and they will be recognized. A lot. And after a while recognizing matches like these, you will start to believe that there is a lot of random in the match engine. And each time you get to hear about another one, it will add to this feeling.
Now, don’t get me wrong here. I’m not blaming you for feeling things the wrong way – absolutely not. In my belief feelings are what’s important in games and what you should take seriously. But when we talk about this topic we have to try to separate these two from each other, as it’s two separate things. And that any attempt of curing will fail if we don’t know what we’re supposed to cure.
Ok, some facts about “real random” then: During the years we’ve deliberately decreased the random factor a couple of times. In all honesty, there’s also been cases where a feature has affected random as a side effect, and in some cases increased the random factor somewhat. But all in all, fact is that the random factor has been decreased over the years and that ratings are much more important these days.
There also been user studies supporting this, a while ago a user compared the number of “surprise matches” (taken from Alltid) from season to season, and there were indeed fewer cases of surprise matches in present times than in past times. Another user compared Hattrick matches compared to real life football, which naturally is in a way impossible to do but nevertheless very interesting. The result? Hattrick matches had fewer surprises than real life football.
But even though these things say there is less random in the game, the feeling you get from reading the forums is the total opposite. And it seems each time we’ve decreased the random factor in the game (and announced it), it seems this feeling gets bigger. And when you think about it, it’s perhaps not that strange: Because when we say something has been decreased – people start to look more intense on this and especially notice examples of the total opposite.
There is also one thing more that is important to be aware about. Teams are much more matured today compared to the past, we have a lot of more matches where team strengths are leveled and almost equal. And as a result, you may feel that random has a more decisive role.
To conclude: The random factor in the match engine is smaller today compared to the past, but the feeling of random in the game has increased (in my belief). And if we should learn something from the past, it’s that a real decrease of random in the game engine doesn’t necessarily mean users will feel it’s less random – it can actually have a fueling effect (I would actually say that the last time we decreased it, it had just that effect). So, before we do anything about the “real random” we should do something about how you experience random instead. To explain it better, to make you see what’s (un)luck and what really wasn’t. That’s where we have to start, and I think we will be able to have something to show you during this year.
At the end of the day I personally think the amount of random (real random) is at a pretty good level, in case you’re wondering. The real challenge for us lies within explaining it better and making it feel more logical, and I would also say that we probably will have to look at the match engine in regard to leveled/equaled teams at some point. But lets take it one step at a time. And remember, no matter what will happen and how much random decreases – there will still be surprising results. And they will be talked about.
This is a very important issue in the game. I think random is not bad, because also in the real football matches random is always present. I suggest you that you have to increase your communication about random in the commentaries of the matches. Explain random through words in the commentary would be better and a lot of people, I think, will be confident in it. Also partial ht-stats during the match (after 20 minutes, after 40, 60, 80, ecc) will be useful. I love Hattrick, but you have to be more transparent in random stuff if you want to keep your users in the game for long time! 🙂
I agree with you SANTE.
The most important problem with random is this “dark” part. Only a few people knows anything about these factors during the match, and sometimes it’s impossible to say if the match I lost its because random factor or my fault.
It would be very interesting if the reports were more accurate about this.
i was very pleased with last week’s post, but this week i must say i was hoping for more. 😉
although, i think it’s very important that you delivered as promised, and i applaud you for that.
for instance, i was hoping to hear about set minimum (and also maximum) probability for something (anything) to happen in match engine… other games also use this concept, so it can’t be your big secret, right?
cause i remember one discussion on global about this, years ago, also involving at least one member of ht-team… later, i’ve tried 2-3 times to lure you (you and flameron, i guess) into discussion about this, but with no luck… so this time my chances are also not so good, i guess. 😉
good luck and keep the good stuff coming!
Hi Tjecken 🙂 I agree with Sante Pollastri about you having to be more transparent about it. And let me use your example. You say “Because when we say something has been decreased – people start to look more intense on this and especially notice examples of the total opposite.” and i ask you: When did you announced that you decreased the Random factor? I don’t rebemeber reading anything about it….so you example doesn’t apply to me and i’m guessing many other users.
I am realy glad you are exposing some thoughts about Random but i still feel doubtful about it. You say “But all in all, fact is that the random factor has been decreased over the years and that ratings are much more important these days.” and i see poor attacking ratings scoring against very good defenses, and very good attacking ratings not scoring (or even scoring 1 goal) against porr defense ratings. In both examples with a balanced midfield.
I guess i have to live with this level of Random a couple more years…lol…and hopefully you will (really) decrease it in the future 🙂
One final question: are you finished with this issue or can i expect more news about it? My belly is not full 🙂
It’s been announced several times, but the last time was when we launched the new engine (Dec 2009) as far as I remember.
I’ve said what I wanted to say, so I would say I’m done for now. But then again, it also depends on if something comes up on the forum which I want to address. 🙂
“Because when we say something has been decreased – people start to look more intense on this and especially notice examples of the total opposite.”
I think that it is not really fair to say so. If we say that announcing sth will make it more severe, we might say when a government say “unemployment rate is reduced to 3%” the problem will actually get worse, as people will look more for it and they will see this 3% more (not the other 97% percent). I agree that it will draw a little bit more attention on the matter, but the fact is that it will clear the air too. Also it will decrease the hard feeling, as we know that you are doing sth and it has been “taken care” of. Although I am a reading type [in Hattrick], I didn’t really recall you pointing out this issue many time. I had sth in the back of my head, which was probably what you said back in Dec 2009. Actually, my concern always was “why HTs aren’t doing anything on this matter?”; so maybe you should think twice before concluding that the problem is based on giving publicity or overemphasis on the matter.
Although I don’t have the inside knowledge that you have, but I suppose that you might be right on the side effect issue which you said. Maybe you have decreased the random, but other tweaks increased this random! Or maybe we are seeing less unpredictable matches, but more “unfair” unpredictable matches?
In regard to a solution, I think that we should move toward a more logical unpredictability! what I mean? you should give us more valid reasons to increase out team`s “chance” of wining; namely by relying more on factors like specialties, experience, free-kick or any new factor. In this way, we can have a influential rule in increasing the number of positive chances in the game and increasing the chance of winning game [specially when we are inferior to the opponent]. Also, the role of analyzing the opponent, predicting their way of play, disabling them, and etc. must be strengthened in the game. In this way, we are not sure of the outcome of the match (as we are not sure if our prediction/opponent’s prediction is right or wrong) and also we have had a valid role on the outcome of the match (when we decide to increase the number of players with head, experience, free-kick, etc!).
“Another user compared Hattrick matches compared to real life football, which naturally is in a way impossible to do but nevertheless very interesting. The result? Hattrick matches had fewer surprises than real life football.”
I had a friend which always said “In real football, the superior team doesn’t always win. But the crucial difference between real football and HT is that in real football, teams play 36-40 matches a season and then see where they are and in HT we play 14 matches and see where we are. When a team lose 2 match out of 38 match based on bad luck, they are able to make up for it as it means losing like 5.2% of points of the whole possible points, but losing 2 match out of 14 matches means 14.2% points of the whole possible points!!! So we should see surprising matches three times less in HT compared real football.”
It is worth mentioning that this former friend lost his interest in the game after he failed to promote to tier one after several tries (which were unsuccessful based on these surprising matches)
“The real challenge for us lies within explaining it better and making it feel more logical, and I would also say that we probably will have to look at the match engine in regard to leveled/equaled teams at some point.”
I completely agree with you on this. It happened many times to me that other users ask me explain to them the reason of their defeat [while they are the superior team]. Sometimes these randoms, were because of SE’s and I easily explain it to them and they, somehow, accept the defeat. But when it happens for some unknown reasons and I say it is how HT is and sometimes you lose games when you are the leading team (random) and they get along. But when this random thing happens a lot, they start to hate it’s influential rule. So, I think, explanatory report/details along with ways of improvement (SP, Specialty, etc.) is certainly an effective solution to the problem.
Thank you
Keep the random, it is important to keep the game interesting.
I really appreciate your arguments. What I think may be really discouraging, especially for less experienced players, is the existence of very big, unjustified defeats. Say, when you’re a clear favourite playing at home, it is acceptable to lose 0:1, 0:2 to an underdog from time to time. But 0:5 looks very unpropable.
I think it’s the same for the winning side – when I score 8-9 goals almost every match in a low division, I lose some piece of interest in the coming games. I guess it wouldn’t happen if the results remained a bit lower, no matter how bad my enemy is (for example 3:0, 4:0 being most propable outcomes, when my chances for victory reach 95%)
How about inventing some mechanism for the match engine to prevent too one-sided results from occuring? It wouldn’t affect the surprise results ratio, but would make all these bad defeats more acceptable.
Generally speaking, there are more goals scored in Hattrick than in real life (by design choice really, it’s fun with more goals). But as you say and I agree with that, when a team is clearly better they tend to score 8-10 goals which only rarely happens irl. We’ve thought of some mechanism around this, but it has never been prioritized.
But remember, decreasing the number of goals in a match will increase the random factor.
Because of the small number of goals in a match, there are more suprising matches in football games like in other ball games, for example handball or basketball.
Well, this is a nice-written personal opinion, but i really can’t see why an opposite opinion cannot be published. It would be better if you published a user’s opinion who feels different about the random, as well. So there would be an open discussion, to help the developers making the game better. Not all users reading the forums, but many of them reading the articles in hattrick news, so all opinions posted in a polite way, had to be published.
Well, in this channel we write what we think and we state our opinion. I mean, that’s what this blog is for. There are lots of places where other opinions can be heard, the forums being one example (I recommended a forum thread myself in the last blog post) but foremost we have Hattrick Press as such a channel.
i agree to sante pollastri.. more transparent would be nice.. as well in other case..
i have an excellent coach, train 100% witz 18 % stamina, but my formlevel is passable (low) .. there are things in this game which are not possible… there is a guide (hattrick book) and then you compare with other teams and you see, that an excellent coach is useless.. and in my opinion ht should push excellent coaches and give them more power…
Random part is necessary in a football game, and surely we all agree with the complications of its implementation…To convert this sport into mathematical equations is more subjective than the feeling of “excess of random”, in fact.
Maybe the most important point is the way we see real football. We analyse a real match in two ways:
Before the match we quantify the potential of both teams, and we could explain this in HT terms (excelent attack, poor defense, etc)
After the match, and depending on the different events and the unpredictable situations that occurs during it (the real world randomness, an enormous number or uncountable variables with little repercussion each one, but in sum they affect seriously the final result), we re-qualify the actual ratings so them fits the real development of the match.
So here is the problem, HT only shows the objective potential ratings resulting of the players skills, not the actual performance of the team after the application of the random part in the process of match calculation…the way this random part has affected the match situations at every moment is transparent to the user.
In other words, when we watch real football we can see an outstanding stiker missing a goal because of a mistaken decision in the moment of the shot, affected by a real life randomness…In HT world we see that we lost a goal chance by the left wing, where we have a mythical attack against an aceptable defence, for example. Our minds can’t understand this, and frustration comes.
Personally, i agree the way you implemented randomness in the context of a virtual environment. Probably, the real challenge is not this implementation itself, but to explain to a huge community how this feature affects to everyone and why.
I agree with the user that wrotes before me, more detailed statistics would be very useful. Maybe partial stats every 15 minutes and every substitution, with % of sucess in goal chances and possesion, would help users to realise why things happens, or not…No place for doubts then.
The final point is…Would you play a game where you can lose without merit it? Everyone has an answer for this question, and unfortunately you can’t do nothing for changing this.
Really nice article! I would say that this should be obligatory reading for all pre-intermediate and intermediate managers! (including the previous article). Great job here.
Totaly with SANTE POLLASTRI comment.
The random is OK, and is Good for the game. I have no doubt about it (And I was “victim” of it this season!). But there is a miscommunication or misunderstanding about it.
I think I understand your intentions (correct me if I’m wrong).
I think we should somehow signal the events that had been affected by random.
not in a completely objective way, but with a little imagination (Citing an error of the referee, for example).
Or…
percentage of “exposure” to the random, the match had 😉
Also partial ht-stats during the match would be important to understand the game. which sometimes seems random, it is not.
Best wishes and good job!
Antonio
Improved match ratings presentation? :p Anyway I do honestly agree that the “random” is acceptable in Hattrick. There are many problems in hattrick, the random is not really one of them.
At the very least not a major one. Here, I mean random here as the probability to get a chance, probability of it going to each sector, and probability to score.
Fix stuff like SE, the tactics, and the super imbalanced ratings contribution of players making central attack effectively useless, and the match engine will be a lot better.
No prob here with “real” random, especially not with leveled teams (in RL after quarters of the CL e.g. luck – in form of posts, refs, lesions etc. – becomes a major factor). Fine wih me (this season I won against a better team by SE and was happy, I lost against a level team by SE, the way the cookie crumbles), and especially from the POV of “abstractions” to mark the game outcomes.
The “subjective” random, though, is another story:
I can understand a guy playing against a bot can be frustrated, when, preparing well and otimizing probabilities he then loses as the bot wins with “his 4%”. Of cause I undersand that 4% are 4%, so he *will* win in 4 of 100 games (and it could be 6 in a row in 200).
Still, the frustration is real and somewhat founded: In RL, if an underdog team prepares well and beats one or more biggies (and be it for mere guts, see 2b league – third – Mirames in the Spanish cup this season taking out one first leaguer after the other for some streak), it is fun understandable and reason for raising your hat to the mister and the players. If the same thing happens in HT it should get the same acclaim, methinks.
OTOH, if you play against a team whos manager does *not* prepare and does *not* chose tactics or does anything to beat you as biggie, it should not happen, period. As in Rl.
Maybe there is a spot to look at (screen online time, preparation, setup, etc.) and add a penalty (or reward) to *active*, win searching managers (my idea would have the same setup as now for active managers, but a – small – probabilities penalty to those that play a whole season with their standard setup).
Another aspect – and this is a change of theme somewhat – is that we dont (as managers) control the “12TH” player (the fans) or only in a standard abstraction with having more posession playing at home. In RL, when the shit hits the fan or the stakes are high, managers do something to get the fans support in an exceptional way, e.g. by lowering ticket price to fill the stadium, running campaigns to get the support, etc. Sometimes this shows results, sometimes not, but it is a manager tool to influence probabilities that I am missing.
Just a few thoughts,
Rattler
I completely agree with author. Balance of random today is pretty good. But in the same time random is a little bit one-sided. I mean there is seasons when nothing unexpected happens and seasons when we have 5 from 14 unexpected matches.
For example I had one season when whole season 4-5 players from general composition was with injury. Or some times random lose happens with opponent which is extremly weaker. That couldn’t happens even in real life.
So I think, things like that are picking up questions about “unfair” or “too much” random.
Hey Tjecken,
more statistics and better transparency are the one side of making random more understandable for the managers. The other side is the “feeling” of a manager when a event with small probability.
If a really bad attack scores against a very high defense, a manager feels a little bit stultified when the match report says something like this:
” In minute 65 The middle defenders had no chance when Player XY dribbled around them easily and made the goal with a cool smile on his lips.”
A sentence like this a very frustrating for the manager with the high defense. In my view it would accept it with a better feeling if the text would be something like this:
“The middle defenders marked Player XY very close, so that he had no chance to score. But in minute 65. an unlucky stray ball landed in front of his feed, the suprised striker only made a kiddy shot, but that was enough to beat the even more suprised keeper.”
The result is the same, the opponent made a goal – but the feeling of the manager reading the second text would be much better.
So I think implementing several text modules for serveral scoring probabilitys is another way to make the feeling for the game much better, espacially for users who don’t want to get evertything explained with math and numbers.
Greetings from Germany
Stefan
ASP.NET-Workmate 😉
Is Random that makes me ansious to see the results of the matches!
Don’t that that from me…
I like underdogs victories in real life and in HT.
In Portugal (see the HT Cups Winners, you’ll see the reandom) we talk too much about it… But now we even have a “sweet” name for random: Randão! Is a personal name, that is said in a similiar way:
Don’t kill Randão! Long lives Randão!
This feeling of random would decrease if we had better match ratings. And I have to say I’m getting bit pessimistic.